From 8ad0e2adeed294bd98910e65d162581d54276263 Mon Sep 17 00:00:00 2001
From: Robert Shaw <rshaw@neuralmagic.com>
Date: Sun, 29 Dec 2024 10:38:27 -0500
Subject: [PATCH] added new configs

---
 ...dtype=fp8_w8a8,block_shape=[128, 128].json | 146 ------------------
 ...dtype=fp8_w8a8,block_shape=[128, 128].json | 146 ++++++++++++++++++
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  66 ++++----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  60 +++----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  62 ++++----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  70 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  60 +++----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  70 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  80 +++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  66 ++++----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json | 146 ++++++++++++++++++
 ...type=fp8_w8a8,block_shape=[128, 128].json} |  74 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  70 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  72 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  80 +++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  70 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  82 +++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  76 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  76 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  68 ++++----
 ...type=fp8_w8a8,block_shape=[128, 128].json} |  62 ++++----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  70 ++++-----
 ...type=fp8_w8a8,block_shape=[128, 128].json} |  70 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  66 ++++----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  60 +++----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  76 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  68 ++++----
 ...type=fp8_w8a8,block_shape=[128, 128].json} |  84 +++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json | 146 ++++++++++++++++++
 ...type=fp8_w8a8,block_shape=[128, 128].json} |  46 +++---
 ...dtype=fp8_w8a8,block_shape=[128, 128].json | 146 ++++++++++++++++++
 ...dtype=fp8_w8a8,block_shape=[128, 128].json | 146 ++++++++++++++++++
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  80 +++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  86 +++++------
 ...dtype=fp8_w8a8,block_shape=[128, 128].json | 146 ++++++++++++++++++
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  76 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  74 ++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json |  78 +++++-----
 ...dtype=fp8_w8a8,block_shape=[128, 128].json | 146 ++++++++++++++++++
 39 files changed, 2121 insertions(+), 1245 deletions(-)
 delete mode 100644 vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
 create mode 100644 vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=1536,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=1536,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (78%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=1536,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (79%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=2048,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (84%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=2048,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (81%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=2304,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (83%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=2304,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (80%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=24576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (77%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=24576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (75%)
 create mode 100644 vllm/model_executor/layers/quantization/utils/configs/configs/N=256,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
 rename vllm/model_executor/layers/quantization/utils/configs/{N=7168,K=16384,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json => configs/N=3072,K=1536,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json} (81%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=3072,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (80%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=3072,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (78%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=32768,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (76%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=32768,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (77%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=36864,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (77%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=36864,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (79%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=4096,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (77%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=4096,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (79%)
 rename vllm/model_executor/layers/quantization/utils/configs/{N=7168,K=1152,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json => configs/N=4608,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json} (78%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=4608,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (78%)
 rename vllm/model_executor/layers/quantization/utils/configs/{N=4608,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json => configs/N=512,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json} (78%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (83%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (81%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=7168,K=1024,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (76%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=7168,K=1024,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (79%)
 rename vllm/model_executor/layers/quantization/utils/configs/{N=7168,K=16384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json => configs/N=7168,K=1152,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json} (78%)
 create mode 100644 vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1152,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
 rename vllm/model_executor/layers/quantization/utils/configs/{N=7168,K=1152,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json => configs/N=7168,K=128,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json} (82%)
 create mode 100644 vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=16384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
 create mode 100644 vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=16384,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=7168,K=18432,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (75%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=7168,K=18432,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (75%)
 create mode 100644 vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=7168,K=2048,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (76%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=7168,K=2304,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json (75%)
 rename vllm/model_executor/layers/quantization/utils/configs/{ => configs}/N=7168,K=2304,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json (77%)
 create mode 100644 vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=256,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json

diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
deleted file mode 100644
index 256c4f6d0f9a0..0000000000000
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ /dev/null
@@ -1,146 +0,0 @@
-{
-    "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "2": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "4": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "8": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "16": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "24": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "32": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "48": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "64": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "96": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "128": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "256": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "512": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "1024": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "1536": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "2048": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "3072": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    },
-    "4096": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
-        "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
-    }
-}
diff --git a/vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=1536,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=1536,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
new file mode 100644
index 0000000000000..6496a38fba8ae
--- /dev/null
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=1536,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -0,0 +1,146 @@
+{
+    "1": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "2": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "4": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "8": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "16": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "24": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "32": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "48": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "64": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "96": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "128": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "256": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "512": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1024": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1536": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "2048": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "3072": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "4096": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    }
+}
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 78%
rename from vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index f89fe2a12cc50..3618053b65831 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,19 +1,19 @@
 {
     "1": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,7 +21,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
@@ -29,47 +29,47 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
@@ -77,7 +77,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "128": {
         "BLOCK_SIZE_M": 64,
@@ -85,15 +85,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,46 +101,46 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 79%
rename from vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index afe0178de0886..46a982f5ee9a4 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=1536,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=1536,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -5,7 +5,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
@@ -13,7 +13,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,39 +21,39 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
@@ -61,7 +61,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
@@ -69,31 +69,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,7 +101,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,38 +109,38 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=2048,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 84%
rename from vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=2048,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index 45d0a6848355f..035ec027fa566 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=2048,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,19 +1,19 @@
 {
     "1": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,7 +21,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
@@ -29,7 +29,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,47 +37,47 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
@@ -85,31 +85,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
@@ -117,7 +117,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
@@ -125,7 +125,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,7 +133,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
@@ -141,6 +141,6 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=2048,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 81%
rename from vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=2048,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index 3dcf850600d53..8b49f2781cb54 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=2048,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=2048,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,19 +1,19 @@
 {
     "1": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,15 +21,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,95 +37,95 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,7 +133,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
@@ -141,6 +141,6 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=2304,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 83%
rename from vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=2304,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index d88bc4b198aec..851bc9f9f0b50 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=2304,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -5,7 +5,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "2": {
         "BLOCK_SIZE_M": 64,
@@ -13,7 +13,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,7 +21,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
@@ -29,7 +29,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,15 +37,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
@@ -53,15 +53,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
@@ -69,23 +69,23 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
@@ -93,7 +93,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,46 +101,46 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=2304,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 80%
rename from vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=2304,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index 3982cae83a00c..d1227c2157990 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=2304,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=2304,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -5,47 +5,47 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
@@ -53,39 +53,39 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
@@ -93,7 +93,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,46 +101,46 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=24576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 77%
rename from vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=24576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index e44ad6757ba86..1c61451fb34e5 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=24576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,43 +1,43 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
-        "num_warps": 4,
-        "num_stages": 2
+        "GROUP_SIZE_M": 1,
+        "num_warps": 8,
+        "num_stages": 4
     },
     "2": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
-        "num_warps": 4,
-        "num_stages": 2
+        "GROUP_SIZE_M": 1,
+        "num_warps": 8,
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "24": {
         "BLOCK_SIZE_M": 64,
@@ -45,15 +45,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "48": {
         "BLOCK_SIZE_M": 64,
@@ -61,39 +61,39 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,7 +101,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,38 +109,38 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=24576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 75%
rename from vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=24576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index 828ea488c37fa..63e661c80de6a 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=24576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=24576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,19 +1,19 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
+        "num_warps": 8,
+        "num_stages": 4
     },
     "2": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
-        "num_warps": 4,
-        "num_stages": 2
+        "num_warps": 8,
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,7 +21,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "8": {
         "BLOCK_SIZE_M": 64,
@@ -29,31 +29,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "48": {
         "BLOCK_SIZE_M": 64,
@@ -61,7 +61,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "64": {
         "BLOCK_SIZE_M": 64,
@@ -69,15 +69,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
@@ -85,15 +85,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,7 +101,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,38 +109,38 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/configs/N=256,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=256,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
new file mode 100644
index 0000000000000..cf354037903c0
--- /dev/null
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=256,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -0,0 +1,146 @@
+{
+    "1": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "2": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "4": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "8": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "16": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "24": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "32": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "48": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "64": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "96": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "128": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "256": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "512": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "1024": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "1536": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "2048": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "3072": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "4096": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    }
+}
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=16384,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=3072,K=1536,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 81%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=16384,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=3072,K=1536,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index 80c87f32c7f8c..eccb86a76df0d 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=16384,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=3072,K=1536,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -3,25 +3,25 @@
         "BLOCK_SIZE_M": 16,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
@@ -29,7 +29,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,110 +37,110 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=3072,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 80%
rename from vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=3072,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index 5801b5d389fd4..88af48431d8b8 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=3072,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -3,17 +3,17 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,23 +21,23 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "24": {
         "BLOCK_SIZE_M": 64,
@@ -45,15 +45,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "48": {
         "BLOCK_SIZE_M": 64,
@@ -61,31 +61,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
@@ -93,7 +93,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,46 +101,46 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=3072,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 78%
rename from vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=3072,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index e55b7758ad53d..dd069726d7ed4 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=3072,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=3072,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -5,31 +5,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,55 +37,55 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
@@ -93,7 +93,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,46 +101,46 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=32768,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 76%
rename from vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=32768,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index b20bd9f2fa0fd..56b939e52fac3 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=32768,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,99 +1,99 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
-        "num_warps": 4,
-        "num_stages": 2
+        "GROUP_SIZE_M": 1,
+        "num_warps": 8,
+        "num_stages": 4
     },
     "2": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
-        "num_warps": 4,
-        "num_stages": 2
+        "GROUP_SIZE_M": 64,
+        "num_warps": 8,
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
-        "num_warps": 4,
-        "num_stages": 2
+        "GROUP_SIZE_M": 1,
+        "num_warps": 8,
+        "num_stages": 3
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,7 +101,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,7 +109,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
@@ -117,15 +117,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,14 +133,14 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=32768,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 77%
rename from vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=32768,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index e51668ff66aaf..63d9a0bf5d79d 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=32768,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=32768,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,35 +1,35 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
+        "num_warps": 8,
+        "num_stages": 4
     },
     "2": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
-        "num_warps": 4,
-        "num_stages": 2
+        "num_warps": 8,
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_N": 256,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
-        "num_warps": 4,
-        "num_stages": 2
+        "GROUP_SIZE_M": 1,
+        "num_warps": 8,
+        "num_stages": 3
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,31 +37,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "64": {
         "BLOCK_SIZE_M": 64,
@@ -69,31 +69,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,7 +101,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,7 +109,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
@@ -117,7 +117,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
@@ -125,7 +125,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,14 +133,14 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=36864,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 77%
rename from vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=36864,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index fa019d1bb0d48..7fa398c15a2a5 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=36864,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,27 +1,27 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
-        "BLOCK_SIZE_M": 16,
+        "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "8": {
         "BLOCK_SIZE_M": 64,
@@ -29,47 +29,47 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "24": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "32": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "64": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "96": {
         "BLOCK_SIZE_M": 64,
@@ -77,70 +77,70 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=36864,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 79%
rename from vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=36864,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index abf60154d8b88..f15d8f64c7090 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=36864,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=36864,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,35 +1,35 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
-        "num_warps": 4,
-        "num_stages": 2
+        "GROUP_SIZE_M": 16,
+        "num_warps": 8,
+        "num_stages": 3
     },
     "2": {
-        "BLOCK_SIZE_M": 16,
+        "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,63 +37,63 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,7 +101,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,38 +109,38 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=4096,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 77%
rename from vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=4096,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index dce5887047674..cd3e07804fdec 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=4096,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,115 +1,115 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
@@ -117,15 +117,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,7 +133,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
@@ -141,6 +141,6 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=4096,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 79%
rename from vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=4096,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index dbc9c91635439..9d5a329d7466a 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=4096,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=4096,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,19 +1,19 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,15 +21,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,31 +37,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
@@ -69,31 +69,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
@@ -101,7 +101,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,23 +109,23 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,14 +133,14 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1152,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=4608,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 78%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1152,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=4608,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index 77b01c3bcc181..96e1594a3eabb 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1152,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=4608,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,19 +1,19 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
+        "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,55 +21,55 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
@@ -77,7 +77,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
@@ -85,23 +85,23 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,7 +109,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
@@ -117,7 +117,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
@@ -125,22 +125,22 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=4608,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=4608,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 78%
rename from vllm/model_executor/layers/quantization/utils/configs/N=4608,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=4608,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index ab84d436021c2..5ffd367df833d 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=4608,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=4608,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -3,17 +3,17 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,55 +21,55 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
@@ -77,7 +77,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
@@ -85,7 +85,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
@@ -93,15 +93,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,38 +109,38 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=4608,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=512,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 78%
rename from vllm/model_executor/layers/quantization/utils/configs/N=4608,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=512,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index 4adc50acf36ce..eabc423949a24 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=4608,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=512,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,43 +1,43 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
+        "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "2": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "24": {
         "BLOCK_SIZE_M": 64,
@@ -45,15 +45,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
@@ -61,7 +61,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
@@ -69,7 +69,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "96": {
         "BLOCK_SIZE_M": 64,
@@ -77,47 +77,47 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "512": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
@@ -125,22 +125,22 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 83%
rename from vllm/model_executor/layers/quantization/utils/configs/N=576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index 30c54b265f8f5..51e237b91b8e7 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -3,17 +3,17 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,15 +21,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,39 +37,39 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
@@ -77,31 +77,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "512": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,38 +109,38 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 81%
rename from vllm/model_executor/layers/quantization/utils/configs/N=576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index 42031762b49b7..6280219c9ee7d 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -3,33 +3,33 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
@@ -37,15 +37,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
@@ -53,7 +53,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
@@ -61,7 +61,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
@@ -69,7 +69,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
@@ -77,15 +77,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
@@ -93,15 +93,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "512": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,23 +109,23 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,14 +133,14 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1024,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1024,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 76%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1024,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1024,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index 7c81a79ad5834..40c01c0b92b4b 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1024,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1024,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -3,129 +3,129 @@
         "BLOCK_SIZE_M": 16,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,14 +133,14 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1024,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1024,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 79%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1024,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1024,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index 844cb0d7cebb3..c6fd3659799bc 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1024,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1024,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,19 +1,19 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,23 +21,23 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
@@ -45,15 +45,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
@@ -61,15 +61,15 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
@@ -77,47 +77,47 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
@@ -125,7 +125,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,14 +133,14 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=16384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1152,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 78%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=16384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1152,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index 45f4174640d30..160f12ed3f95a 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=16384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1152,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,51 +1,51 @@
 {
     "1": {
         "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
@@ -53,94 +53,94 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1152,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1152,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
new file mode 100644
index 0000000000000..e5c4a1d2c94e5
--- /dev/null
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=1152,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -0,0 +1,146 @@
+{
+    "1": {
+        "BLOCK_SIZE_M": 16,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "2": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "4": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "8": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "16": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "24": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "32": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "48": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "64": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "96": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "128": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "256": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "512": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1024": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1536": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "2048": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "3072": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "4096": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 3
+    }
+}
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1152,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=128,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 82%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1152,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=128,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index afbcd7a2ba58c..2bf5eb27e3820 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=1152,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=128,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -5,10 +5,10 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2": {
-        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_M": 16,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
@@ -19,15 +19,15 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
-        "num_warps": 4,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 8,
         "num_stages": 2
     },
     "8": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
         "num_stages": 2
     },
@@ -35,7 +35,7 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
         "num_stages": 2
     },
@@ -43,7 +43,7 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
         "num_stages": 2
     },
@@ -59,7 +59,7 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
         "num_stages": 2
     },
@@ -67,17 +67,17 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
         "num_stages": 2
     },
     "96": {
-        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_M": 128,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "128": {
         "BLOCK_SIZE_M": 64,
@@ -91,20 +91,20 @@
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
         "num_stages": 2
     },
     "512": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_M": 128,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
-        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_M": 128,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
@@ -112,7 +112,7 @@
         "num_stages": 2
     },
     "1536": {
-        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_M": 128,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
@@ -120,15 +120,15 @@
         "num_stages": 2
     },
     "2048": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_M": 128,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
         "num_stages": 2
     },
     "3072": {
-        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_M": 128,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
@@ -136,8 +136,8 @@
         "num_stages": 2
     },
     "4096": {
-        "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_M": 128,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
diff --git a/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=16384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=16384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
new file mode 100644
index 0000000000000..0a1e14cffbb2a
--- /dev/null
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=16384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -0,0 +1,146 @@
+{
+    "1": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 8,
+        "num_stages": 5
+    },
+    "2": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "4": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "8": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "16": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "24": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "32": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "48": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "64": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "96": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "128": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "256": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "512": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1024": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1536": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "2048": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "3072": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "4096": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    }
+}
diff --git a/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=16384,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=16384,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
new file mode 100644
index 0000000000000..15b1c93f60fc5
--- /dev/null
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=16384,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -0,0 +1,146 @@
+{
+    "1": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "2": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "4": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "8": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "16": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "24": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "32": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "48": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "64": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "96": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "128": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "256": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "512": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1024": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1536": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "2048": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "3072": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "4096": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 3
+    }
+}
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=18432,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=18432,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 75%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=18432,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=18432,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index c8ef92170e8f2..8ff12e64c172f 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=18432,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=18432,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,107 +1,107 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "16": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "32": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "64": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
@@ -109,31 +109,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
@@ -141,6 +141,6 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=18432,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=18432,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 75%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=18432,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=18432,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index 1abab67d093f4..4532f93681e2b 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=18432,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=18432,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,123 +1,123 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
@@ -125,22 +125,22 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
new file mode 100644
index 0000000000000..ca7f32b9552b4
--- /dev/null
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -0,0 +1,146 @@
+{
+    "1": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 8,
+        "num_stages": 5
+    },
+    "2": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "4": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "8": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "16": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "24": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "32": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "48": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "64": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "96": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "128": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "256": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "512": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1024": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1536": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "2048": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "3072": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "4096": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 3
+    }
+}
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2048,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2048,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 76%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2048,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2048,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index c4e52ecc6b85d..5acea242cc0ad 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2048,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2048,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,11 +1,11 @@
 {
     "1": {
-        "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
-        "num_warps": 4,
-        "num_stages": 2
+        "GROUP_SIZE_M": 1,
+        "num_warps": 8,
+        "num_stages": 4
     },
     "2": {
         "BLOCK_SIZE_M": 64,
@@ -13,7 +13,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "4": {
         "BLOCK_SIZE_M": 64,
@@ -21,31 +21,31 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 5
     },
     "8": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
@@ -53,71 +53,71 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
@@ -125,7 +125,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,7 +133,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
@@ -141,6 +141,6 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2304,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2304,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 75%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2304,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2304,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json
index 2a1af0a7119ba..58cdd93e90b8c 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2304,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2304,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,131 +1,131 @@
 {
     "1": {
         "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
@@ -133,7 +133,7 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
@@ -141,6 +141,6 @@
         "BLOCK_SIZE_K": 128,
         "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2304,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2304,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
similarity index 77%
rename from vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2304,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
rename to vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2304,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
index de3a5858ee469..b72e0371d1421 100644
--- a/vllm/model_executor/layers/quantization/utils/configs/N=7168,K=2304,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=2304,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -1,146 +1,146 @@
 {
     "1": {
         "BLOCK_SIZE_M": 16,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "4": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "8": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "16": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 16,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "24": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "32": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 32,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "48": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "64": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 32,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "96": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "128": {
         "BLOCK_SIZE_M": 64,
-        "BLOCK_SIZE_N": 32,
+        "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 4
     },
     "256": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "512": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1024": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 16,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "1536": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 32,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "2048": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "3072": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 64,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 1,
+        "GROUP_SIZE_M": 64,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     },
     "4096": {
         "BLOCK_SIZE_M": 64,
         "BLOCK_SIZE_N": 128,
         "BLOCK_SIZE_K": 128,
-        "GROUP_SIZE_M": 64,
+        "GROUP_SIZE_M": 1,
         "num_warps": 4,
-        "num_stages": 2
+        "num_stages": 3
     }
 }
diff --git a/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=256,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=256,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json
new file mode 100644
index 0000000000000..293adce387e06
--- /dev/null
+++ b/vllm/model_executor/layers/quantization/utils/configs/configs/N=7168,K=256,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json	
@@ -0,0 +1,146 @@
+{
+    "1": {
+        "BLOCK_SIZE_M": 16,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "2": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "4": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "8": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "16": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "24": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "32": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "48": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "64": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "96": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "128": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 32,
+        "num_warps": 4,
+        "num_stages": 5
+    },
+    "256": {
+        "BLOCK_SIZE_M": 128,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 64,
+        "num_warps": 4,
+        "num_stages": 4
+    },
+    "512": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 128,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 16,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1024": {
+        "BLOCK_SIZE_M": 64,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 3
+    },
+    "1536": {
+        "BLOCK_SIZE_M": 128,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 2
+    },
+    "2048": {
+        "BLOCK_SIZE_M": 128,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 2
+    },
+    "3072": {
+        "BLOCK_SIZE_M": 128,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 2
+    },
+    "4096": {
+        "BLOCK_SIZE_M": 128,
+        "BLOCK_SIZE_N": 64,
+        "BLOCK_SIZE_K": 128,
+        "GROUP_SIZE_M": 1,
+        "num_warps": 4,
+        "num_stages": 2
+    }
+}