[ Kernel ] Enable Dynamic Per Token fp8 (vllm-project#6547)

.buildkite/lm-eval-harness/configs/Meta-Llama-3-8B-Instruct-Channelwise-compressed-tensors.yaml

@@ -0,0 +1,11 @@
+# bash .buildkite/lm-eval-harness/run-lm-eval-gsm-vllm-baseline.sh -m nm-testing/Meta-Llama-3-8B-Instruct-W8A8-FP8-Channelwise-compressed-tensors -b auto -l 1000 -f 5 -t 1
+model_name: "nm-testing/Meta-Llama-3-8B-Instruct-W8A8-FP8-Channelwise-compressed-tensors"
+tasks:
+- name: "gsm8k"
+  metrics:
+  - name: "exact_match,strict-match"
+    value: 0.769
+  - name: "exact_match,flexible-extract"
+    value: 0.769
+limit: 1000
+num_fewshot: 5

.buildkite/lm-eval-harness/configs/models-small.txt

@@ -3,4 +3,5 @@ Meta-Llama-3-8B-Instruct-FP8.yaml
 Meta-Llama-3-8B-Instruct-FP8-compressed-tensors.yaml
 Meta-Llama-3-8B-Instruct-INT8-compressed-tensors.yaml
 Meta-Llama-3-8B-Instruct-nonuniform-compressed-tensors.yaml
+Meta-Llama-3-8B-Instruct-Channelwise-compressed-tensors.yaml
 Qwen2-1.5B-Instruct-INT8-compressed-tensors.yaml

tests/kernels/test_fp8_quant.py

@@ -27,7 +27,8 @@ def test_dynamic_per_token_fp8_quant(num_tokens: int, hidden_size: int,
                    device="cuda") + 1e-6  # avoid nans
 
     ref_out, ref_scales = ref_dynamic_per_token_quant(x, torch.float8_e4m3fn)
-    ops_out, ops_scales = ops.dynamic_per_token_scaled_fp8_quant(x)
+    ops_out, ops_scales = ops.scaled_fp8_quant(x,
+                                               use_per_token_if_dynamic=True)
 
     assert torch.allclose(ref_scales, ops_scales)
     assert torch.allclose(ref_out.to(dtype=torch.float32),

vllm/_custom_ops.py

@@ -300,6 +300,7 @@ def scaled_fp8_quant(
     input: torch.Tensor,
     scale: Optional[torch.Tensor] = None,
     batch_dim_padding: Optional[int] = None,
+    use_per_token_if_dynamic: bool = False,
 ) -> Tuple[torch.Tensor, torch.Tensor]:
     """
     Quantize input tensor to FP8 and return quantized tensor and scale.
@@ -315,6 +316,8 @@ def scaled_fp8_quant(
         scale: Optional scaling factor for the FP8 quantization
         batch_dim_padding: If specified, pad the first dimension
             of the output to at least this value.
+        use_per_token_if_dynamic: Whether to do per_tensor or per_token 
+            in the dynamic quantization case.
 
     Returns:
         Tuple[torch.Tensor, torch.Tensor]: The output tensor in FP8 and
@@ -328,22 +331,19 @@ def scaled_fp8_quant(
     else:
         output = torch.empty_like(input, dtype=torch.float8_e4m3fn)
     if scale is None:
-        scale = torch.zeros(1, device=input.device, dtype=torch.float32)
-        torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
+        if use_per_token_if_dynamic:
+            scale = torch.empty((input.numel() // input.shape[-1], 1),
+                                device=input.device,
+                                dtype=torch.float32)
+            torch.ops._C.dynamic_per_token_scaled_fp8_quant(
+                output, input, scale)
+        else:
+            scale = torch.zeros(1, device=input.device, dtype=torch.float32)
+            torch.ops._C.dynamic_scaled_fp8_quant(output, input, scale)
     else:
         torch.ops._C.static_scaled_fp8_quant(output, input, scale)
-    return output, scale
-
 
-def dynamic_per_token_scaled_fp8_quant(
-        input: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]:
-
-    output = torch.empty_like(input, dtype=torch.float8_e4m3fn)
-    scales = torch.empty((input.numel() // input.shape[-1], 1),
-                         device=input.device,
-                         dtype=torch.float32)
-    torch.ops._C.dynamic_per_token_scaled_fp8_quant(output, input, scales)
-    return output, scales
+    return output, scale
 
 
 # int8

vllm/model_executor/layers/quantization/compressed_tensors/schemes/compressed_tensors_w8a8_fp8.py

@@ -103,4 +103,5 @@ def apply_weights(self,
             weight_scale=layer.weight_scale,
             input_scale=layer.input_scale,
             bias=bias,
-            cutlass_fp8_supported=self.cutlass_fp8_supported)
+            cutlass_fp8_supported=self.cutlass_fp8_supported,
+            use_per_token_if_dynamic=True)

vllm/model_executor/layers/quantization/fp8.py

@@ -214,7 +214,8 @@ def apply(self,
             weight_scale=layer.weight_scale,
             input_scale=layer.input_scale,
             bias=bias,
-            cutlass_fp8_supported=self.cutlass_fp8_supported)
+            cutlass_fp8_supported=self.cutlass_fp8_supported,
+            use_per_token_if_dynamic=False)
 
 
 class Fp8MoEMethod(FusedMoEMethodBase):

vllm/model_executor/layers/quantization/utils/w8a8_utils.py

@@ -107,41 +107,55 @@ def apply_fp8_linear(
     input_scale: torch.Tensor,
     bias: Optional[torch.Tensor] = None,
     cutlass_fp8_supported: bool = True,
+    use_per_token_if_dynamic: bool = False,
 ) -> torch.Tensor:
     # ops.scaled_fp8_quant supports both dynamic and static quant.
     #   If dynamic, layer.input_scale is None and x_scale computed from x.
     #   If static, layer.input_scale is scalar and x_scale is input_scale.
 
+    # cutlass_scaled_mm supports per tensor/channel W and per tensor/token A
     if cutlass_fp8_supported:
-        qinput, x_scale = ops.scaled_fp8_quant(input, input_scale)
+        qinput, x_scale = ops.scaled_fp8_quant(
+            input,
+            input_scale,
+            use_per_token_if_dynamic=use_per_token_if_dynamic)
 
         # Fused GEMM_DQ
-        output = ops.cutlass_scaled_mm(qinput,
-                                       weight,
-                                       out_dtype=input.dtype,
-                                       scale_a=x_scale,
-                                       scale_b=weight_scale,
-                                       bias=bias)
-
+        return ops.cutlass_scaled_mm(qinput,
+                                     weight,
+                                     out_dtype=input.dtype,
+                                     scale_a=x_scale,
+                                     scale_b=weight_scale,
+                                     bias=bias)
+
+    # torch.scaled_mm supports per tensor weights + activations only
+    # so fallback to naive if per channel or per token
     else:
         # Note: we pad the input because torch._scaled_mm is more performant
         # for matrices with batch dimension > 16.
         # This could change in the future.
-        qinput, x_scale = ops.scaled_fp8_quant(input,
-                                               input_scale,
-                                               batch_dim_padding=17)
+        qinput, x_scale = ops.scaled_fp8_quant(
+            input,
+            input_scale,
+            batch_dim_padding=17,
+            use_per_token_if_dynamic=use_per_token_if_dynamic)
+
+        per_tensor_weights = (weight_scale.numel() == 1)
+        per_tensor_activations = (x_scale.numel() == 1)
 
-        if weight_scale.numel() == 1:
+        if per_tensor_weights and per_tensor_activations:
             # Fused GEMM_DQ
             output, _ = torch._scaled_mm(qinput,
                                          weight,
                                          out_dtype=input.dtype,
                                          scale_a=x_scale,
                                          scale_b=weight_scale,
                                          bias=bias)
+            return torch.narrow(output, 0, 0, input.shape[0])
+
         else:
-            # Fallback for channelwise case, where the weight scales are
-            # applied separately.
+            # Fallback for channelwise case, where we use unfused DQ
+            # due to limitations with scaled_mm
 
             # Symmetric quantized GEMM by definition computes the following:
             #   C = (s_x * X) (s_w * W) + bias
@@ -155,21 +169,21 @@ def apply_fp8_linear(
             # For the scaled_mm fallback case, we break this down, since it
             # does not support s_w being a vector.
 
-            # This computes C = sx * (X * W).
+            # GEMM
+            # This computes C = (X * W).
             # Output in fp32 to allow subsequent ops to happen in-place
             output, _ = torch._scaled_mm(qinput,
                                          weight,
-                                         out_dtype=torch.float32,
-                                         scale_a=x_scale)
+                                         out_dtype=torch.float32)
+            # Unpad (undo batch_dim_padding)
+            output = torch.narrow(output, 0, 0, input.shape[0])
 
-            # C = sw * sx * (X * W)
-            output = output * weight_scale.t()
+            # DQ
+            # C = sw * sx * (X * W) + bias
+            output = output * x_scale * weight_scale.t()
             if bias is not None:
-                # C = sw * sx * (X * W) + bias
                 output = output + bias
-            output = output.to(dtype=input.dtype)
-
-    return torch.narrow(output, 0, 0, input.shape[0])
+            return output.to(dtype=input.dtype)
 
 
 def apply_int8_linear(