Merge remote-tracking branch 'upstream/main' into luka/rms-norm-fusion

CMakeLists.txt

@@ -169,12 +169,12 @@ endif()
 
 #
 # Use FetchContent for C++ dependencies that are compiled as part of vLLM's build process.
-# Configure it to place files in vllm/.deps, in order to play nicely with sccache.
+# setup.py will override FETCHCONTENT_BASE_DIR to play nicely with sccache.
+# Each dependency that produces build artifacts should override its BINARY_DIR to avoid
+# conflicts between build types. It should instead be set to ${CMAKE_BINARY_DIR}/<dependency>.
 #
 include(FetchContent)
-get_filename_component(PROJECT_ROOT_DIR "${CMAKE_CURRENT_SOURCE_DIR}" ABSOLUTE)
-file(MAKE_DIRECTORY "${FETCHCONTENT_BASE_DIR}")
-set(FETCHCONTENT_BASE_DIR "${PROJECT_ROOT_DIR}/.deps")
+file(MAKE_DIRECTORY ${FETCHCONTENT_BASE_DIR}) # Ensure the directory exists
 message(STATUS "FetchContent base directory: ${FETCHCONTENT_BASE_DIR}")
 
 #
@@ -253,7 +253,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
     message(STATUS "Building Marlin kernels for archs: ${MARLIN_ARCHS}")
   else()
     message(STATUS "Not building Marlin kernels as no compatible archs found"
-                   "in CUDA target architectures")
+                   " in CUDA target architectures")
   endif()
 
   #
@@ -433,7 +433,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
     message(STATUS "Building Marlin MOE kernels for archs: ${MARLIN_MOE_ARCHS}")
   else()
     message(STATUS "Not building Marlin MOE kernels as no compatible archs found"
-                   "in CUDA target architectures")
+                   " in CUDA target architectures")
   endif()
 endif()
 
@@ -510,6 +510,8 @@ else()
           GIT_REPOSITORY https://github.com/vllm-project/flash-attention.git
           GIT_TAG 013f0c4fc47e6574060879d9734c1df8c5c273bd
           GIT_PROGRESS TRUE
+          # Don't share the vllm-flash-attn build between build types
+          BINARY_DIR ${CMAKE_BINARY_DIR}/vllm-flash-attn
   )
 endif()
 

benchmarks/benchmark_latency.py

@@ -1,5 +1,6 @@
 """Benchmark the latency of processing a single batch of requests."""
 import argparse
+import dataclasses
 import json
 import time
 from pathlib import Path
@@ -10,43 +11,19 @@
 from tqdm import tqdm
 
 from vllm import LLM, SamplingParams
-from vllm.engine.arg_utils import DEVICE_OPTIONS, EngineArgs
+from vllm.engine.arg_utils import EngineArgs
 from vllm.inputs import PromptType
-from vllm.model_executor.layers.quantization import QUANTIZATION_METHODS
 from vllm.utils import FlexibleArgumentParser
 
 
 def main(args: argparse.Namespace):
     print(args)
 
+    engine_args = EngineArgs.from_cli_args(args)
+
     # NOTE(woosuk): If the request cannot be processed in a single batch,
     # the engine will automatically process the request in multiple batches.
-    llm = LLM(
-        model=args.model,
-        speculative_model=args.speculative_model,
-        num_speculative_tokens=args.num_speculative_tokens,
-        speculative_draft_tensor_parallel_size=\
-            args.speculative_draft_tensor_parallel_size,
-        tokenizer=args.tokenizer,
-        quantization=args.quantization,
-        tensor_parallel_size=args.tensor_parallel_size,
-        trust_remote_code=args.trust_remote_code,
-        dtype=args.dtype,
-        max_model_len=args.max_model_len,
-        enforce_eager=args.enforce_eager,
-        kv_cache_dtype=args.kv_cache_dtype,
-        quantization_param_path=args.quantization_param_path,
-        device=args.device,
-        ray_workers_use_nsight=args.ray_workers_use_nsight,
-        enable_chunked_prefill=args.enable_chunked_prefill,
-        download_dir=args.download_dir,
-        block_size=args.block_size,
-        gpu_memory_utilization=args.gpu_memory_utilization,
-        load_format=args.load_format,
-        distributed_executor_backend=args.distributed_executor_backend,
-        otlp_traces_endpoint=args.otlp_traces_endpoint,
-        enable_prefix_caching=args.enable_prefix_caching,
-    )
+    llm = LLM(**dataclasses.asdict(engine_args))
 
     sampling_params = SamplingParams(
         n=args.n,
@@ -125,19 +102,6 @@ def run_to_completion(profile_dir: Optional[str] = None):
     parser = FlexibleArgumentParser(
         description='Benchmark the latency of processing a single batch of '
         'requests till completion.')
-    parser.add_argument('--model', type=str, default='facebook/opt-125m')
-    parser.add_argument('--speculative-model', type=str, default=None)
-    parser.add_argument('--num-speculative-tokens', type=int, default=None)
-    parser.add_argument('--speculative-draft-tensor-parallel-size',
-                        '-spec-draft-tp',
-                        type=int,
-                        default=None)
-    parser.add_argument('--tokenizer', type=str, default=None)
-    parser.add_argument('--quantization',
-                        '-q',
-                        choices=[*QUANTIZATION_METHODS, None],
-                        default=None)
-    parser.add_argument('--tensor-parallel-size', '-tp', type=int, default=1)
     parser.add_argument('--input-len', type=int, default=32)
     parser.add_argument('--output-len', type=int, default=128)
     parser.add_argument('--batch-size', type=int, default=8)
@@ -154,45 +118,6 @@ def run_to_completion(profile_dir: Optional[str] = None):
                         type=int,
                         default=30,
                         help='Number of iterations to run.')
-    parser.add_argument('--trust-remote-code',
-                        action='store_true',
-                        help='trust remote code from huggingface')
-    parser.add_argument(
-        '--max-model-len',
-        type=int,
-        default=None,
-        help='Maximum length of a sequence (including prompt and output). '
-        'If None, will be derived from the model.')
-    parser.add_argument(
-        '--dtype',
-        type=str,
-        default='auto',
-        choices=['auto', 'half', 'float16', 'bfloat16', 'float', 'float32'],
-        help='data type for model weights and activations. '
-        'The "auto" option will use FP16 precision '
-        'for FP32 and FP16 models, and BF16 precision '
-        'for BF16 models.')
-    parser.add_argument('--enforce-eager',
-                        action='store_true',
-                        help='enforce eager mode and disable CUDA graph')
-    parser.add_argument(
-        '--kv-cache-dtype',
-        type=str,
-        choices=['auto', 'fp8', 'fp8_e5m2', 'fp8_e4m3'],
-        default="auto",
-        help='Data type for kv cache storage. If "auto", will use model '
-        'data type. CUDA 11.8+ supports fp8 (=fp8_e4m3) and fp8_e5m2. '
-        'ROCm (AMD GPU) supports fp8 (=fp8_e4m3)')
-    parser.add_argument(
-        '--quantization-param-path',
-        type=str,
-        default=None,
-        help='Path to the JSON file containing the KV cache scaling factors. '
-        'This should generally be supplied, when KV cache dtype is FP8. '
-        'Otherwise, KV cache scaling factors default to 1.0, which may cause '
-        'accuracy issues. FP8_E5M2 (without scaling) is only supported on '
-        'cuda version greater than 11.8. On ROCm (AMD GPU), FP8_E4M3 is '
-        'instead supported for common inference criteria.')
     parser.add_argument(
         '--profile',
         action='store_true',
@@ -203,78 +128,12 @@ def run_to_completion(profile_dir: Optional[str] = None):
         default=None,
         help=('path to save the pytorch profiler output. Can be visualized '
               'with ui.perfetto.dev or Tensorboard.'))
-    parser.add_argument("--device",
-                        type=str,
-                        default="auto",
-                        choices=DEVICE_OPTIONS,
-                        help='device type for vLLM execution')
-    parser.add_argument('--block-size',
-                        type=int,
-                        default=16,
-                        help='block size of key/value cache')
-    parser.add_argument(
-        '--enable-chunked-prefill',
-        action='store_true',
-        help='If True, the prefill requests can be chunked based on the '
-        'max_num_batched_tokens')
-    parser.add_argument("--enable-prefix-caching",
-                        action='store_true',
-                        help="Enable automatic prefix caching")
-    parser.add_argument(
-        "--ray-workers-use-nsight",
-        action='store_true',
-        help="If specified, use nsight to profile ray workers",
-    )
-    parser.add_argument('--download-dir',
-                        type=str,
-                        default=None,
-                        help='directory to download and load the weights, '
-                        'default to the default cache dir of huggingface')
     parser.add_argument(
         '--output-json',
         type=str,
         default=None,
         help='Path to save the latency results in JSON format.')
-    parser.add_argument('--gpu-memory-utilization',
-                        type=float,
-                        default=0.9,
-                        help='the fraction of GPU memory to be used for '
-                        'the model executor, which can range from 0 to 1.'
-                        'If unspecified, will use the default value of 0.9.')
-    parser.add_argument(
-        '--load-format',
-        type=str,
-        default=EngineArgs.load_format,
-        choices=[
-            'auto', 'pt', 'safetensors', 'npcache', 'dummy', 'tensorizer',
-            'bitsandbytes'
-        ],
-        help='The format of the model weights to load.\n\n'
-        '* "auto" will try to load the weights in the safetensors format '
-        'and fall back to the pytorch bin format if safetensors format '
-        'is not available.\n'
-        '* "pt" will load the weights in the pytorch bin format.\n'
-        '* "safetensors" will load the weights in the safetensors format.\n'
-        '* "npcache" will load the weights in pytorch format and store '
-        'a numpy cache to speed up the loading.\n'
-        '* "dummy" will initialize the weights with random values, '
-        'which is mainly for profiling.\n'
-        '* "tensorizer" will load the weights using tensorizer from '
-        'CoreWeave. See the Tensorize vLLM Model script in the Examples'
-        'section for more information.\n'
-        '* "bitsandbytes" will load the weights using bitsandbytes '
-        'quantization.\n')
-    parser.add_argument(
-        '--distributed-executor-backend',
-        choices=['ray', 'mp'],
-        default=None,
-        help='Backend to use for distributed serving. When more than 1 GPU '
-        'is used, will be automatically set to "ray" if installed '
-        'or "mp" (multiprocessing) otherwise.')
-    parser.add_argument(
-        '--otlp-traces-endpoint',
-        type=str,
-        default=None,
-        help='Target URL to which OpenTelemetry traces will be sent.')
+
+    parser = EngineArgs.add_cli_args(parser)
     args = parser.parse_args()
     main(args)

benchmarks/benchmark_prefix_caching.py

@@ -25,6 +25,7 @@
         --input-length-range 128:256
 """
 
+import dataclasses
 import json
 import random
 import time
@@ -33,6 +34,7 @@
 from transformers import PreTrainedTokenizerBase
 
 from vllm import LLM, SamplingParams
+from vllm.engine.arg_utils import EngineArgs
 from vllm.utils import FlexibleArgumentParser
 
 try:
@@ -129,12 +131,9 @@ def main(args):
         filtered_datasets = [(PROMPT, prompt_len, args.output_len)
                              ] * args.num_prompts
 
-    llm = LLM(model=args.model,
-              tokenizer_mode='auto',
-              trust_remote_code=True,
-              enforce_eager=True,
-              tensor_parallel_size=args.tensor_parallel_size,
-              enable_prefix_caching=args.enable_prefix_caching)
+    engine_args = EngineArgs.from_cli_args(args)
+
+    llm = LLM(**dataclasses.asdict(engine_args))
 
     sampling_params = SamplingParams(temperature=0, max_tokens=args.output_len)
 
@@ -162,18 +161,11 @@ def main(args):
     parser = FlexibleArgumentParser(
         description=
         'Benchmark the performance with or without automatic prefix caching.')
-    parser.add_argument('--model',
-                        type=str,
-                        default='baichuan-inc/Baichuan2-13B-Chat')
     parser.add_argument("--dataset-path",
                         type=str,
                         default=None,
                         help="Path to the dataset.")
-    parser.add_argument('--tensor-parallel-size', '-tp', type=int, default=1)
     parser.add_argument('--output-len', type=int, default=10)
-    parser.add_argument('--enable-prefix-caching',
-                        action='store_true',
-                        help='enable prefix caching')
     parser.add_argument('--num-prompts',
                         type=int,
                         default=1,
@@ -190,9 +182,7 @@ def main(args):
                         default='128:256',
                         help='Range of input lengths for sampling prompts,'
                         'specified as "min:max" (e.g., "128:256").')
-    parser.add_argument("--seed",
-                        type=int,
-                        default=0,
-                        help='Random seed for reproducibility')
+
+    parser = EngineArgs.add_cli_args(parser)
     args = parser.parse_args()
     main(args)