[Core] Deprecating block manager v1 and make block manager v2 default (#8704)
Removing the block manager v1. This is the initial piece of prefix-caching-centric design. In order to achieve prefix-caching-centric design, we need to simplify the code path so that we only use v2 block manager (which has much higher performance on prefix caching).
This commit is contained in:
@@ -38,7 +38,6 @@ def main(args: argparse.Namespace):
|
||||
quantization_param_path=args.quantization_param_path,
|
||||
device=args.device,
|
||||
ray_workers_use_nsight=args.ray_workers_use_nsight,
|
||||
use_v2_block_manager=args.use_v2_block_manager,
|
||||
enable_chunked_prefill=args.enable_chunked_prefill,
|
||||
download_dir=args.download_dir,
|
||||
block_size=args.block_size,
|
||||
@@ -221,9 +220,6 @@ if __name__ == '__main__':
|
||||
parser.add_argument("--enable-prefix-caching",
|
||||
action='store_true',
|
||||
help="Enable automatic prefix caching")
|
||||
parser.add_argument('--use-v2-block-manager',
|
||||
action='store_true',
|
||||
default=EngineArgs.use_v2_block_manager)
|
||||
parser.add_argument(
|
||||
"--ray-workers-use-nsight",
|
||||
action='store_true',
|
||||
|
||||
@@ -33,7 +33,6 @@ from typing import List, Optional, Tuple
|
||||
from transformers import PreTrainedTokenizerBase
|
||||
|
||||
from vllm import LLM, SamplingParams
|
||||
from vllm.engine.arg_utils import EngineArgs
|
||||
from vllm.utils import FlexibleArgumentParser
|
||||
|
||||
try:
|
||||
@@ -134,7 +133,6 @@ def main(args):
|
||||
tokenizer_mode='auto',
|
||||
trust_remote_code=True,
|
||||
enforce_eager=True,
|
||||
use_v2_block_manager=args.use_v2_block_manager,
|
||||
tensor_parallel_size=args.tensor_parallel_size,
|
||||
enable_prefix_caching=args.enable_prefix_caching)
|
||||
|
||||
@@ -176,10 +174,6 @@ if __name__ == "__main__":
|
||||
parser.add_argument('--enable-prefix-caching',
|
||||
action='store_true',
|
||||
help='enable prefix caching')
|
||||
parser.add_argument('--use-v2-block-manager',
|
||||
action='store_true',
|
||||
default=EngineArgs.use_v2_block_manager,
|
||||
help='Use BlockSpaceMangerV2')
|
||||
parser.add_argument('--num-prompts',
|
||||
type=int,
|
||||
default=1,
|
||||
|
||||
@@ -86,7 +86,6 @@ def run_vllm(
|
||||
distributed_executor_backend: Optional[str],
|
||||
gpu_memory_utilization: float = 0.9,
|
||||
num_scheduler_steps: int = 1,
|
||||
use_v2_block_manager: bool = False,
|
||||
download_dir: Optional[str] = None,
|
||||
load_format: str = EngineArgs.load_format,
|
||||
disable_async_output_proc: bool = False,
|
||||
@@ -113,7 +112,6 @@ def run_vllm(
|
||||
distributed_executor_backend=distributed_executor_backend,
|
||||
load_format=load_format,
|
||||
num_scheduler_steps=num_scheduler_steps,
|
||||
use_v2_block_manager=use_v2_block_manager,
|
||||
disable_async_output_proc=disable_async_output_proc,
|
||||
)
|
||||
|
||||
@@ -176,7 +174,6 @@ async def run_vllm_async(
|
||||
distributed_executor_backend: Optional[str],
|
||||
gpu_memory_utilization: float = 0.9,
|
||||
num_scheduler_steps: int = 1,
|
||||
use_v2_block_manager: bool = False,
|
||||
download_dir: Optional[str] = None,
|
||||
load_format: str = EngineArgs.load_format,
|
||||
disable_async_output_proc: bool = False,
|
||||
@@ -204,7 +201,6 @@ async def run_vllm_async(
|
||||
distributed_executor_backend=distributed_executor_backend,
|
||||
load_format=load_format,
|
||||
num_scheduler_steps=num_scheduler_steps,
|
||||
use_v2_block_manager=use_v2_block_manager,
|
||||
disable_async_output_proc=disable_async_output_proc,
|
||||
worker_use_ray=False,
|
||||
disable_log_requests=True,
|
||||
@@ -341,8 +337,7 @@ def main(args: argparse.Namespace):
|
||||
args.enable_prefix_caching, args.enable_chunked_prefill,
|
||||
args.max_num_batched_tokens, args.distributed_executor_backend,
|
||||
args.gpu_memory_utilization, args.num_scheduler_steps,
|
||||
args.use_v2_block_manager, args.download_dir, args.load_format,
|
||||
args.disable_async_output_proc
|
||||
args.download_dir, args.load_format, args.disable_async_output_proc
|
||||
]
|
||||
|
||||
if args.async_engine:
|
||||
@@ -471,10 +466,6 @@ if __name__ == "__main__":
|
||||
type=int,
|
||||
default=1,
|
||||
help="Maximum number of forward steps per scheduler call.")
|
||||
parser.add_argument("--use-v2-block-manager",
|
||||
action='store_true',
|
||||
default=EngineArgs.use_v2_block_manager,
|
||||
help="Enable block manager v2.")
|
||||
parser.add_argument(
|
||||
"--enable-prefix-caching",
|
||||
action='store_true',
|
||||
|
||||
@@ -16,7 +16,6 @@ def main(args):
|
||||
enforce_eager=True,
|
||||
enable_prefix_caching=True,
|
||||
tensor_parallel_size=args.tensor_parallel_size,
|
||||
use_v2_block_manager=args.use_v2_block_manager,
|
||||
)
|
||||
|
||||
sampling_params = SamplingParams(temperature=0, max_tokens=args.output_len)
|
||||
@@ -56,8 +55,5 @@ if __name__ == "__main__":
|
||||
parser.add_argument('--enable-prefix-caching',
|
||||
action='store_true',
|
||||
help='enable prefix caching')
|
||||
parser.add_argument('--use-v2-block-manager',
|
||||
action='store_true',
|
||||
help='Use BlockSpaceMangerV2')
|
||||
args = parser.parse_args()
|
||||
main(args)
|
||||
|
||||
Reference in New Issue
Block a user