ef6abf3dc7
Adds ALL files needed for Dockerfile build:
- qwen3_6_scripts/ (baseline patches + our optimizations)
- vllm/ (full vllm package)
- paged_attention_v2_pytorch.py (V2 with single-bmm optimization)
- Dockerfile + computility-run.yaml
Our optimizations vs baseline:
1. paged_attn.py: pre-gathered context KV (eliminates 194 gather calls),
Triton try/fallback, V2 heuristic, threshold 32K→64K
2. paged_attention_v2_pytorch.py: fills NotImplementedError,
single-bmm Phase 1 (195 launches → 3)
3. patch_enable_triton.py: HAS_TRITON=True with safety fallback
4. patch_triton_tuning.py: BLOCK=64, NUM_WARPS=4 for BI-V100
5. computility-run.yaml: gpu-memory-utilization 0.9→0.95,
max-num-batched-tokens 8192→16384
This repo can now be submitted to dev.modelhub.org.cn as-is.
37 lines
1.1 KiB
Python
37 lines
1.1 KiB
Python
import os
|
|
from functools import lru_cache
|
|
|
|
import torch
|
|
|
|
from vllm.logger import init_logger
|
|
|
|
from .interface import DeviceCapability, Platform, PlatformEnum
|
|
|
|
logger = init_logger(__name__)
|
|
|
|
if os.environ.get("VLLM_WORKER_MULTIPROC_METHOD", None) in ["fork", None]:
|
|
logger.warning("`fork` method is not supported by ROCm. "
|
|
"VLLM_WORKER_MULTIPROC_METHOD is overridden to"
|
|
" `spawn` instead.")
|
|
os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"
|
|
|
|
|
|
class RocmPlatform(Platform):
|
|
_enum = PlatformEnum.ROCM
|
|
|
|
@classmethod
|
|
@lru_cache(maxsize=8)
|
|
def get_device_capability(cls, device_id: int = 0) -> DeviceCapability:
|
|
major, minor = torch.cuda.get_device_capability(device_id)
|
|
return DeviceCapability(major=major, minor=minor)
|
|
|
|
@classmethod
|
|
@lru_cache(maxsize=8)
|
|
def get_device_name(cls, device_id: int = 0) -> str:
|
|
return torch.cuda.get_device_name(device_id)
|
|
|
|
@classmethod
|
|
def get_device_total_memory(cls, device_id: int = 0) -> int:
|
|
device_props = torch.cuda.get_device_properties(device_id)
|
|
return device_props.total_memory
|