Files
aituner/runs/frontier-qwen30-vllm020-profile-v1/profile_vllm020_flashattn.py

287 lines
11 KiB
Python

#!/usr/bin/env python3
"""Profile the exact vLLM 0.20 FlashAttention backend at TP-local shapes.
This deliberately uses vLLM's own v0.20.0 attention benchmark runner instead
of Frontier's FlashInfer-only attention wrapper. The output is raw evidence;
projection into Frontier's split attention CSV schema is a separate step.
"""
from __future__ import annotations
import argparse
import json
import statistics
import subprocess
import sys
import types
from pathlib import Path
import torch
import vllm
VLLM_VERSION = "0.20.0"
VLLM_COMMIT = "88d34c6409e9fb3c7b8ca0c04756f061d2099eb1"
def parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser()
parser.add_argument("--vllm-source", type=Path, required=True)
parser.add_argument("--model", type=Path, required=True)
parser.add_argument("--output", type=Path, required=True)
parser.add_argument("--tp", type=int, choices=(1, 2, 4), nargs="+", default=[1, 2, 4])
parser.add_argument(
"--batch-specs",
nargs="+",
default=["q128", "4q1s128", "q128_4q1s128"],
)
parser.add_argument("--warmup-iters", type=int, default=3)
parser.add_argument("--repeats", type=int, default=5)
parser.add_argument("--device", default="cuda:0")
parser.add_argument("--profile-kv-update", action="store_true")
return parser.parse_args()
def git_head(repo: Path) -> str:
return subprocess.check_output(
["git", "-C", str(repo), "rev-parse", "HEAD"], text=True
).strip()
def json_default(value: object) -> object:
if isinstance(value, (Path, torch.dtype, torch.device)):
return str(value)
item = getattr(value, "item", None)
if callable(item):
return item()
raise TypeError(f"cannot serialize {type(value).__name__}")
def main() -> None:
args = parse_args()
if vllm.__version__ != VLLM_VERSION:
raise SystemExit(f"expected vLLM {VLLM_VERSION}, got {vllm.__version__}")
source_head = git_head(args.vllm_source)
if source_head != VLLM_COMMIT:
raise SystemExit(f"expected vLLM source {VLLM_COMMIT}, got {source_head}")
if not args.model.joinpath("config.json").is_file():
raise SystemExit(f"missing model config: {args.model / 'config.json'}")
bench_dir = args.vllm_source / "benchmarks" / "attention_benchmarks"
sys.path.insert(0, str(bench_dir))
import runner # type: ignore[import-not-found] # noqa: PLC0415
from batch_spec import parse_batch_spec # type: ignore[import-not-found] # noqa: PLC0415
from common import BenchmarkConfig # type: ignore[import-not-found] # noqa: PLC0415
from vllm.config import ( # noqa: PLC0415
CacheConfig,
CompilationConfig,
DeviceConfig,
LoadConfig,
ModelConfig,
ParallelConfig,
SchedulerConfig,
VllmConfig,
set_current_vllm_config,
)
from vllm.v1.attention.backends.utils import ( # noqa: PLC0415
get_kv_cache_layout,
set_kv_cache_layout,
)
from vllm.v1.kv_cache_interface import FullAttentionSpec # noqa: PLC0415
from vllm.v1.worker.workspace import init_workspace_manager # noqa: PLC0415
def create_vllm_config(config: BenchmarkConfig, max_num_blocks: int) -> VllmConfig:
model_config = ModelConfig(
model=str(args.model),
tokenizer=str(args.model),
trust_remote_code=False,
dtype="bfloat16",
seed=0,
max_model_len=40960,
)
cache_config = CacheConfig(block_size=config.block_size, cache_dtype="auto")
cache_config.num_gpu_blocks = max_num_blocks
cache_config.num_cpu_blocks = 0
parallel_config = ParallelConfig(tensor_parallel_size=1)
scheduler_config = SchedulerConfig(
max_num_seqs=256,
max_num_batched_tokens=8192,
max_model_len=40960,
is_encoder_decoder=False,
enable_chunked_prefill=True,
)
model_config.get_num_layers = types.MethodType(
lambda self: config.num_layers, model_config
)
model_config.get_sliding_window_for_layer = types.MethodType(
lambda self, i: None, model_config
)
model_config.get_logits_soft_cap_for_layer = types.MethodType(
lambda self, i: 0.0, model_config
)
model_config.get_sm_scale_for_layer = types.MethodType(
lambda self, i: 1.0 / config.head_dim**0.5, model_config
)
model_config.get_num_attention_heads = types.MethodType(
lambda self, parallel_config=None: config.num_q_heads, model_config
)
model_config.get_num_kv_heads = types.MethodType(
lambda self, parallel_config=None: config.num_kv_heads, model_config
)
model_config.get_head_size = types.MethodType(
lambda self: config.head_dim, model_config
)
model_config.get_sliding_window = types.MethodType(
lambda self: None, model_config
)
return VllmConfig(
model_config=model_config,
cache_config=cache_config,
parallel_config=parallel_config,
scheduler_config=scheduler_config,
device_config=DeviceConfig(),
load_config=LoadConfig(),
compilation_config=CompilationConfig(),
)
runner._create_vllm_config = create_vllm_config
init_workspace_manager(args.device)
def profile_kv_cache_update(config: BenchmarkConfig) -> dict[str, float]:
device = torch.device(config.device)
requests = parse_batch_spec(config.batch_spec)
total_q = sum(request.q_len for request in requests)
max_kv = max(request.kv_len for request in requests)
max_blocks_per_request = (max_kv + config.block_size - 1) // config.block_size
max_num_blocks = len(requests) * max_blocks_per_request
vllm_config = create_vllm_config(config, max_num_blocks)
dtype = vllm_config.model_config.dtype
with set_current_vllm_config(vllm_config):
backend_config = runner._get_backend_config(config.backend)
backend_class, impl, layer = runner._create_backend_impl(
backend_config, config, device, dtype
)
required_layout = backend_class.get_required_kv_cache_layout()
if required_layout is not None:
set_kv_cache_layout(required_layout)
get_kv_cache_layout.cache_clear()
common_metadata = runner._build_common_attn_metadata(
[request.q_len for request in requests],
[request.kv_len for request in requests],
config.block_size,
device,
)
kv_cache_spec = FullAttentionSpec(
block_size=config.block_size,
num_kv_heads=config.num_kv_heads,
head_size=config.head_dim,
dtype=dtype,
)
layer._kv_cache_spec = kv_cache_spec
_, key_list, value_list = runner._create_input_tensors(
config, total_q, device, dtype
)
cache_list = runner._create_kv_cache(
config, max_num_blocks, backend_class, device, dtype
)
for _ in range(config.warmup_iters):
for layer_index in range(config.num_layers):
impl.do_kv_cache_update(
layer,
key_list[layer_index],
value_list[layer_index],
cache_list[layer_index],
common_metadata.slot_mapping,
)
torch.accelerator.synchronize()
samples: list[float] = []
for _ in range(config.repeats):
start = torch.cuda.Event(enable_timing=True)
end = torch.cuda.Event(enable_timing=True)
start.record()
for layer_index in range(config.num_layers):
impl.do_kv_cache_update(
layer,
key_list[layer_index],
value_list[layer_index],
cache_list[layer_index],
common_metadata.slot_mapping,
)
end.record()
torch.accelerator.synchronize()
samples.append(float(start.elapsed_time(end)) / config.num_layers)
return {
"min_ms": min(samples),
"max_ms": max(samples),
"mean_ms": statistics.fmean(samples),
"median_ms": statistics.median(samples),
"std_ms": statistics.pstdev(samples),
}
rows: list[dict[str, object]] = []
for tp in args.tp:
for batch_spec in args.batch_specs:
config = BenchmarkConfig(
backend="FLASH_ATTN",
batch_spec=batch_spec,
num_layers=1,
head_dim=128,
num_q_heads=32 // tp,
num_kv_heads=4 // tp,
block_size=16,
device=args.device,
dtype=torch.bfloat16,
repeats=args.repeats,
warmup_iters=args.warmup_iters,
profile_memory=True,
kv_cache_dtype="auto",
use_cuda_graphs=False,
)
result = runner.run_attention_benchmark(config)
row = result.to_dict()
row["tensor_parallel_size"] = tp
row["attention_core_excludes_kv_cache_update"] = True
if args.profile_kv_update:
row["kv_cache_update_time"] = profile_kv_cache_update(config)
rows.append(row)
print(
json.dumps(
{
"tp": tp,
"batch_spec": batch_spec,
"mean_time_s": result.mean_time,
"error": result.error,
},
sort_keys=True,
default=json_default,
),
flush=True,
)
if not result.success:
raise SystemExit(f"attention profile failed: {row}")
payload = {
"schema_version": "qwen30_vllm020_flashattn_raw.v1",
"environment": {
"vllm_version": vllm.__version__,
"vllm_source_commit": source_head,
"torch_version": torch.__version__,
"torch_cuda": torch.version.cuda,
"gpu": torch.cuda.get_device_name(torch.device(args.device)),
"model": str(args.model),
"dtype": "bfloat16",
"attention_backend": "FLASH_ATTN",
"block_size": 16,
"profile_kv_update": args.profile_kv_update,
},
"rows": rows,
}
args.output.parent.mkdir(parents=True, exist_ok=True)
args.output.write_text(
json.dumps(payload, indent=2, sort_keys=True, default=json_default) + "\n"
)
if __name__ == "__main__":
main()