biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Enable scaled FP8 (e4m3fn) KV cache on ROCm (AMD GPU) (#3290)

Browse Source 
Co-authored-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
Co-authored-by: HaiShaw <hixiao@gmail.com>
Co-authored-by: AdrianAbeyta <Adrian.Abeyta@amd.com>
Co-authored-by: Matthew Wong <Matthew.Wong2@amd.com>
Co-authored-by: root <root@gt-pla-u18-08.pla.dcgpu>
Co-authored-by: mawong-amd <156021403+mawong-amd@users.noreply.github.com>
Co-authored-by: ttbachyinsda <ttbachyinsda@outlook.com>
Co-authored-by: guofangze <guofangze@kuaishou.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: jacobthebanana <50071502+jacobthebanana@users.noreply.github.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Adrian Abeyta
2024-04-03 16:15:55 -05:00
committed by GitHub
parent 3dcb3e8b98
commit 2ff767b513
41 changed files with 2592 additions and 142 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
benchmarks/benchmark_latency.py
View File

@@ -24,6 +24,7 @@ def main(args: argparse.Namespace):
dtype=args.dtype,
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,
@@ -127,10 +128,23 @@ if __name__ == '__main__':
parser.add_argument(
"--kv-cache-dtype",
type=str,
choices=['auto', 'fp8_e5m2'],
choices=['auto', 'fp8'],
default='auto',
help=
'Data type for kv cache storage. If "auto", will use model data type.')
'Data type for kv cache storage. If "auto", will use model data type. '
'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(
'--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',