wangbo/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2026-01-10 14:20:49 +08:00
Code Issues Actions 11 Packages Projects Releases Wiki Activity

add amd gfx1102 arch for pytorch attention

Browse Source
This commit is contained in:
doctorpangloss 2025-10-21 14:33:54 -07:00
parent d9269785d3
commit 99be2c40fa
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
comfy/model_management.py
View File

@ -383,7 +383,7 @@ SUPPORT_FP8_OPS = args.supports_fp8_compute
try:
if is_amd():
torch.backends.cudnn.enabled = False # Seems to improve things a lot on AMD
logging.info("Set: torch.backends.cudnn.enabled = False for better AMD performance.")
logger.info("Set: torch.backends.cudnn.enabled = False for better AMD performance.")
try:
rocm_version = tuple(map(int, str(torch.version.hip).split(".")[:2]))
except:
@ -394,7 +394,7 @@ try:
if args.use_split_cross_attention == False and args.use_quad_cross_attention == False:
if importlib.util.find_spec('triton') is not None: # AMD efficient attention implementation depends on triton. TODO: better way of detecting if it's compiled in or not.
if torch_version_numeric >= (2, 7): # works on 2.6 but doesn't actually seem to improve much
if any((a in arch) for a in ["gfx90a", "gfx942", "gfx1100", "gfx1101", "gfx1151"]): # TODO: more arches, TODO: gfx950
if any((a in arch) for a in ["gfx90a", "gfx942", "gfx1100", "gfx1101", "gfx1102", "gfx1151"]): # TODO: more arches, TODO: gfx950
ENABLE_PYTORCH_ATTENTION = True
if rocm_version >= (7, 0):
if any((a in arch) for a in ["gfx1201"]):