Skip to content

Only disable cudnn on newer AMD GPUs. #10437

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
comfyanonymous merged 1 commit into from
Oct 21, 2025
Merged

Only disable cudnn on newer AMD GPUs. #10437

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 10 additions & 4 deletions comfy/model_management.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -330,15 +330,21 @@ def amd_min_version(device=None, min_rdna_version=0):


SUPPORT_FP8_OPS = args.supports_fp8_compute

AMD_RDNA2_AND_OLDER_ARCH = ["gfx1030", "gfx1031", "gfx1010", "gfx1011", "gfx1012", "gfx906", "gfx900", "gfx803"]

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.")
arch = torch.cuda.get_device_properties(get_torch_device()).gcnArchName
if not (any((a in arch) for a in AMD_RDNA2_AND_OLDER_ARCH)):
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.")

try:
rocm_version = tuple(map(int, str(torch.version.hip).split(".")[:2]))
except:
rocm_version = (6, -1)
arch = torch.cuda.get_device_properties(get_torch_device()).gcnArchName

logging.info("AMD arch: {}".format(arch))
logging.info("ROCm version: {}".format(rocm_version))
if args.use_split_cross_attention == False and args.use_quad_cross_attention == False:
Expand Down Expand Up @@ -1331,7 +1337,7 @@ def should_use_bf16(device=None, model_params=0, prioritize_performance=True, ma

if is_amd():
arch = torch.cuda.get_device_properties(device).gcnArchName
if any((a in arch) for a in ["gfx1030", "gfx1031", "gfx1010", "gfx1011", "gfx1012", "gfx906", "gfx900", "gfx803"]): # RDNA2 and older don't support bf16
if any((a in arch) for a in AMD_RDNA2_AND_OLDER_ARCH): # RDNA2 and older don't support bf16
if manual_cast:
return True
return False
Expand Down