Add a ModelAttentionBackend node to manually select the attention for models in the workflows. Currently supports pytorch attention or comfy kitchen attention.
Add --use-ck-attention to enable comfy kitchen attention as the default attention backend for all models (might break some).
Per AGENTS.md: prefer specific exception types in new code. AttributeError
covers torch.version.cuda being None, ValueError covers an unparseable
version string.
cuBLAS FP4 matmul kernels (cublasLtMatmulAlgoGetHeuristic) require
CUDA 13.0+. On Blackwell GPUs with torch built against CUDA <13
(e.g. cu128), the previous check only looked at compute capability
and let native NVFP4 through, causing CUBLAS_STATUS_NOT_SUPPORTED
errors or VRAM blowups at matmul time. Now falls back to the
regular quantized-storage path when the CUDA build is too old.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This avoids name collision (circular imports) for external custom nodes,
for which the comfy path is pushed into sys.path so Python's own logging module
is shadowed otherwise.
fixes: #15229
This priority scheme was broken in the case where you have pin
registration exhaustion while loading a VBAR that gets a big evicition.
The weight would stay in the loaded set but inherit the MRU priority
against other workflow models WRT pin registration which leads to async
offload without pinning.
Fix by universally promiting active pin registration above workflow
pins without concern for the weights/weights-loaded split. This diverges
from the actual budgeting where the split still makes sense.
Changes:
Remove sequential scan hint
Prefer NVML pressure on windows
Add async malloc clamp option (unused by comfy so far)
Workaround AMD windows GPU virtual address space leak
The largest change is the NVML pressure, which works around a cuMemGetInfo
drift from actual VRAM in some circumstances.
Windows has proven this logic works for a long time and there are
corner cases where this materialization actual consumes real RAM
on linux.
Its not as bad as the original windows commit charge surge, but
its still a detectable transient leak. So simplify and unify.
Alan Li
Jukka Seppänen
Barish Ozbay
comfyanonymous
rattus
Alexander Brown
Alex Harper
Alexis Rolland
blepping
endman100
rivadart
Oliver Freyermuth