Skip to content

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

Speed up inference on nvidia 10 series on Linux.

comfyanonymous8d31a6632fdf
diff --git a/comfy/model_management.py b/comfy/model_management.py
index f6a76a572..11683a905 100644
--- a/comfy/model_management.py
+++ b/comfy/model_management.py
@@ -370,8 +370,10 @@ def offloaded_memory(loaded_models, device):
offloaded_mem += m.model_offloaded_memory()
return offloaded_mem
+WINDOWS = any(platform.win32_ver())
+
EXTRA_RESERVED_VRAM = 400 * 1024 * 1024
-if any(platform.win32_ver()):
+if WINDOWS:
EXTRA_RESERVED_VRAM = 600 * 1024 * 1024 #Windows is higher because of the shared vram issue
if args.reserve_vram is not None:
@@ -1002,7 +1004,10 @@ def should_use_fp16(device=None, model_params=0, prioritize_performance=True, ma
nvidia_10_series = ["1080", "1070", "titan x", "p3000", "p3200", "p4000", "p4200", "p5000", "p5200", "p6000", "1060", "1050", "p40", "p100", "p6", "p4"]
for x in nvidia_10_series:
if x in props.name.lower():
- return True
+ if WINDOWS or manual_cast:
+ return True
+ else:
+ return False #weird linux behavior where fp32 is faster
if manual_cast:
free_model_memory = maximum_vram_for_weights(device)