Add a supports_fp64 function. (Comfy-Org#13368)

Author: comfyanonymous

comfy/ldm/flux/math.py

@@ -16,7 +16,7 @@ def attention(q: Tensor, k: Tensor, v: Tensor, pe: Tensor, mask=None, transforme
 
 def rope(pos: Tensor, dim: int, theta: int) -> Tensor:
     assert dim % 2 == 0
-    if comfy.model_management.is_device_mps(pos.device) or comfy.model_management.is_intel_xpu() or comfy.model_management.is_directml_enabled():
+    if not comfy.model_management.supports_fp64(pos.device):
         device = torch.device("cpu")
     else:
         device = pos.device

comfy/model_management.py

@@ -1732,6 +1732,21 @@ def supports_mxfp8_compute(device=None):
 
     return True
 
+def supports_fp64(device=None):
+    if is_device_mps(device):
+        return False
+
+    if is_intel_xpu():
+        return False
+
+    if is_directml_enabled():
+        return False
+
+    if is_ixuca():
+        return False
+
+    return True
+
 def extended_fp16_support():
     # TODO: check why some models work with fp16 on newer torch versions but not on older
     if torch_version_numeric < (2, 7):