comfy-aimdo 0.2 - Improved pytorch allocator integration (#12557)

Integrate comfy-aimdo 0.2 which takes a different approach to
installing the memory allocator hook. Instead of using the complicated
and buggy pytorch MemPool+CudaPluggableAlloctor, cuda is directly hooked
making the process much more transparent to both comfy and pytorch. As
far as pytorch knows, aimdo doesnt exist anymore, and just operates
behind the scenes.

Remove all the mempool setup stuff for dynamic_vram and bump the
comfy-aimdo version. Remove the allocator object from memory_management
and demote its use as an enablment check to a boolean flag.

Comfy-aimdo 0.2 also support the pytorch cuda async allocator, so
remove the dynamic_vram based force disablement of cuda_malloc and
just go back to the old settings of allocators based on command line
input.

cuda_malloc.py

@@ -1,10 +1,8 @@
 import os
 import importlib.util
-from comfy.cli_args import args, PerformanceFeature, enables_dynamic_vram
+from comfy.cli_args import args, PerformanceFeature
 import subprocess
 
-import comfy_aimdo.control
-
 #Can't use pytorch to get the GPU names because the cuda malloc has to be set before the first import.
 def get_gpu_names():
     if os.name == 'nt':
@@ -87,10 +85,6 @@ if not args.cuda_malloc:
     except:
         pass
 
-if enables_dynamic_vram() and comfy_aimdo.control.init():
-    args.cuda_malloc = False
-    os.environ['PYTORCH_CUDA_ALLOC_CONF'] = ""
-
 if args.disable_cuda_malloc:
     args.cuda_malloc = False