Move at::{Refcounted,}MapAllocator to c10 (#109881)

`libshm.so` depends on the torch library exclusively for `at::RefcountedMapAllocator`,
 so it makes sense to move it to c10 along with the other memory allocators.

This means `libshm.so` only depends on `c10` and we don't need to relink
`libshm.so` for every ATen change.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/109881
Approved by: https://github.com/albanD
This commit is contained in:
Peter Bell
2023-10-09 17:43:42 +00:00
committed by PyTorch MergeBot
parent 3704bf4ee8
commit 0341deb1c7
17 changed files with 1154 additions and 798 deletions

View File

@ -1001,7 +1001,6 @@ aten_cpu_source_non_codegen_list = [
"aten/src/ATen/FunctionalTensorWrapper.cpp",
"aten/src/ATen/FunctionalizeFallbackKernel.cpp",
"aten/src/ATen/MemoryOverlap.cpp",
"aten/src/ATen/MapAllocator.cpp",
"aten/src/ATen/NamedTensorUtils.cpp",
"aten/src/ATen/NestedTensorImpl.cpp",
"aten/src/ATen/ParallelCommon.cpp",