Skip to content

Commit cf47da7

Browse files
committed
Fix cuTENSORMp CUDA runtime preload
CUDA 12's libcutensorMp.so.2 directly depends on libcudart.so.12, but its wheel RUNPATH does not include the cuda_runtime package. Pathfinder can therefore find and preload cuTENSOR, cuBLASLt, and NCCL yet still fail to dlopen cuTENSORMp in a wheel-only environment unless libcudart is already available from an ambient toolkit. Add cudart to the descriptor dependency list so Pathfinder resolves the packaged CUDA runtime before opening cuTENSORMp. This makes CUDA 12 wheel-only loading independent of LD_LIBRARY_PATH while preserving CUDA 13 and local-toolkit behavior.
1 parent 75f221c commit cf47da7

1 file changed

Lines changed: 1 addition & 1 deletion

File tree

‎cuda_pathfinder/cuda/pathfinder/_dynamic_libs/descriptor_catalog.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -380,7 +380,7 @@ class DescriptorSpec:
380380
packaged_with="other",
381381
linux_sonames=("libcutensorMp.so.2",),
382382
site_packages_linux=("cutensor/lib",),
383-
dependencies=("cutensor", "cublasLt", "nccl"),
383+
dependencies=("cutensor", "cublasLt", "cudart", "nccl"),
384384
),
385385
DescriptorSpec(
386386
name="nccl",

0 commit comments

Comments
 (0)