jax
Fix batched Cholesky OOM on ROCm by bypassing hipSolver's external hipMalloc
#717
Open

Fix batched Cholesky OOM on ROCm by bypassing hipSolver's external hipMalloc #717

FlemingH wants to merge 43 commits into ROCm:amd-main from FlemingH:amd-main
FlemingH
charleshofer Remove nvidia_wheel_versions
097ff78e
charleshofer Make jaxlib targets visible
19a87c5b
charleshofer hipblas typedef fix
35b23689
charleshofer No GPU fail
9bf2dbf6
mminutoli Wrap HIP inline functions in anonymous namespaces in vendor.h
7d1708e4
dsicarov-amd SWDEV-512768 - Replace hipGetLastError with hipExtGetLastError
30d7f94d
charleshofer Add shared utility function get_rocm_version to test_util.py
a5377e54
phambinhfin Fix hipSparse CSR algorithm mappings for ROCm 7
db30afa3
phambinhfin Fix v_pages quantization and adjust test params for ROCm compatibilit…
a44f9428
Arech8 Address LLVM assertion failure due to a multithreaded use. Update .gi…
01746ea7
Arech8 Add skip of test_is_finite() on Cuda (#565)
f5555631
AratiGanesh Add rocm test requirements file (#570)
8cf787ad
charleshofer Let the unit tests use build.py for setting up Bazel commands for uni…
17e6022f
gulsumgudukbay adding abort logic to rocm/jax (#590)
b600136d
phambinhfin Skip is_finite tests on ROCm (not in Triton lowering for jax 0.8.0) (…
02399d0b
phambinhfin Fix shared memory limit check for ROCm in test_dot (#596)
0959b0fe
magaonka-amd Fix Numpy signatures test (#598)
b43ca180
Ruturaj4 fix merge arts
cdb5bcb5
gulsumgudukbay Enable RngShardingTests (#644)
de1ef417
mminutoli Enable test_variadic_reduce_window on ROCm (#647)
d8179cdf
magaonka-amd Skip sparse tests on ROCm due to hipSPARSE issue (#652)
c5016efc
magaonka-amd Update sparse test skip messages in v0.8.2 (#653)
4e6626ea
magaonka-amd Skip sparse tests on ROCm due to hipSPARSE issue (#652)
694e861d
magaonka-amd Update sparse test skip messages in v0.8.2 (#653)
12e07fb9
magaonka-amd Skip sparse tests on ROCm due to hipSPARSE issue (#652)
76e576f6
magaonka-amd Update sparse test skip messages in v0.8.2 (#653)
237e5ad8
AratiGanesh Enable testMultivariateNormalSingularCovariance on ROCm (#666)
da3a3cce
AratiGanesh Skip test_tridiagonal_solve on ROCm due to hipSPARSE numerical errors…
06d459eb
gulsumgudukbay Update Skip Reason Outputs (#663)
c30a4494
magaonka-amd Skip sparse tests on ROCm due to hipSPARSE issue (#652)
58ce4e11
magaonka-amd Update sparse test skip messages in v0.8.2 (#653)
e8307d2c
magaonka-amd Skip testCudaArrayInterfaceOnNonCudaFails on ROCm platform (#677)
64ee74ec
magaonka-amd Skip sparse tests on ROCm due to hipSPARSE issue (#652)
f1441325
magaonka-amd Update sparse test skip messages in v0.8.2 (#653)
9dd1698e
magaonka-amd Skip sparse tests on ROCm due to hipSPARSE issue (#652)
fd1195e3
magaonka-amd Update sparse test skip messages in v0.8.2 (#653)
4af53277
AratiGanesh Add ROCm encoding for test_struct_encoding_determinism (#683)
bfe02082
magaonka-amd Remove 'mean' from unsupported params for jnp.var (#689)
44b8a6c3
magaonka-amd Implement approx_tanh for ROCm using OCML tanh function (#691)
7bd4a132
AratiGanesh Skipping testEighTinyNorm due to hipSolver issues (#697)
95ae9faa
gulsumgudukbay Abort detection CI workflow (#688)
e355fcd6
gulsumgudukbay Abort-Detection: Fix halt-for-connection input (#712)
d36ebc22
hawkinsp
hawkinsp commented on 2026-02-27
FlemingH FlemingH closed this 27 days ago
FlemingH FlemingH force pushed from 2ccdebda to d36ebc22 27 days ago
FlemingH Fix batched Cholesky OOM on ROCm by routing hipSOLVER workspace throu…
5b647ae9
FlemingH FlemingH reopened this 27 days ago
mminutoli mminutoli force-pushed the amd-main branch from d36ebc22 to 7d684aa5 26 days ago
mminutoli mminutoli force-pushed the amd-main branch from 7d684aa5 to a0cd2621 17 days ago
mminutoli mminutoli force-pushed the amd-main branch from 69c86c70 to 2c2d75c5 10 days ago
mminutoli mminutoli force-pushed the amd-main branch from 2c2d75c5 to 7d9fd7e1 8 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone