Hey Cory,
On 2024-02-28 21:16, Cordell Bloor wrote:
> This segfault does seem to be caused by mixing clang-15 and clang-17 in
> the HIP RTC codepath. When libamdhip64 from ROCm 5.6.1 (built with the
> same clang-17 as rocm-compilersupport 6.0+git20231212.4510c28+dfsg-1) is
> used, the segfault dis
This segfault does seem to be caused by mixing clang-15 and clang-17 in
the HIP RTC codepath. When libamdhip64 from ROCm 5.6.1 (built with the
same clang-17 as rocm-compilersupport 6.0+git20231212.4510c28+dfsg-1) is
used, the segfault disappeared [1].
Sincerely,
Cory Bloor
[1]:
https://ci.ro
The segfault on the very first rocfft test (and in no other library) is
probably a good indication that HIP runtime compilation (RTC) is broken.
That is a feature that is used for every rocFFT function, but not used
by any other ROCm library.
Whether this bug is because the HIP stack is curren
Package: libamd-comgr2
Version: 6.0+git20231212.4510c28+dfsg-1~exp2
Severity: important
X-Debbugs-Cc: c...@slerp.xyz
Dear Maintainer,
The rocfft tests began segfaulting on all architectures when
rocm-compilersupport 6.0+git20231212.4510c28+dfsg-1~exp2 was uploaded to
unstable. You can see it from
4 matches
Mail list logo
