[Bug 2251086] Review Request: rocfft - ROCm Fast Fourier Transforms (FFT) library

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



https://bugzilla.redhat.com/show_bug.cgi?id=2251086



--- Comment #11 from Tim Flink <tflink@xxxxxxxxxx> ---
(In reply to Tom Rix from comment #6)
> For Ben's question about %check
> Checking depends on having an AMD GPU in the build machine, that is not
> possible.
> The -test subpackage is consistent with how other rocm packages have handled
> this.
> However in the other packages -test is off by default and not expected to be
> part of the offical fedora release.
> We should be consistent, can rocfft also have -test off by default ? 

I probably should have started a conversation about this before submitting this
for review, but I actually want to change all the rocm packages to build the
-test subpackages by default unless the behavior of rocfft is significantly
different from the other rocm packages. There's no other reasonable way I can
see to test the binaries shipped with Fedora - cmake will detect the lack of a
build to test against if you build the tests alone and build from source rather
than use an installed binary. You can use the rpm libraries for testing but
building the entire library for every test seems like an incredible waste,
especially if we end up enabling the various cache features at any point. The
build time for rocfft with kernel db enabled is over an hour on my machine
IIRC.

> New comments
> CMAKE_BUILD_TYPE=Debug
> We have not figured out how or if to do this.
> I know from my own experiments that this will cause packing errors when the
> rpm tooling does not handle some clang-ism.
> And if we did have a -debug, I would not want to *-d.so lib names.
> 
> Above, I also changed the RELEASE to RelWithDebInfo so it would be
> consistent with other packages.

Thanks, I didn't realize that was an option and have already had a bit of fun
with the changing library names when debug was enabled.

> The explicit list of AMDGPU_TARGETS is missing some, you can see in the
> cmake output that gfx94x is checked.
> If we need a full list, I think it would be useful to add that to
> rocm-rpm-macros package macros.rocm file.

I'm unclear on why we want to do this. I know that there are some rocm packages
where the full library is too large when all of the gpu targets are enabled but
rocfft isn't one of those. I get disabling some older targets to make the
library size reasonable but why disable older card families if we don't have
to?


-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
https://bugzilla.redhat.com/show_bug.cgi?id=2251086

Report this comment as SPAM: https://bugzilla.redhat.com/enter_bug.cgi?product=Bugzilla&format=report-spam&short_desc=Report%20of%20Bug%202251086%23c11
--
_______________________________________________
package-review mailing list -- package-review@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to package-review-leave@xxxxxxxxxxxxxxxxxxxxxxx
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/package-review@xxxxxxxxxxxxxxxxxxxxxxx
Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue




[Index of Archives]     [Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite Conditions]     [KDE Users]

  Powered by Linux