-
Notifications
You must be signed in to change notification settings - Fork 8
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Remove CUDA from the test extras #1791
Conversation
The tests are still pretty slow, but I've got some ideas on how we can improve things. |
This should at least improve code loading time when running ClimaCore's test suite (it does seem to start much more quickly with |
I'm going to peel some pieces off from this.. |
8f2012e
to
0c1e5b5
Compare
bc52c60
to
bcecc63
Compare
bcecc63
to
cf36886
Compare
This PR is not ideal in that it uses That said, this does remove CUDA from the test environment. |
If you are not using macros from CUDA, you can access it with |
cf36886
to
7d6d846
Compare
db90343
to
c5a01be
Compare
c5a01be
to
5ed9017
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yay!
wip Use get_extension instead of eval-parse Add composed function stencil benchmark
5ed9017
to
efe07b7
Compare
This PR removes CUDA from the test environment.
I did comment out
Operators/spectralelement/opt.jl
of runtests.jl (it's still run on buildkite), due to aviatesk/JET.jl#602.I think soon we'll split off our tests into
runtests_opt.jl
that performs all allocation, flop-counting, and inference tests.runtests_benchmark.jl
that performs all benchmarksruntests_convergence.jl
that performs all convergence tests(#1784)