{2026.06}[2026.1] SciPy-bundle 2026.05 - #1585
Conversation
|
bot: build repo:eessi.io-2026.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
bot: build repo:eessi.io-2026.06-software instance:eessi-bot-riscv-prod for:arch=riscv64/generic |
|
New job on instance
|
|
Failing tests on RISC-V: |
|
Going to ignore RISC-V for now. bot: build repo:eessi.io-2026.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen3 |
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
There are test failures for scipy for various targets. aarch64/generic + neoverse n1: neoverse v1 test failures: |
|
@bedroge Wouldn't be the first time we need to ignore some test failures on It is interesting that all tests are passing for Any more details on how the tests fail exactly? Is it persistent? |
|
bot: build repo:eessi.io-2026.06-software instance:eessi-bot-mc-aws for:arch=aarch64/generic |
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
True, though we didn't have to do this for recent versions, and neither for generic/neoverse n1: https://github.com/EESSI/software-layer-scripts/blob/main/eb_hooks.py#L1852. Most of them are numerical errors: |
|
The ones for neoverse v1: @boegel These are actually very similar to the ones that you reported in scipy/scipy#25346, and for which a patch was added in easybuilders/easybuild-easyconfigs#26223. |
|
Tried with some different OpenBLAS versions (also the one that works fine in 2025.06), but that didn't solve it. Now trying different numpy/scipy versions... |
|
@bedroge Let's open issues upstream for this, see what they provide as feedback. I would make a separate issue for the failing And then also an issue for the bunch of failing tests we see on |
|
OpenBLAS has been rebuilt, and this should solve the issue for bot: build repo:eessi.io-2026.06-software instance:eessi-bot-mc-aws for:arch=aarch64/generic |
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
@bedroge looks good to deploy? edit: ah, no, |
|
bot: build repo:eessi.io-2026.06-software instance:eessi-bot-aws-eu-south for:arch=x86_64/intel/graniterapids |
|
New job on instance
|
|
@bedroge Did you check what the problem is for |
Yeah, see numpy/numpy#32376. |
OK, that shouldn't block this PR then... |
|
staging PR merged |
No description provided.