Skip to content
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

Upgrade numpy to version 1.17.0 #6494

Closed
wants to merge 3 commits into from
Closed

Upgrade numpy to version 1.17.0 #6494

wants to merge 3 commits into from

Conversation

cmnbroad
Copy link
Collaborator

Fixes #6480. Note that xbyak issue is unchanged and still exists with this upgrade.

This also contains a test for some of the key python packages to verify that they are resolved by Conda to the version of the package requested in the .yml file.

@cmnbroad
Copy link
Collaborator Author

@samuelklee @mwalker174 @lucidtronix Tests seem to pass with this (numpy 1.18 seems to cause runtime errors for CNN, but 1.17 seems to work ok), so this should at least get us up to a version that addresses the gcc issues in this forum post.

@samuelklee
Copy link
Contributor

Great, thanks for adding this @cmnbroad! Let me manually check that there weren't any numerical changes in the gCNV WDL tests. @asmirnov239 and Jack might also want to test on some small, real data. Probably overkill, but just to be safe... We'll get back to you!

@samuelklee
Copy link
Contributor

Did some numerical spot checks against master, looks like we're typically good to 10+ decimal places for the very small gCNV WDL test data. I'm fine with this going in without checking on real data first, but I'll tell Andrey and Jack to keep an eye out. Thanks again for updating!

samuelklee
samuelklee previously approved these changes Mar 13, 2020
@samuelklee
Copy link
Contributor

samuelklee commented Mar 16, 2020

Wait, actually, let me check some issues discussed in #5781 and #5764. Is the pip-installed version of numpy compiled against MKL? I didn't realize the change in #5781 was reverted in #6413 while I was on leave.

@cmnbroad cmnbroad force-pushed the cn_upgrade_numpy branch 2 times, most recently from ad3bc5d to e5ca093 Compare March 16, 2020 20:15
@gatk-bot
Copy link

Travis reported job failures from build 29532
Failures in the following jobs:

Test Type JDK Job ID Logs
python openjdk8 29532.5 logs

@samuelklee samuelklee self-requested a review March 20, 2020 16:59
@samuelklee samuelklee dismissed their stale review March 20, 2020 17:00

Noticed inconsistencies, trying to fix over in sl_reshape.

@samuelklee
Copy link
Contributor

Reminder to myself to incorporate version checks from this branch over in #5026.

@samuelklee
Copy link
Contributor

@cmnbroad I incorporated the version checks over in #5026. Going to go ahead and close this, thanks again for putting it together.

@samuelklee samuelklee closed this Apr 2, 2020
@samuelklee samuelklee deleted the cn_upgrade_numpy branch April 2, 2020 17:30
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Upgrade numpy in the conda environment
3 participants