-
Notifications
You must be signed in to change notification settings - Fork 22
Modify M preset for workloads to better fit systems with integrated GPU and 16GB of ram #287
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
Modify M preset for workloads to better fit systems with integrated GPU and 16GB of ram #287
Conversation
|
@diptorupd @ZzEeKkAa dpnp uses this numbers to report performance. Previous numbers causes swaps on 16GB ram systems preventing measurements. |
f300af1 to
90e95eb
Compare
|
@diptorupd pairwise distance fails on ci on numba_dpex_p. Should I add it to expected failures? |
0baafac to
6179764
Compare
|
rambo fails on windows with numba_dpex_p |
antonwolfy
left a comment
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.
What is about to halve the M for l2_norm?
37d6a32 to
12d21e0
Compare
|
@antonwolfy Fixed l2_norm. |
12d21e0 to
5bff057
Compare
I've checked on the laptop with |
|
@diptorupd @ZzEeKkAa ping |
diptorupd
left a comment
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.
Thank you!
|
CI failures are unrelated to the change, so going ahead and merging. |
|
This PR is good to go, the current issue is similar to one that was fixed in IntelPython/numba-dpex#1093. For testing CI installs old version of numba - 0.56.4. We can merge this one and I'll work on PR that fixes this issue. |
|
@ZzEeKkAa I can't merge this PR since I don't have sufficient permissions |
|
@AlexanderKalistratov I've set it for auto-mega, once conflicts are resolved. I'm working on those issues in separate PR and should finish it by EOD. |
5bff057 to
e4f1682
Compare
Adding laptop friendly preset size M16Gb to some workloads