-
Notifications
You must be signed in to change notification settings - Fork 3
Panan Optimization #113
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
base: main
Are you sure you want to change the base?
Panan Optimization #113
Conversation
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
!redeploy |
|
🚀 Attempted to deploy 🖥️
|
|
!redeploy |
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
!redeploy |
|
🚀 Attempted to deploy 🖥️
|
|
!redeploy |
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
|
🚀 Attempted to deploy 🖥️
|
This is picking up from @minghangli-uni's work for helping optimize @claireyung's panan runs in preparation for her 10-year run.
Live results (5-day run):
pr75-1best of 3pr113-2best of 1pr113-2has a few changes: ifx instead of ifort,-{march,mtune}=sapphirerapidsand spack build changes. I was a little worried since sometimesifxcan be a lot slower thanifort.pr113-3best of 1-O3. Not much difference. This is because-O3isn't registered in most of our builds as-O2is hardcoded and placed after spack user-suppliedfflags.pr113-4best of 1-march=sapphirerapidsto-march=cascadelakefor compatibility with cascadelake nodes. A bit of a slow down is not surprising but still disappointing.pr113-10best of 2-flto -fuse-ld=lldto everything except CICE6. Disappointing result.pr113-20best of 1-qopt-prefetchto everything, resulting in a small improvement.pr113-21best of 1-O3takes precedence. Got a small boost.pr113-25best of 1-fltoto see if performance is better (since previous-fltoresult showed a slowdown). Does actually slow down, so would prefer to keep it.pr113-26best of 1[email protected]after manodeep's encouraging findings. Not much difference.pr113-27best of 1-fltoWhich seemed to help a decent amount.Current most preferred:
pr113-27for backwards compatibility.🚀 The latest prerelease
access-om3/pr113-3at 4899d3f is here: #113 (comment) 🚀🚀 The latest prerelease
access-om3/pr113-4at 89258cb is here: #113 (comment) 🚀🚀 The latest prerelease
access-om3/pr113-14at 8f4870d is here: #113 (comment) 🚀🚀 The latest prerelease
access-om3/pr113-18at 0377ee8 is here: #113 (comment) 🚀🚀 The latest prerelease
access-om3/pr113-20at 0377ee8 is here: #113 (comment) 🚀🚀 The latest prerelease
access-om3/pr113-21at f176d6b is here: #113 (comment) 🚀🚀 The latest prerelease
access-om3/pr113-22at 6d8dab5 is here: #113 (comment) 🚀🚀 The latest prerelease
access-om3/pr113-27at 11938bd is here: #113 (comment) 🚀🚀 The latest prerelease
access-om3/pr113-28at 7793329 is here: #113 (comment) 🚀