-
Notifications
You must be signed in to change notification settings - Fork 577
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
Use ninja job pool limits to fix cuda+rdc+static builds on 'ride' and 'waterman' (#4502) #4761
Use ninja job pool limits to fix cuda+rdc+static builds on 'ride' and 'waterman' (#4502) #4761
Conversation
…odules Now this is back to a full SEMS set of modules.
Origin repo remote tracking branch: 'github/master' Origin repo remote repo URL: 'github = git@github.com:TriBITSPub/TriBITS.git' At commit: commit 7619c97a3991115e60e886cc711cf30ae7d58a0b Author: Roscoe A. Bartlett <rabartl@sandia.gov> Date: Thu Mar 28 16:18:43 2019 -0600 Summary: Add vars for ninja job pool limits and always enable ninja testing if can be found (trilinos#281)
…mits (trilinos#4502) Bring in update of TriBITS to allow limiting number of link processes for ATDM cuda+rdc+static builds.
This makes it so that you just terminate the <build-name> with '_pt' or '-pt' and the configuration will be switched to the Primary Tested packages with no disables and enabling Fortran.
This will allow limiting parallel link jobs for cuda+rdc builds.
…4502) This reduces the overall ninja parallel build level from 64 to 32 processes and the parallel link level from 64 to 16.
…nos#4502) This reduces the overall ninja parallel build level from 64 to 32 processes and the parallel link level from 64 to 16.
# Currently, EMPIRE configures of Trilinos have this enabled by default. But | ||
# really we should elevate every subpackage that ATDM uses to Primary Tested. | ||
# That is the right solution. | ||
ATDM_SET_ENABLE(Trilinos_ENABLE_SECONDARY_TESTED_CODE ON) |
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.
do we know what's not PT that empire uses?
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.
@rppawlo asked:
do we know what's not PT that empire uses?
That is what needs to be nailed down in #2597. We already know of some for SPARC and Gemma. For some reason, the EMPIRE configuration was explicitly setting Trilinos_ENABLE_SECONDARY_TESTED_CODE=ON.
It would not be hard to figure that out what ST SE packages (if any) EMPIRE is using that are not PT SE packages. Some quick configure experiments setting Trilinos_ENABLE_SECONDARY_TESTED_CODE to on an off would tell you that.
So many things to do ...
Status Flag 'Pre-Test Inspection' - Auto Inspected - Inspection Is Not Necessary for this Pull Request. |
Status Flag 'Pull Request AutoTester' - Testing Jenkins Projects: Pull Request Auto Testing STARTING (click to expand)Build InformationTest Name: Trilinos_pullrequest_gcc_4.8.4
Jenkins Parameters
Build InformationTest Name: Trilinos_pullrequest_intel_17.0.1
Jenkins Parameters
Build InformationTest Name: Trilinos_pullrequest_gcc_4.9.3_SERIAL
Jenkins Parameters
Build InformationTest Name: Trilinos_pullrequest_gcc_7.2.0
Jenkins Parameters
Build InformationTest Name: Trilinos_pullrequest_cuda_9.2
Jenkins Parameters
Using Repos:
Pull Request Author: bartlettroscoe |
Status Flag 'Pull Request AutoTester' - Jenkins Testing: all Jobs PASSED Pull Request Auto Testing has PASSED (click to expand)Build InformationTest Name: Trilinos_pullrequest_gcc_4.8.4
Jenkins Parameters
Build InformationTest Name: Trilinos_pullrequest_intel_17.0.1
Jenkins Parameters
Build InformationTest Name: Trilinos_pullrequest_gcc_4.9.3_SERIAL
Jenkins Parameters
Build InformationTest Name: Trilinos_pullrequest_gcc_7.2.0
Jenkins Parameters
Build InformationTest Name: Trilinos_pullrequest_cuda_9.2
Jenkins Parameters
|
Status Flag 'Pre-Merge Inspection' - SUCCESS: The last commit to this Pull Request has been INSPECTED AND APPROVED by [ rppawlo ]! |
Status Flag 'Pull Request AutoTester' - Pull Request will be Automerged |
Merge on Pull Request# 4761: IS A SUCCESS - Pull Request successfully merged |
This PR should fix the cuda+rdc+static builds on 'waterman' and 'ride' (#4502). I built from complete scratch on 'ride' and 'waterman' and the builds completed with no build errors at all and the only test failures where those that we already knew about for the
Trilinos-atdm-white-ride-cuda-9.2-gnu-7.2.0-rdc-release-debug-pt
build on 'ride' and 'white'.This required an update to TriBITS to take advantage of the CMake Ninja job pool levels. See:
After this merge, we will be able to promote the builds:
to the "ATDM" CDash group since they should be 100% clean (at least they are for this Trilinos version). Then later perhaps set up a cuda+rdc+shared PR build to protect RDC before merging to 'develop'.
Info about the tests run are detailed below with test build and test results.
Build and test results summary (click to expand)
Testing out on 'ride':
That actually passed showing:
It is good news that passed. It means the ATDM Trilinos builds targeting the ATDM APPs passes with RDC after this change!
Unfortunately, that is not all of the PT packages :-( Something went wrong wtih the new 'pt' build-name keyword support. I see the error. The var name was wrong in:
I fixed that and am now trying again on 'ride':
This time it is not disabling the packages ATDM is not using:
Thus gave the result:
We already know about those failing tests in issues #4042, #3749, #3542.
What is interesting is there the STK test in issue #3544 is not failing.
I made the same changes on 'waterman' and am now testing there as well:
That fully passed with the result:
So we can go ahead and promote that build once this merges!