-
Notifications
You must be signed in to change notification settings - Fork 908
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
Correctly compile benchmarks #7485
Correctly compile benchmarks #7485
Conversation
CPM will force these variables to OFF and add them to the CMakeCache. This causes problems when you re-execute cmake in the same build directory as now they are set to OFF, and CUDF will not enable benchmarks/tests.
Does the same transitive propagation issue exist with |
Good question. I will update those just in case. |
Codecov Report
@@ Coverage Diff @@
## branch-0.19 #7485 +/- ##
===============================================
+ Coverage 81.85% 82.26% +0.40%
===============================================
Files 101 101
Lines 16883 17244 +361
===============================================
+ Hits 13819 14185 +366
+ Misses 3064 3059 -5
Continue to review full report at Codecov.
|
@gpucibot merge |
The refactor in #7107 introduced two subtle bugs that caused benchmarks to not build correctly.
benchmark_main
target needs to be explicitly placed on the final executable link line as it injects object files, and that doesn't work via transitive propagation.BUILD_BENCHMARKS
andBUILD_TESTING
are shared between RMM and CUDF. When CUDF builds RMM via CPM it will force cache these variables toOFF
. This means that if a developer try to runcmake -DBUILD_BENCHMARKS=ON .
any subsequent execution of justcmake .
will cause CUDF to forget it should have kept benchmarks enabled.