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
[RF] RooProdPdf test failure on s390x #12430
Comments
guitargeek
changed the title
RooFit test failure on s390x
[RF] RooProdPdf test failure on s390x
Mar 7, 2023
guitargeek
added a commit
to guitargeek/root
that referenced
this issue
Mar 7, 2023
In the ROOT 6.26 development cycle, the RooProdPdf was partially rewritten in moden C++ with less manual memory allocation to improve performance (PR root-project#7907). In that PR, a unit test that verifies the RooProdPdf can correctly deal with factorizing PDFs was implemented. However, that test used an arbitrary PDF where the correct factorization was checked in a rather crude way: check by hashing the content of the RooProdPdf cache element for a given normalization set that said PR doesn't change any behavior (the reference hash was hardcoded in the unit test). This commit suggests a better alternative for the unit test, checking for a multidimensional product pdf of factorizing uniform pdfs that the pdf values for differenc normalization sets is as expected. This should cover the same functionality and is less fragile and implementation dependend than hashing the cache elements. This closes GitHub issue root-project#12430, as the rewritten test is not affected anymore by the problem reported in that issue.
guitargeek
added a commit
that referenced
this issue
Mar 7, 2023
In the ROOT 6.26 development cycle, the RooProdPdf was partially rewritten in moden C++ with less manual memory allocation to improve performance (PR #7907). In that PR, a unit test that verifies the RooProdPdf can correctly deal with factorizing PDFs was implemented. However, that test used an arbitrary PDF where the correct factorization was checked in a rather crude way: check by hashing the content of the RooProdPdf cache element for a given normalization set that said PR doesn't change any behavior (the reference hash was hardcoded in the unit test). This commit suggests a better alternative for the unit test, checking for a multidimensional product pdf of factorizing uniform pdfs that the pdf values for differenc normalization sets is as expected. This should cover the same functionality and is less fragile and implementation dependend than hashing the cache elements. This closes GitHub issue #12430, as the rewritten test is not affected anymore by the problem reported in that issue.
guitargeek
added a commit
to guitargeek/root
that referenced
this issue
Mar 31, 2023
In the ROOT 6.26 development cycle, the RooProdPdf was partially rewritten in moden C++ with less manual memory allocation to improve performance (PR root-project#7907). In that PR, a unit test that verifies the RooProdPdf can correctly deal with factorizing PDFs was implemented. However, that test used an arbitrary PDF where the correct factorization was checked in a rather crude way: check by hashing the content of the RooProdPdf cache element for a given normalization set that said PR doesn't change any behavior (the reference hash was hardcoded in the unit test). This commit suggests a better alternative for the unit test, checking for a multidimensional product pdf of factorizing uniform pdfs that the pdf values for differenc normalization sets is as expected. This should cover the same functionality and is less fragile and implementation dependend than hashing the cache elements. This closes GitHub issue root-project#12430, as the rewritten test is not affected anymore by the problem reported in that issue.
guitargeek
added a commit
to guitargeek/root
that referenced
this issue
Mar 31, 2023
In the ROOT 6.26 development cycle, the RooProdPdf was partially rewritten in moden C++ with less manual memory allocation to improve performance (PR root-project#7907). In that PR, a unit test that verifies the RooProdPdf can correctly deal with factorizing PDFs was implemented. However, that test used an arbitrary PDF where the correct factorization was checked in a rather crude way: check by hashing the content of the RooProdPdf cache element for a given normalization set that said PR doesn't change any behavior (the reference hash was hardcoded in the unit test). This commit suggests a better alternative for the unit test, checking for a multidimensional product pdf of factorizing uniform pdfs that the pdf values for differenc normalization sets is as expected. This should cover the same functionality and is less fragile and implementation dependend than hashing the cache elements. This closes GitHub issue root-project#12430, as the rewritten test is not affected anymore by the problem reported in that issue.
guitargeek
added a commit
that referenced
this issue
Mar 31, 2023
In the ROOT 6.26 development cycle, the RooProdPdf was partially rewritten in moden C++ with less manual memory allocation to improve performance (PR #7907). In that PR, a unit test that verifies the RooProdPdf can correctly deal with factorizing PDFs was implemented. However, that test used an arbitrary PDF where the correct factorization was checked in a rather crude way: check by hashing the content of the RooProdPdf cache element for a given normalization set that said PR doesn't change any behavior (the reference hash was hardcoded in the unit test). This commit suggests a better alternative for the unit test, checking for a multidimensional product pdf of factorizing uniform pdfs that the pdf values for differenc normalization sets is as expected. This should cover the same functionality and is less fragile and implementation dependend than hashing the cache elements. This closes GitHub issue #12430, as the rewritten test is not affected anymore by the problem reported in that issue.
omazapa
pushed a commit
to omazapa/root
that referenced
this issue
Apr 13, 2023
In the ROOT 6.26 development cycle, the RooProdPdf was partially rewritten in moden C++ with less manual memory allocation to improve performance (PR root-project#7907). In that PR, a unit test that verifies the RooProdPdf can correctly deal with factorizing PDFs was implemented. However, that test used an arbitrary PDF where the correct factorization was checked in a rather crude way: check by hashing the content of the RooProdPdf cache element for a given normalization set that said PR doesn't change any behavior (the reference hash was hardcoded in the unit test). This commit suggests a better alternative for the unit test, checking for a multidimensional product pdf of factorizing uniform pdfs that the pdf values for differenc normalization sets is as expected. This should cover the same functionality and is less fragile and implementation dependend than hashing the cache elements. This closes GitHub issue root-project#12430, as the rewritten test is not affected anymore by the problem reported in that issue.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Describe the bug
The gtest-roofit-roofitcore-test-testRooProdPdf test fails on s390x.
This is a big-endian architecture.
See the log below.
The assert does not fail because the two values are the same with differnt byte order. The issue is more complicated than that::
$ printf '%x\n%x\n' 3649468902 2448666198
d98679e6
91f3ae56
Expected behavior
Passing test on s390x
To Reproduce
Setup
Additional context
Thefailing test:
The text was updated successfully, but these errors were encountered: