Mark the TriangularSolverMatrix "kernel" functions as EIGEN_DEVICE_FUNC

[fwyzard]

This should fix cms-sw/cmssw#42460 .

For reference, the changes are cms-externals/eigen-git-mirror@e021252 :

diff --git a/Eigen/src/Core/products/TriangularSolverMatrix.h b/Eigen/src/Core/products/TriangularSolverMatrix.h
index 9f5ead2d6..7354d8df2 100644
--- a/Eigen/src/Core/products/TriangularSolverMatrix.h
+++ b/Eigen/src/Core/products/TriangularSolverMatrix.h
@@ -21,6 +21,7 @@ template <typename Scalar, typename Index, int Mode, bool Conjugate, int TriStor
 struct trsmKernelL {
   // Generic Implementation of triangular solve for triangular matrix on left and multiple rhs.
   // Handles non-packed matrices.
+  EIGEN_DEVICE_FUNC
   static void kernel(
     Index size, Index otherSize,
     const Scalar* _tri, Index triStride,
@@ -31,6 +32,7 @@ template <typename Scalar, typename Index, int Mode, bool Conjugate, int TriStor
 struct trsmKernelR {
   // Generic Implementation of triangular solve for triangular matrix on right and multiple lhs.
   // Handles non-packed matrices.
+  EIGEN_DEVICE_FUNC
   static void kernel(
     Index size, Index otherSize,
     const Scalar* _tri, Index triStride,
@@ -38,6 +40,7 @@ struct trsmKernelR {
 };
 
 template <typename Scalar, typename Index, int Mode, bool Conjugate, int TriStorageOrder,int OtherInnerStride>
+EIGEN_DEVICE_FUNC
 EIGEN_STRONG_INLINE void trsmKernelL<Scalar, Index, Mode, Conjugate, TriStorageOrder, OtherInnerStride>::kernel(
     Index size, Index otherSize,
     const Scalar* _tri, Index triStride,
@@ -89,6 +92,7 @@ EIGEN_STRONG_INLINE void trsmKernelL<Scalar, Index, Mode, Conjugate, TriStorageO
 
 
 template <typename Scalar, typename Index, int Mode, bool Conjugate, int TriStorageOrder, int OtherInnerStride>
+EIGEN_DEVICE_FUNC
 EIGEN_STRONG_INLINE void trsmKernelR<Scalar, Index, Mode, Conjugate, TriStorageOrder, OtherInnerStride>::kernel(
     Index size, Index otherSize,
     const Scalar* _tri, Index triStride,

[fwyzard]

enable gpu

[fwyzard]

please test

[cmsbuild]

A new Pull Request was created by @fwyzard (Andrea Bocci) for branch IB/CMSSW_13_3_X/master.

@smuzaffar, @iarspider, @aandvalenzuela can you please review it and eventually sign? Thanks.
@sextonkennedy, @antoniovilela, @rappoccio you are the release manager for this.
cms-bot commands are listed here

[fwyzard]

Note that the compiler would warn about the underlying error, but Eigen silences those warnings.
Maybe we should remove that silencing ?

The culprit seems to be https://github.com/cms-externals/eigen-git-mirror/blob/e021252e5767868dac39ab75a73616ec871460ec/Eigen/src/Core/util/DisableStupidWarnings.h#L117 :

  EIGEN_NV_DIAG_SUPPRESS(20011)

[cmsbuild]

-1

Failed Tests: UnitTests GpuUnitTests
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-895c67/35367/summary.html
COMMIT: 50889cb
CMSSW: CMSSW_13_3_X_2023-10-23-2300/el8_amd64_gcc12
Additional Tests: GPU
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmsdist/8782/35367/install.sh to create a dev area with all the needed externals and cmssw changes.

Unit Tests

I found 1 errors in the following unit tests:

---> test TestGeneratorInterfacePythia8ConcurrentGeneratorFilter had ERRORS

GPU Unit Tests

I found 1 errors in the following unit tests:

---> test testTFVisibleDevicesCUDA had ERRORS

Comparison Summary

Summary:

• You potentially removed 7 lines from the logs
• Reco comparison results: 4 differences found in the comparisons
• DQMHistoTests: Total files compared: 50
• DQMHistoTests: Total histograms compared: 3357400
• DQMHistoTests: Total failures: 3
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3357375
• DQMHistoTests: Total skipped: 22
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 49 files compared)
• Checked 214 log files, 167 edm output root files, 50 DQM output files
• TriggerResults: no differences found

GPU Comparison Summary

Summary:

• No significant changes to the logs found
• Reco comparison results: 0 differences found in the comparisons
• DQMHistoTests: Total files compared: 3
• DQMHistoTests: Total histograms compared: 39740
• DQMHistoTests: Total failures: 21
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 39719
• DQMHistoTests: Total skipped: 0
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 2 files compared)
• Checked 8 log files, 10 edm output root files, 3 DQM output files
• TriggerResults: no differences found

[fwyzard]

The failing tests are unrelated to these changes, while testEigenGPUNoFit_t now passes: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-895c67/35367/gpuUnitTests/src/RecoTracker/PixelTrackFitting/test/testEigenGPUNoFit_t/testing.log .

[fwyzard]

@smuzaffar can this be merged ?

[smuzaffar]

+externals

unit tests failures are not related to this PR

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next IB/CMSSW_13_3_X/master IBs (but tests are reportedly failing). This pull request will now be reviewed by the release team before it's merged. @rappoccio, @sextonkennedy, @antoniovilela (and backports should be raised in the release meeting by the corresponding L2)