Add Similarity_2_5
Merge request reports
Activity
- [2017-11-07 00:36] Validation started with lhcb-head#1653
- [2017-11-07 00:36] Validation started with lhcb-lcg-dev4#337
- [2017-11-07 00:36] Validation started with lhcb-lcg-dev3#337
- [2017-11-07 00:37] Validation started with lhcb-gaudi-head-noavx#187
- [2017-11-07 00:37] Validation started with lhcb-gaudi-head-noavx2#168
- [2017-11-07 00:37] Validation started with lhcb-clang-test#762
- [2017-11-07 00:37] Validation started with lhcb-gaudi-head#1651
- [2017-11-08 00:04] Validation started with lhcb-head#1654
- [2017-11-08 00:05] Validation started with lhcb-gaudi-head-noavx2#169
- [2017-11-08 00:05] Validation started with lhcb-clang-test#763
- [2017-11-08 00:05] Validation started with lhcb-gaudi-head-noavx#188
- [2017-11-08 00:06] Validation started with lhcb-lcg-dev3#338
- [2017-11-08 00:06] Validation started with lhcb-lcg-dev4#338
- [2017-11-08 00:06] Validation started with lhcb-gaudi-head#1652
- [2017-11-08 16:10] Validation started with lhcb-gaudi-merge#275
- [2017-11-08 18:19] Validation started with lhcb-gaudi-merge#276
- [2017-11-08 21:49] Validation started with lhcb-gaudi-merge#278
- [2017-11-08 21:49] Validation started with lhcb-gaudi-merge#277
- [2017-11-08 21:53] Validation started with lhcb-gaudi-merge#279
- [2017-11-09 00:05] Validation started with lhcb-lcg-dev3#339
- [2017-11-09 00:05] Validation started with lhcb-lcg-dev4#339
- [2017-11-09 00:05] Validation started with lhcb-clang-test#764
- [2017-11-09 00:05] Validation started with lhcb-head#1655
- [2017-11-09 00:06] Validation started with lhcb-gaudi-head-noavx#189
- [2017-11-09 00:06] Validation started with lhcb-gaudi-head#1653
- [2017-11-09 00:06] Validation started with lhcb-gaudi-head-noavx2#170
- [2017-11-09 08:15] Validation started with lhcb-gaudi-merge#281
- [2017-11-09 08:24] Validation started with lhcb-gaudi-merge#280
- [2017-11-09 10:14] Validation started with lhcb-head#1656
Edited by Software for LHCbCan you please add the
similarity_2_5
use case also to the test, https://gitlab.cern.ch/lhcb/LHCb/blob/master/Kernel/LHCbMath/tests/TestSimilarity.cpp (and corresponding reference files).The test is broken (did you try running it yesterday?)
In the opt builds, it fails with
*** Break *** segmentation violation #4 <signal handler called> #5 0x0000000000406eb6 in void fillSMatrixSymWithCondNumber<ROOT::Math::SMatrix<double, 2u, 2u, ROOT::Math::MatRepStd<double, 2u, 2u> >, ROOT::Math::SMatrix<double, 2u, 2u, ROOT::Math::MatRepSym<double, 2u> > >(ROOT::Math::SMatrix<double, 2u, 2u, ROOT::Math::MatRepSym<double, 2u> >&, TRandom&, double) ()
In the dbg builds the stderr may be more informative:
terminate called after throwing an instance of 'std::out_of_range' what(): map::at
(last stdout in above case was
============= Similarity_2_5 Test ============= Has SSE3: 1 Has AVX: 1 Has AVX2: 1 ISet = 8 | DoSSE3: 0 DoAVX: 0 Do AVX2: 1 Generic similarity transform result [ 0.346839 0.426094 0.447029 0.446696 0.155146 0.426094 0.572084 0.421402 0.423952 0.232367 0.447029 0.421402 0.91194 0.903735 0.0902007 0.446696 0.423952 0.903735 0.89571 0.0925936 0.155146 0.232367 0.0902007 0.0925936 0.105279 ]
See e.g. https://lhcb-nightlies.cern.ch/logs/tests/nightly/lhcb-head/1654/x86_64-slc6-gcc62-dbg/LHCb/ and https://lhcb-nightlies.cern.ch/logs/tests/nightly/lhcb-head/1654/x86_64-slc6-gcc62-opt/LHCb/
I have not tried running it yesterday. Actually I did not know how to do this. The problem is probably a hard coded matrix dimension in https://gitlab.cern.ch/lhcb/LHCb/blob/f4df8fb276545c2ede6fafa908212d8ac96c9c45/Kernel/LHCbMath/tests/TestSimilarity.cpp#L186
I will change it and try to figure out how to run the test.
Another problem is, that in the current TestSimilarity.cpp version it is not possible to test only the generic version (which probably also makes no sense). One could either remove again the test for Similarity_2_5, or use the generic version as place holder for possible later implementations of other versions. Any preferences?
mentioned in commit a766fe1b
mentioned in commit 11d69ca0
mentioned in merge request !1792 (closed)