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
Stabilize deltaPhi #12843
Stabilize deltaPhi #12843
Conversation
A new Pull Request was created by @VinInn (Vincenzo Innocente) for CMSSW_8_0_X. It involves the following packages: Alignment/OfflineValidation @civanch, @diguida, @cvuosalo, @cerminar, @monttj, @cmsbuild, @franzoni, @mdhildreth, @slava77, @vadler, @mmusich, @davidlange6 can you please review it and eventually sign? Thanks. Following commands in first line of a comment are recognized
|
@cmsbuild, please test |
The tests are being triggered in jenkins. |
You can force compiler to produce "compatible" results between run-time and compile-time, but that mostly means utilising less efficient instructions (e.g. it could disable FMA). |
@cmsbuild , please test |
The tests are being triggered in jenkins. |
-1 >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/HiGenCleaner.cc >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/MultipleAlgoIterator.cc >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/ParametrizedSubtractor.cc >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/ParticleTowerProducer.cc In file included from /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/ParticleTowerProducer.cc:35:0: /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/poison/RecoHI/HiJetAlgos/plugins/ParticleTowerProducer.h:1:2: error: #error THIS FILE HAS BEEN REMOVED FROM THE PACKAGE. #error THIS FILE HAS BEEN REMOVED FROM THE PACKAGE. ^ >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/ReflectedIterator.cc >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/UEParameters.cc >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/VoronoiSubtractor.cc you can see the results of the tests here: |
-1 >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/MultipleAlgoIterator.cc Copying tmp/slc6_amd64_gcc493/src/CondFormats/SiPixelObjects/test/testSerializationSiPixelObjects/testSerializationSiPixelObjects to productstore area: >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/ParametrizedSubtractor.cc Leaving library rule at src/CondFormats/SiPixelObjects/test >> Compiling /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/ParticleTowerProducer.cc /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/ParticleTowerProducer.cc:35:63: fatal error: RecoHI/HiJetAlgos/interface/ParticleTowerProducer.h: No such file or directory #include "RecoHI/HiJetAlgos/interface/ParticleTowerProducer.h" ^ compilation terminated. /tmp/cmsbuild/workspace/ib-any-integration/CMSSW_8_0_X_2015-12-22-1100/src/RecoHI/HiJetAlgos/src/ParticleTowerProducer.cc:35:63: fatal error: RecoHI/HiJetAlgos/interface/ParticleTowerProducer.h: No such file or directory #include "RecoHI/HiJetAlgos/interface/ParticleTowerProducer.h" you can see the results of the tests here: |
I'm not sure why we should leave a broken implementation of acos. |
ok, I already made the change will copy and commit here... |
@cmsbuild, please test |
The tests are being triggered in jenkins. |
There are no regression. |
btw I made a quick study of the origin of large Phi-windows... |
there are some regressions in my local tests. |
+1
|
+1 |
implement deltaPhi w/o while loop (required among others by Ofast to avoid infinite loops due to NaN)
took longer because
Geom::Phi (that I will evict from CMSSW at some point) is incompatible with the new implementation
HI compiles some CGAL stuff with rounding-math that is incompatible with the use of compile-time expressions
deltaPhi-like algorithms exist in many places including two identical implementation of normalizedPhi
fixed also in TripletSeed and the origin of nan in tkdetUtil
needless to say that it will recompile everything
expect regression due to possible small numerical differences