fix susy gg_tt builds and add it to the repo #625

The default HRDCOD=0 build presently fails ccache g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o In file included from Parameters_MSSM_SLHA2.cc:8: Parameters_MSSM_SLHA2.h:19:2: error: #error This non-SM physics process only supports MGONGPU_HARDCODE_PARAM builds (madgraph5#439): please run "make HRDCOD=1" 19 | #error This non-SM physics process only supports MGONGPU_HARDCODE_PARAM builds (madgraph5#439): please run "make HRDCOD=1" | ^~~~~ In file included from Parameters_MSSM_SLHA2.cc:8: Parameters_MSSM_SLHA2.h: In function ‘const Parameters_MSSM_SLHA2_dependentCouplings::DependentCouplings_sv Parameters_MSSM_SLHA2_dependentCouplings::computeDependentCouplings_fromG(const fptype_sv&)’: Parameters_MSSM_SLHA2.h:806:56: error: conversion from ‘fptype_sv’ {aka ‘__vector(4) double’} to non-scalar type ‘const mgOnGpu::cxsmpl<double>’ requested 806 | constexpr cxsmpl<double> mdl_G__exp__2 = ( ( G ) * ( G ) ); | ~~~~~~~~^~~~~~~~~ Parameters_MSSM_SLHA2.h:809:31: error: ‘mdl_I51x11’ was not declared in this scope 809 | out.GC_51 = -( cI * G * mdl_I51x11 ); | ^~~~~~~~~~ Parameters_MSSM_SLHA2.cc: In member function ‘void Parameters_MSSM_SLHA2::setIndependentParameters(SLHAReader&)’: Parameters_MSSM_SLHA2.cc:67:3: error: ‘indices’ was not declared in this scope 67 | indices[0] = 3; | ^~~~~~~ make[1]: *** [cudacpp_src.mk:236: Parameters_MSSM_SLHA2.o] Error 1 The non-default HRDCOD=1 however also fails, the first error being ccache g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -DMGONGPU_HARDCODE_PARAM -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o In file included from Parameters_MSSM_SLHA2.cc:8: Parameters_MSSM_SLHA2.h:380:51: error: call to non-‘constexpr’ function ‘mgOnGpu::cxsmpl<FP> mgOnGpu::conj(const mgOnGpu::cxsmpl<FP>&) [with FP = double]’ 380 | constexpr cxsmpl<double> mdl_conjg__yu3x3 = conj( mdl_yu3x3 ); | ~~~~^~~~~~~~~~~~~ In file included from Parameters_MSSM_SLHA2.h:13, from Parameters_MSSM_SLHA2.cc:8:

Now valid code? In the past this failed as "a constexpr function cannot have a nonliteral return type mgOnGpu::cxsmpl"

… particles in HRDCOD=1

…sa (why is this different here??) Replace constexpr cxsmpl<double> mdl_G__exp__2 = ( ( G ) * ( G ) ); by const fptype_sv mdl_G__exp__2 = ( ( G ) * ( G ) );

…es from printParameters to setParameters The HRDCOD=1 build now fails with ccache /usr/local/cuda-12.0/bin/nvcc -O3 -lineinfo -I. -I../../src -I../../../../../tools -I/usr/local/cuda-12.0/include/ -DUSE_NVTX -gencode arch=compute_70,code=compute_70 -gencode arch=compute_70,code=sm_70 -use_fast_math -std=c++17 -ccbin /usr/lib64/ccache/g++ -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -DMGONGPU_HARDCODE_PARAM -Xcompiler -fPIC -c gcheck_sa.cu -o gcheck_sa.o ../../src/Parameters_MSSM_SLHA2.h(310): error: expression must have a constant value ../../src/Parameters_MSSM_SLHA2.h(310): note #2703-D: cannot call non-constexpr function "atan(double) noexcept(true)" /usr/local/cuda-12.0/include/crt/math_functions.h(4137): here ../../src/Parameters_MSSM_SLHA2.h(725): error: expression must have a constant value ../../src/Parameters_MSSM_SLHA2.h(725): note #2703-D: cannot call non-constexpr function "cos(double) noexcept(true)" /usr/local/cuda-12.0/include/crt/math_functions.h(553): here ../../src/Parameters_MSSM_SLHA2.h(726): error: expression must have a constant value ../../src/Parameters_MSSM_SLHA2.h(726): note #2703-D: cannot call non-constexpr function "sin(double) noexcept(true)" /usr/local/cuda-12.0/include/crt/math_functions.h(520): here 3 errors detected in the compilation of "gcheck_sa.cu". The CUDA_HOME=none HRDCOD=0 build also fails with ccache g++ -O3 -std=c++17 -I. -I../../src -I../../../../../tools -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -DMGONGPU_HARDCODE_PARAM -DMGONGPU_HAS_NO_CURAND -fPIC -c CPPProcess.cc -o CPPProcess.o CPPProcess.cc: In function ‘void mg5amcCpu::calculate_wavefunctions(int, const fptype*, const fptype*, mgOnGpu::fptype*, fptype_sv*, int)’: CPPProcess.cc:241:81: error: wrong type argument to unary minus 241 | FFV1_0<W_ACCESS, A_ACCESS, CD_ACCESS>( w_fp[3], w_fp[2], w_fp[4], -COUPs[1], &amp_fp[0] ); | ~~~~~~~^ CPPProcess.cc:251:62: error: wrong type argument to unary minus 251 | FFV1_1<W_ACCESS, CD_ACCESS>( w_fp[2], w_fp[0], -COUPs[1], cIPD[0], cIPD[1], w_fp[4] ); | ~~~~~~~^ CPPProcess.cc:254:81: error: wrong type argument to unary minus 254 | FFV1_0<W_ACCESS, A_ACCESS, CD_ACCESS>( w_fp[3], w_fp[4], w_fp[1], -COUPs[1], &amp_fp[0] ); | ~~~~~~~^ CPPProcess.cc:263:62: error: wrong type argument to unary minus 263 | FFV1_2<W_ACCESS, CD_ACCESS>( w_fp[3], w_fp[0], -COUPs[1], cIPD[0], cIPD[1], w_fp[4] ); | ~~~~~~~^ CPPProcess.cc:266:81: error: wrong type argument to unary minus 266 | FFV1_0<W_ACCESS, A_ACCESS, CD_ACCESS>( w_fp[4], w_fp[2], w_fp[1], -COUPs[1], &amp_fp[0] ); | ~~~~~~~^

This is cmd="import model MSSM_SLHA2; define q = u c d s u~ c~ d~ s~; generate g q > t t~ l- l+ q" This is one of the first processes I tested (inspired from Nathan's tests). But in the end many issues are visible in SM gq_ttq and gq_ttllq, and in SUSY gg_tt I keep this commit for the record, but I will revert it

Revert "[susy2] in CODEGEN, add SUSY process susy_gq_ttllq" This reverts commit 01f5bbc.

…pply old changes on top 'make HRDCOD=0' fails with ccache /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -fPIC -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o In file included from Parameters_MSSM_SLHA2.cc:15: Parameters_MSSM_SLHA2.h:26:2: error: #error This non-SM physics process only supports MGONGPU_HARDCODE_PARAM builds (madgraph5#439): please run "make HRDCOD=1" 26 | #error This non-SM physics process only supports MGONGPU_HARDCODE_PARAM builds (madgraph5#439): please run "make HRDCOD=1" | ^~~~~ 'make HRDCOD=1' fails with ccache /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -DMGONGPU_HARDCODE_PARAM -fPIC -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o In file included from Parameters_MSSM_SLHA2.cc:15: Parameters_MSSM_SLHA2.h:403:53: error: call to non-'constexpr' function 'mgOnGpu::cxsmpl<FP> mgOnGpu::conj(const cxsmpl<FP>&) [with FP = double]' 403 | constexpr cxsmpl<double> mdl_conjg__yu3x3 = conj( mdl_yu3x3 ); | ~~~~^~~~~~~~~~~~~

Previously untracked files, now added with git add: susy_gg_tt.sa/.gitignore susy_gg_tt.sa/COPYING susy_gg_tt.sa/COPYING.LESSER susy_gg_tt.sa/COPYRIGHT susy_gg_tt.sa/SubProcesses/CommonRandomNumberKernel.cc susy_gg_tt.sa/SubProcesses/CommonRandomNumbers.h susy_gg_tt.sa/SubProcesses/CurandRandomNumberKernel.cc susy_gg_tt.sa/SubProcesses/GpuAbstraction.h susy_gg_tt.sa/SubProcesses/GpuRuntime.h susy_gg_tt.sa/SubProcesses/HiprandRandomNumberKernel.cc susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/.gitignore susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CommonRandomNumberKernel.cc susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CommonRandomNumbers.h susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CurandRandomNumberKernel.cc susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/GpuAbstraction.h susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/GpuRuntime.h susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/HiprandRandomNumberKernel.cc susy_gg_tt.sa/mg5.in susy_gg_tt.sa/test/

…alid code? In the past this failed as "a constexpr function cannot have a nonliteral return type mgOnGpu::cxsmpl") Cherry-pick commit 0fd793b from Mon Apr 3 13:55:33 2023 +0200, no conflicts Change the date with 'git commit --amend --date="$(date -R)"' The above comments come from the original commit. Updates on 14-Feb-2024 below. 'make HRDCOD=1' fails with ccache /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -DMGONGPU_HARDCODE_PARAM -fPIC -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o In file included from Parameters_MSSM_SLHA2.cc:15: Parameters_MSSM_SLHA2.h:758:5: error: expected unqualified-id before 'if' 758 | if( mdl_Mneu2 < 0 ) | ^~

… particles in HRDCOD=1 Cherry-pick commit bd5db8f from Mon Apr 3 14:06:20 2023 +0200, fix conflicts with HIP support PR Change the date with 'git commit --amend --date="$(date -R)"' The above comments come from the original commit. Updates on 14-Feb-2024 below. 'make HRDCOD=1' fails with ccache /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -DMGONGPU_HARDCODE_PARAM -fPIC -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o In file included from Parameters_MSSM_SLHA2.cc:15: Parameters_MSSM_SLHA2.h: In function 'const mg5amcCpu::Parameters_MSSM_SLHA2_dependentCouplings::DependentCouplings_sv mg5amcCpu::Parameters_MSSM_SLHA2_dependentCouplings::computeDependentCouplings_fromG(const mg5amcCpu::fptype_sv&)': Parameters_MSSM_SLHA2.h:839:58: error: conversion from 'mg5amcCpu::fptype_sv' {aka '__vector(4) double'} to non-scalar type 'const mgOnGpu::cxsmpl<double>' requested 839 | constexpr cxsmpl<double> mdl_G__exp__2 = ( ( G ) * ( G ) ); | ~~~~~~~~^~~~~~~~~ Parameters_MSSM_SLHA2.h:839:58: error: could not convert '(((mg5amcCpu::fptype_sv)G) * ((mg5amcCpu::fptype_sv)G))' from 'mg5amcCpu::fptype_sv' {aka '__vector(4) double'} to 'const mgOnGpu::cxsmpl<double>' 839 | constexpr cxsmpl<double> mdl_G__exp__2 = ( ( G ) * ( G ) ); | ~~~~~~~~^~~~~~~~~ | | | mg5amcCpu::fptype_sv {aka __vector(4) double}

…sa (why is this different here??) Replace constexpr cxsmpl<double> mdl_G__exp__2 = ( ( G ) * ( G ) ); by const fptype_sv mdl_G__exp__2 = ( ( G ) * ( G ) ); Cherry-pick commit 36dfe05 from Mon Apr 3 14:09:36 2023 +0200, fix a conflict with HIP support PR Change the date with 'git commit --amend --date="$(date -R)"' The above comments come from the original commit. Updates on 14-Feb-2024 below. 'make HRDCOD=1' fails with ccache /usr/local/cuda-12.0/bin/nvcc -Xcompiler -O3 -lineinfo -I. -I../../src -I/usr/local/cuda-12.0/include/ -DUSE_NVTX -gencode arch=compute_70,code=compute_70 -gencode arch=compute_70,code=sm_70 -use_fast_math -std=c++17 -ccbin /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -DMGONGPU_HARDCODE_PARAM -Xcompiler -fPIC -c -x cu Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2_cu.o Parameters_MSSM_SLHA2.h(333): error: expression must have a constant value Parameters_MSSM_SLHA2.h(333): note #2703-D: cannot call non-constexpr function "atan(double) noexcept(true)" /usr/local/cuda-12.0/include/crt/math_functions.h(4137): here Parameters_MSSM_SLHA2.h(748): error: expression must have a constant value Parameters_MSSM_SLHA2.h(748): note #2703-D: cannot call non-constexpr function "cos(double) noexcept(true)" /usr/local/cuda-12.0/include/crt/math_functions.h(553): here Parameters_MSSM_SLHA2.h(749): error: expression must have a constant value Parameters_MSSM_SLHA2.h(749): note #2703-D: cannot call non-constexpr function "sin(double) noexcept(true)" /usr/local/cuda-12.0/include/crt/math_functions.h(520): here 3 errors detected in the compilation of "Parameters_MSSM_SLHA2.cc". Note however that 'CUDA_HOME=none make HRDCOD=1 -j' does succeed! This means that constexpr sin, cos, atan exist in gcc but not in nvcc? Maybe they can be simply hardcoded for nvcc as they are for gcc? Also, check.exe succeeds, while runTest.exe fails because the reference logfile is missing. Note also: cannot cherry-pick commit 586195d from Mon Apr 3 14:13:36 2023 +0200 This would result in an empty commit: I imagine that the upstream mg5anlo has modified this part of the generated code

CUDA_HOME=none make HRDCOD=1 -j CUDACPP_RUNTEST_DUMPEVENTS=1 ./runTest.exe

…ent out "#error only HRDCOD=1 is supported" 'make HRDCOD=0' fails with ccache /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -fPIC -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o In file included from Parameters_MSSM_SLHA2.cc:15: Parameters_MSSM_SLHA2.h: In function 'const mg5amcCpu::Parameters_MSSM_SLHA2_dependentCouplings::DependentCouplings_sv mg5amcCpu::Parameters_MSSM_SLHA2_dependentCouplings::computeDependentCouplings_fromG(const mg5amcCpu::fptype_sv&)': Parameters_MSSM_SLHA2.h:842:33: error: 'mdl_I51x11' was not declared in this scope 842 | out.GC_51 = -( cI * G * mdl_I51x11 ); | ^~~~~~~~~~ Parameters_MSSM_SLHA2.cc: In member function 'void mg5amcCpu::Parameters_MSSM_SLHA2::setIndependentParameters(SLHAReader&)': Parameters_MSSM_SLHA2.cc:80:3: error: 'indices' was not declared in this scope 80 | indices[0] = 3; | ^~~~~~~ make[1]: *** [cudacpp_src.mk:251: Parameters_MSSM_SLHA2.o] Error 1

…SLHA2::getInstance()->mdl_I51x11 visible in HRDCOD=0 builds 'make HRDCOD=0' now fails with ccache /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -fPIC -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o Parameters_MSSM_SLHA2.cc: In member function 'void mg5amcCpu::Parameters_MSSM_SLHA2::setIndependentParameters(SLHAReader&)': Parameters_MSSM_SLHA2.cc:80:3: error: 'indices' was not declared in this scope 80 | indices[0] = 3; | ^~~~~~~

…at was previously commented out 'make HRDCOD=0' now succeeds! however it gives a warning for cuda: ccache /usr/local/cuda-12.0/bin/nvcc -Xcompiler -O3 -lineinfo -I. -I../../src -I/usr/local/cuda-12.0/include/ -DUSE_NVTX -gencode arch=compute_70,code=compute_70 -gencode arch=compute_70,code=sm_70 -use_fast_math -std=c++17 -ccbin /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -Xcompiler -fPIC -c -x cu CPPProcess.cc -o CPPProcess_cu.o ../../src/Parameters_MSSM_SLHA2.h(828): warning #20011-D: calling a __host__ function("mg5amcGpu::Parameters_MSSM_SLHA2::getInstance()") from a __host__ __device__ function("mg5amcGpu::Parameters_MSSM_SLHA2_dependentCouplings::computeDependentCouplings_fromG") is not allowed ../../src/Parameters_MSSM_SLHA2.h(828): remark: The warnings can be suppressed with "-diag-suppress <warning-number>" The warning is probably very relevant because computed MEs are all 0 at runtime! At runtime, gcheck.exe gives 0 MEs. At runtime, runTest.exe fails because it gives 0 MEs. [----------] 1 test from SIGMA_MSSM_SLHA2_GG_TTX_GPU/MadgraphTest [ RUN ] SIGMA_MSSM_SLHA2_GG_TTX_GPU/MadgraphTest.CompareMomentaAndME/0 INFO: Opening reference file ../../test/ref/dump_CPUTest.Sigma_MSSM_SLHA2_gg_ttx.txt MadgraphTest.h:310: Failure The difference between testDriver->getMatrixElement( ievt ) and referenceData[iiter].MEs[ievt] is 2.0052779755903329, which exceeds toleranceMEs * referenceData[iiter].MEs[ievt], where testDriver->getMatrixElement( ievt ) evaluates to 0, referenceData[iiter].MEs[ievt] evaluates to 2.0052779755903329, and toleranceMEs * referenceData[iiter].MEs[ievt] evaluates to 2.0052779755903328e-06. Google Test trace: MadgraphTest.h:289: In comparing event 0 from iteration 0 0 7.500000000000000e+02 0.000000000000000e+00 0.000000000000000e+00 7.500000000000000e+02 ref0 7.500000000000000e+02 0.000000000000000e+00 0.000000000000000e+00 7.500000000000000e+02 1 7.500000000000000e+02 0.000000000000000e+00 0.000000000000000e+00 -7.500000000000000e+02 ref1 7.500000000000000e+02 0.000000000000000e+00 0.000000000000000e+00 -7.500000000000000e+02 2 7.500000000000001e+02 5.849331413473453e+02 -3.138365726669761e+02 -3.490842674916367e+02 ref2 7.500000000000000e+02 5.849331413473452e+02 -3.138365726669761e+02 -3.490842674916366e+02 3 7.500000000000002e+02 -5.849331413473453e+02 3.138365726669762e+02 3.490842674916365e+02 ref3 7.500000000000001e+02 -5.849331413473452e+02 3.138365726669761e+02 3.490842674916364e+02 ME 0.000000000000000e+00 r.ME 2.005277975590333e+00 [ FAILED ] SIGMA_MSSM_SLHA2_GG_TTX_GPU/MadgraphTest.CompareMomentaAndME/0, where GetParam() = 0x2074830 (29 ms) NB: the problems above only concern CUDA. For C++-only builds, HRDCOD=0 builds now fully succeed After `CUDA_HOME=none make HRDCOD=0 -j`, both check.exe and runTest.exe succeed at runtime. This means that C++ builds are now essentially ok for both HRDCOD=0 and HRDCOD=1. The code needs some cleanup however and a backport to code generation, with cross-tests for other processes.

…11 in cIPD on device memory 'make HRDCOD=0' now succeeds and gives no warning. Both CPU and GPU tests succeed in the combined runTest.exe. Both check.exe and gcheck.exe succeed and give the same ME as each other. (And they give the same ME as that from check.exe in 'CUDA_HOME make HRDCOD=1' if common random are used). The other pending problem now is HRDCOD=1 builds in CUDA, where no constexpr sin/cos/atan are available. Note: all tests so far were done with FPTYPE=d. Builds with f and m precision still need to be fully tested and fixed. 'make HRDCOD=0 FPTYPE=f' builds fail: ccache /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_FLOAT -DMGONGPU_FPTYPE2_FLOAT -fPIC -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o In file included from Parameters_MSSM_SLHA2.cc:15: Parameters_MSSM_SLHA2.h: In function 'const mg5amcCpu::Parameters_MSSM_SLHA2_dependentCouplings::DependentCouplings_sv mg5amcCpu::Parameters_MSSM_SLHA2_dependentCouplings::computeDependentCouplings_fromG(const mg5amcCpu::fptype_sv&, const mgOnGpu::fptype*)': Parameters_MSSM_SLHA2.h:861:66: error: the value of 'G' is not usable in a constant expression 861 | constexpr cxsmpl<double> mdl_G__exp__2 = ( ( G ) * ( G ) ); | ^ Parameters_MSSM_SLHA2.h:857:23: note: 'G' was not initialized with a constant expression 857 | const fptype& G = G_sv[i]; | ^

…cuda tests (will be reverted, committed for the record) (This also works but is an overkill, moving a lot of code from Parameters_MSSM_SLHA2.h to CPPProcess.cc. I developed this because I had introduced a silly bug in the previous simpler approach, which I then fixed.)

…e approach This reverts commit 94408b8.

…xpr to const also for FPTYPE=f 'make HRDCOD=0 FPTYPE=f' now fails with: ccache /cvmfs/sft.cern.ch/lcg/releases/gcc/12.1.0-57c96/x86_64-centos9/bin/g++ -O3 -std=c++17 -I. -fPIC -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_FLOAT -DMGONGPU_FPTYPE2_FLOAT -fPIC -c Parameters_MSSM_SLHA2.cc -o Parameters_MSSM_SLHA2.o In file included from Parameters_MSSM_SLHA2.cc:15: Parameters_MSSM_SLHA2.h: In function 'const mg5amcCpu::Parameters_MSSM_SLHA2_dependentCouplings::DependentCouplings_sv mg5amcCpu::Parameters_MSSM_SLHA2_dependentCouplings::computeDependentCouplings_fromG(const mg5amcCpu::fptype_sv&, const mgOnGpu::fptype*)': Parameters_MSSM_SLHA2.h:861:49: error: cannot convert 'mgOnGpu::fptype' {aka 'float'} to 'const mg5amcCpu::fptype_sv' {aka 'const __vector(8) float'} in initialization 861 | const fptype_sv mdl_G__exp__2 = ( ( G ) * ( G ) ); | ~~~~~~~~^~~~~~~~~ | | | mgOnGpu::fptype {aka float}

…for mdl_G__exp__2 in FPTYPE=f 'make HRDCOD=0 FPTYPE=f -j' builds now succeed and the tests succeed too. 'make HRDCOD=0 FPTYPE=m -j' builds now also succeed and the tests succeed too. 'CUDA_HOME=none make HRDCOD=1' builds and tests also succeed with FPTYPE=d,f,m The only pending problem are CUDA HRDCOD=1 builds because of missing constexpr sin/cos/atan (madgraph5#627) This is probably ready to be backported to CODEGEN and eventually merged

…pl::conj function (Question: this is now valid code? In the past this failed as "a constexpr function cannot have a nonliteral return type mgOnGpu::cxsmpl")

…omment out "#error only HRDCOD=1 is supported" and add a comment about non-SM physics processes

…2 instead of fptype[_sv], to ease backport to code generation Note, this implies using cxsml which is a class and not a type, hence a different warning is issues and must be protected against "warning: variable ‘mdl_G__exp__2’ set but not used [-Wunused-but-set-variable]"

…atting

…iously commented out (this is needed for susy_gg_tt)

…__exp__2 instead of cxtype[_sv], because this really is an fptype and not a cxtype

…DEGEN Parameters.h, go back to fptype[_sv] for mdl_G__exp__2 instead of cxtype[_sv], because this really is an fpt>

…still missing: need to add a third parameter to cIPC and related variables) If I regenerate the code and diff to what it should be, this is what I get ______________________________________________________________________________ git diff /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CPPProcess.cc ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ epochX/cudacpp/susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CPPProcess.cc /tmp/git-blob-b4Knci/CPPProcess.cc 0651b576632e9d50c3ed2bf9461bd41309c90821 100644 epochX/cudacpp/susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CPPProcess.cc 0000000000000000000000000000000000000000 100644 79c79 < __device__ const fptype cIPD[3] = { (fptype)Parameters_MSSM_SLHA2::mdl_MT, (fptype)Parameters_MSSM_SLHA2::mdl_WT, (fptype)Parameters_MSSM_SLHA2::mdl_I51x11 }; --- > __device__ const fptype cIPD[2] = { (fptype)Parameters_MSSM_SLHA2::mdl_MT, (fptype)Parameters_MSSM_SLHA2::mdl_WT }; 83c83 < __device__ __constant__ fptype cIPD[3]; --- > __device__ __constant__ fptype cIPD[2]; 86c86 < static fptype cIPD[3]; --- > static fptype cIPD[2]; 505c505 < const fptype tIPD[3] = { (fptype)m_pars->mdl_MT, (fptype)m_pars->mdl_WT, (fptype)m_pars->mdl_I51x11 }; --- > const fptype tIPD[2] = { (fptype)m_pars->mdl_MT, (fptype)m_pars->mdl_WT }; 508c508 < gpuMemcpyToSymbol( cIPD, tIPD, 3 * sizeof( fptype ) ); --- > gpuMemcpyToSymbol( cIPD, tIPD, 2 * sizeof( fptype ) ); 511c511 < memcpy( cIPD, tIPD, 3 * sizeof( fptype ) ); --- > memcpy( cIPD, tIPD, 2 * sizeof( fptype ) ); 514c514 < //for ( i=0; i<3; i++ ) std::cout << std::setprecision(17) << "tIPD[i] = " << tIPD[i] << std::endl; --- > //for ( i=0; i<2; i++ ) std::cout << std::setprecision(17) << "tIPD[i] = " << tIPD[i] << std::endl; ______________________________________________________________________________ git diff /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h /tmp/git-blob-OpNqx8/Parameters_MSSM_SLHA2.h 68d2e3d1bd385d2728480d8e1d1efe8b18563c04 100644 epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h 0000000000000000000000000000000000000000 100644 830,831c830 < //const double mdl_I51x11 = Parameters_MSSM_SLHA2::getInstance()->mdl_I51x11; // fix HRDCOD=0 susy builds < const fptype mdl_I51x11 = cIPD[2]; // fix HRDCOD=0 susy builds --- > // ???

…int out the relevant independent parameters Now the .h file difference is the following. I will move the cIPD handling to an additional parallel mechanism in Parameters.h. ______________________________________________________________________________ git diff /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h /tmp/git-blob-J5f0OS/Parameters_MSSM_SLHA2.h 68d2e3d1bd385d2728480d8e1d1efe8b18563c04 100644 epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h 0000000000000000000000000000000000000000 100644 830,831c830 < //const double mdl_I51x11 = Parameters_MSSM_SLHA2::getInstance()->mdl_I51x11; // fix HRDCOD=0 susy builds < const fptype mdl_I51x11 = cIPD[2]; // fix HRDCOD=0 susy builds --- > const fptype mdl_I51x11;

…ons for HRDCOD=1 cuda builds (madgraph5#627) Builds still fail because assert is not constexpr (I need to extend the function to cover any value and not just 0 to pi/2). Also missing is atan.

…rig functions for HRDCOD=1 cuda builds (madgraph5#627) Extend coverage of sin/cos/tan to any value, and add an atan function. Now the build fails with the following: HRDCOD=1 make ccache /usr/local/cuda-12.0/bin/nvcc -Xcompiler -O3 -lineinfo -I. -I../../src -I/usr/local/cuda-12.0/include/ -DUSE_NVTX -gencode arch=compute_70,code=compute_70 -gencode arch=compute_70,code=sm_70 -use_fast_math -std=c++17 -ccbin /usr/lib64/ccache/g++ -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -DMGONGPU_HARDCODE_PARAM -Xcompiler -fPIC -c -x cu CPPProcess.cc -o CPPProcess_cu.o ../../src/Parameters_MSSM_SLHA2.h(818): error: identifier "mg5amcGpu::Parameters_MSSM_SLHA2::mdl_I51x11" is undefined in device code

…builds (madgraph5#627): it was enought to define mdl_I51x11 as a "__device__" constexpr. NB: Now the build fully succeeds and runTest.exe also succeeds for HRDCOD=1!

… constexpr trig functions - I will clean it up

…ter more intense testing

…in testmisc.cc - still need to clean up Note: I get here a test failure when x is within ~1E-15 of the horizontal or vertical axis (e.g. when it is equivalent to 3pi/2). ./runTest.exe --gtest_filter=*misc Running main() from /data/avalassi/GPU2023/madgraph4gpuBis/test/googletest/googletest/src/gtest_main.cc Note: Google Test filter = *misc [==========] Running 2 tests from 2 test suites. [----------] Global test environment set-up. [----------] 1 test from SIGMA_MSSM_SLHA2_GG_TTX_CPU_MISC [ RUN ] SIGMA_MSSM_SLHA2_GG_TTX_CPU_MISC.testmisc testSinCosTanX: xx= 4.712388980384687897640105802565813064575 x= 4.712388980384687897640105802565813064575 testSinCosTanX: xx= 4.712388980384687898073786671560014838178 x= 4.712388980384687897640105802565813064575 testSinCosTanX: xx= 4.712388980384687897640105802565813064575 x= 4.712388980384687897640105802565813064575 testmisc.cc:390: Failure The difference between std::cos( x ) and constexpr_cos( x ) is 1.7764320877644403e-15, which exceeds std::abs( std::cos( x ) * tolerance ), where std::cos( x ) evaluates to -1.8369701987210297e-16, constexpr_cos( x ) evaluates to -1.9601291076365435e-15, and std::abs( std::cos( x ) * tolerance ) evaluates to 3.6739403974420597e-19. x=4.712388980384687897640105802565813064575, x(0to2Pi)=4.712388980384687897640105802565813064575, istep=55, distance4=1.960020687419294915798673173412680625916e-15 testmisc.cc:390: Failure The difference between std::cos( x ) and constexpr_cos( x ) is 1.7762314254599337e-15, which exceeds std::abs( std::cos( x ) * tolerance ), where std::cos( x ) evaluates to 3.0616169978683831e-16, constexpr_cos( x ) evaluates to -1.4700697256730955e-15, and std::abs( std::cos( x ) * tolerance ) evaluates to 6.123233995736766e-19. x=7.853981633974484566351748071610927581787, x(0to2Pi)=1.570796326794898089326130152798555172922, istep=65, distance4=1.470069725673095462070705252699553966522e-15 testmisc.cc:390: Failure The difference between std::cos( x ) and constexpr_cos( x ) is 1.7765324189166937e-15, which exceeds std::abs( std::cos( x ) * tolerance ), where std::cos( x ) evaluates to -4.2862637970157361e-16, constexpr_cos( x ) evaluates to -2.2051587986182675e-15, and std::abs( std::cos( x ) * tolerance ) evaluates to 8.5725275940314727e-19. x=10.99557428756427412963603273965418338776, x(0to2Pi)=4.712388980384687652610414820841810978891, istep=75, distance4=2.205700899704510220544761978089809417725e-15 testmisc.cc:390: Failure The difference between std::cos( x ) and constexpr_cos( x ) is 1.7761310943076804e-15, which exceeds std::abs( std::cos( x ) * tolerance ), where std::cos( x ) evaluates to 5.5109105961630896e-16, constexpr_cos( x ) evaluates to -1.2250400346913715e-15, and std::abs( std::cos( x ) * tolerance ) evaluates to 1.1021821192326179e-18. x=14.13716694115407079834767500869929790497, x(0to2Pi)=1.570796326794897844296439171074553087237, istep=85, distance4=1.22504003469137145998502091970294713974e-15 [ FAILED ] SIGMA_MSSM_SLHA2_GG_TTX_CPU_MISC.testmisc (0 ms) [----------] 1 test from SIGMA_MSSM_SLHA2_GG_TTX_CPU_MISC (0 ms total) [----------] 1 test from SIGMA_MSSM_SLHA2_GG_TTX_GPU_MISC [ RUN ] SIGMA_MSSM_SLHA2_GG_TTX_GPU_MISC.testmisc testSinCosTanX: xx= 4.712388980384687897640105802565813064575 x= 4.712388980384687897640105802565813064575 testSinCosTanX: xx= 4.712388980384687898073786671560014838178 x= 4.712388980384687897640105802565813064575 testSinCosTanX: xx= 4.712388980384687897640105802565813064575 x= 4.712388980384687897640105802565813064575 [ OK ] SIGMA_MSSM_SLHA2_GG_TTX_GPU_MISC.testmisc (0 ms) [----------] 1 test from SIGMA_MSSM_SLHA2_GG_TTX_GPU_MISC (0 ms total) [----------] Global test environment tear-down [==========] 2 tests from 2 test suites ran. (0 ms total) [ PASSED ] 1 test. [ FAILED ] 1 test, listed below: [ FAILED ] SIGMA_MSSM_SLHA2_GG_TTX_CPU_MISC.testmisc 1 FAILED TEST

…tic cases (horizontal/vertical axis within a few E-15)

…testing

…mplementation of constexpr trig functions for HRDCOD=1 cuda builds (madgraph5#627) Also add it to output.py in CODEGEN

… from susyggtt.sa (madgraph5#627)

…onstexpr_math.h tests from susyggtt.sa (madgraph5#627)

…n the CODEGEN backport

…and pow (now moved to constexpr_math.h as in susyggtt.sa)

… used madgraph5#627 Also add a comment about std::complex (replace is done twice) The differences in generated code are now the following: ______________________________________________________________________________ git diff /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CPPProcess.cc ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ epochX/cudacpp/susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CPPProcess.cc /tmp/git-blob-dFUXLJ/CPPProcess.cc 0651b576632e9d50c3ed2bf9461bd41309c90821 100644 epochX/cudacpp/susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CPPProcess.cc 0000000000000000000000000000000000000000 100644 79c79 < __device__ const fptype cIPD[3] = { (fptype)Parameters_MSSM_SLHA2::mdl_MT, (fptype)Parameters_MSSM_SLHA2::mdl_WT, (fptype)Parameters_MSSM_SLHA2::mdl_I51x11 }; --- > __device__ const fptype cIPD[2] = { (fptype)Parameters_MSSM_SLHA2::mdl_MT, (fptype)Parameters_MSSM_SLHA2::mdl_WT }; 83c83 < __device__ __constant__ fptype cIPD[3]; --- > __device__ __constant__ fptype cIPD[2]; 86c86 < static fptype cIPD[3]; --- > static fptype cIPD[2]; 505c505 < const fptype tIPD[3] = { (fptype)m_pars->mdl_MT, (fptype)m_pars->mdl_WT, (fptype)m_pars->mdl_I51x11 }; --- > const fptype tIPD[2] = { (fptype)m_pars->mdl_MT, (fptype)m_pars->mdl_WT }; 508c508 < gpuMemcpyToSymbol( cIPD, tIPD, 3 * sizeof( fptype ) ); --- > gpuMemcpyToSymbol( cIPD, tIPD, 2 * sizeof( fptype ) ); 511c511 < memcpy( cIPD, tIPD, 3 * sizeof( fptype ) ); --- > memcpy( cIPD, tIPD, 2 * sizeof( fptype ) ); 514c514 < //for ( i=0; i<3; i++ ) std::cout << std::setprecision(17) << "tIPD[i] = " << tIPD[i] << std::endl; --- > //for ( i=0; i<2; i++ ) std::cout << std::setprecision(17) << "tIPD[i] = " << tIPD[i] << std::endl; ______________________________________________________________________________ git diff /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h /tmp/git-blob-fVhC4J/Parameters_MSSM_SLHA2.h 2805ef7d81bd8340b2a77b2ec16df7639165e30c 100644 epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h 0000000000000000000000000000000000000000 100644 23,24d22 < #include "constexpr_math.h" < 295c293 < __device__ constexpr double mdl_I51x11 = 1.; --- > constexpr double mdl_I51x11 = 1.; 370c368 < constexpr cxsmpl<double> mdl_bb = ( ( -mdl_mHd2 + mdl_mHu2 - mdl_MZ__exp__2 * constexpr_cos( 2. * mdl_beta ) ) * constexpr_tan( 2. * mdl_beta ) ) / 2.; --- > constexpr cxsmpl<double> mdl_bb = ( ( -mdl_mHd2 + mdl_mHu2 - mdl_MZ__exp__2 * constexpr_cos( 2. * mdl_beta ) ) * tan( 2. * mdl_beta ) ) / 2.; 801,802c799 < //const double mdl_I51x11 = Parameters_MSSM_SLHA2::getInstance()->mdl_I51x11; // fix HRDCOD=0 susy builds < const fptype mdl_I51x11 = cIPD[2]; // fix HRDCOD=0 susy builds --- > const fptype mdl_I51x11;

…s requires moving earlier on the mechanism to identify parameters such as this one)

The differences are now reduced to this ______________________________________________________________________________ git diff /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CPPProcess.cc ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ epochX/cudacpp/susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CPPProcess.cc /tmp/git-blob-YTajRy/CPPProcess.cc 0651b576632e9d50c3ed2bf9461bd41309c90821 100644 epochX/cudacpp/susy_gg_tt.sa/SubProcesses/P1_Sigma_MSSM_SLHA2_gg_ttx/CPPProcess.cc 0000000000000000000000000000000000000000 100644 79c79 < __device__ const fptype cIPD[3] = { (fptype)Parameters_MSSM_SLHA2::mdl_MT, (fptype)Parameters_MSSM_SLHA2::mdl_WT, (fptype)Parameters_MSSM_SLHA2::mdl_I51x11 }; --- > __device__ const fptype cIPD[2] = { (fptype)Parameters_MSSM_SLHA2::mdl_MT, (fptype)Parameters_MSSM_SLHA2::mdl_WT }; 83c83 < __device__ __constant__ fptype cIPD[3]; --- > __device__ __constant__ fptype cIPD[2]; 86c86 < static fptype cIPD[3]; --- > static fptype cIPD[2]; 505c505 < const fptype tIPD[3] = { (fptype)m_pars->mdl_MT, (fptype)m_pars->mdl_WT, (fptype)m_pars->mdl_I51x11 }; --- > const fptype tIPD[2] = { (fptype)m_pars->mdl_MT, (fptype)m_pars->mdl_WT }; 508c508 < gpuMemcpyToSymbol( cIPD, tIPD, 3 * sizeof( fptype ) ); --- > gpuMemcpyToSymbol( cIPD, tIPD, 2 * sizeof( fptype ) ); 511c511 < memcpy( cIPD, tIPD, 3 * sizeof( fptype ) ); --- > memcpy( cIPD, tIPD, 2 * sizeof( fptype ) ); 514c514 < //for ( i=0; i<3; i++ ) std::cout << std::setprecision(17) << "tIPD[i] = " << tIPD[i] << std::endl; --- > //for ( i=0; i<2; i++ ) std::cout << std::setprecision(17) << "tIPD[i] = " << tIPD[i] << std::endl; ______________________________________________________________________________ git diff /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h /tmp/git-blob-yCgmp8/Parameters_MSSM_SLHA2.h 2805ef7d81bd8340b2a77b2ec16df7639165e30c 100644 epochX/cudacpp/susy_gg_tt.sa/src/Parameters_MSSM_SLHA2.h 0000000000000000000000000000000000000000 100644 801,802c801 < //const double mdl_I51x11 = Parameters_MSSM_SLHA2::getInstance()->mdl_I51x11; // fix HRDCOD=0 susy builds < const fptype mdl_I51x11 = cIPD[2]; // fix HRDCOD=0 susy builds --- > const fptype mdl_I51x11;

…to try and make this easier to code generate (NB: buils and tests are ok for both HARDCOD=0 and =1)

… handling of BSM parameters to try and make this easier to code generate (NB: buils and tests are ok > For HRDCOD=1 this builds and runs ok For HRDCOD=0 this builds but it fails at runtime with invalid device memory access

…SM parameters to try and make this easier to code generate (NB: buils and tests are ok >

… "#define MGONGPUCPP_NBSMINDEPPARAM_GT_0 1" if nBsm>0 This ugly #define was the only way I found to avoid creating arrays[nBsm] in CPPProcess.cc if nBsm is 0 The problem is that nBsm is determined when generating Parameters.h, which happens after CPPProcess.cc has already been generated There may be a better way, but this will do for the moment...

…d another check if nBsm>0

…t from susy_tt.sa)

…cate bsm

… will rerun tests (and try smeft too)

…etting it as a nullptr, and for simplicity add the same code as in susy Now eemumu.sa builds and tests ok for both HRDCOD=0 and =1.

…ed bsmIndepParam as a nullptr, and for simplicity add the same code as in susy Now both susy_gg_tt and sm ee_mumu build and test ok for hrdcod=0/1

…ok now

…builds

…ot defined on Mac

…i constants as M_PIl is not defined on Mac

…o all processes for f in $(git ls-tree --name-only HEAD */src/constexpr_math.h); do \cp susy_gg_tt.sa/src/constexpr_math.h $f; done

…isc.cc, as M_PIl is not defined on Mac (previously fixed constexpr_math.h)

…stants also in testmisc.cc, as M_PIl is not defined on Mac (previously fixed constexpr_math.h)

for f in $(git ls-tree --name-only HEAD */SubProcesses/testmisc.cc); do \cp susy_gg_tt.sa/SubProcesses/testmisc.cc $f; done

…lerances for Mac

for f in $(git ls-tree --name-only HEAD */SubProcesses/testmisc.cc); do \cp susy_gg_tt.sa/SubProcesses/testmisc.cc $f; done

…sons git checkout 04e31d7 tmad/logs_* git checkout c27cb7e tput/logs_*

STARTED AT Thu Feb 29 11:07:30 PM CET 2024 ./tput/teeThroughputX.sh -mix -hrd -makej -eemumu -ggtt -ggttg -ggttgg -gqttq -ggttggg -makeclean ENDED(1) AT Fri Mar 1 02:43:24 AM CET 2024 [Status=0] ./tput/teeThroughputX.sh -flt -hrd -makej -eemumu -ggtt -ggttgg -inlonly -makeclean ENDED(2) AT Fri Mar 1 03:09:38 AM CET 2024 [Status=0] ./tput/teeThroughputX.sh -makej -eemumu -ggtt -ggttg -gqttq -ggttgg -ggttggg -flt -bridge -makeclean ENDED(3) AT Fri Mar 1 03:19:22 AM CET 2024 [Status=0] ./tput/teeThroughputX.sh -eemumu -ggtt -ggttgg -flt -rmbhst ENDED(4) AT Fri Mar 1 03:22:39 AM CET 2024 [Status=0] ./tput/teeThroughputX.sh -eemumu -ggtt -ggttgg -flt -curhst ENDED(5) AT Fri Mar 1 03:25:53 AM CET 2024 [Status=0] ./tput/teeThroughputX.sh -eemumu -ggtt -ggttgg -flt -common ENDED(6) AT Fri Mar 1 03:29:12 AM CET 2024 [Status=0] No errors found in logs

STARTED AT Fri Mar 1 03:29:12 AM CET 2024 ENDED AT Fri Mar 1 07:51:32 AM CET 2024 Status=0 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_m_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0.txt 24 /data/avalassi/GPU2023/madgraph4gpuBis/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd0.txt

…and comment it out (this is not supported in tput/throughputX.sh and in cudacpp anyway)

… not mentioned in usage, and is not supported in cudacpp code anyway)

…ttq) (1) Build tests on login node (~22h) STARTED AT Fri 01 Mar 2024 12:11:58 AM EET ./tput/teeThroughputX.sh -mix -hrd -makej -eemumu -ggtt -ggttg -ggttgg -gqttq -ggttggg -makeclean -makeonly ENDED(1) AT Fri 01 Mar 2024 09:30:18 PM EET [Status=0] ./tput/teeThroughputX.sh -flt -hrd -makej -eemumu -ggtt -ggttgg -inlonly -makeclean -makeonly ENDED(2) AT Fri 01 Mar 2024 09:59:02 PM EET [Status=0] ./tput/teeThroughputX.sh -makej -eemumu -ggtt -ggttg -gqttq -ggttgg -ggttggg -flt -bridge -makeclean -makeonly ENDED(3) AT Fri 01 Mar 2024 10:11:51 PM EET [Status=0] ./tput/teeThroughputX.sh -eemumu -ggtt -ggttgg -flt -rmbhst -makeonly ENDED(4) AT Fri 01 Mar 2024 10:14:10 PM EET [Status=0] SKIP './tput/teeThroughputX.sh -eemumu -ggtt -ggttgg -flt -common -makeonly' ENDED(5) AT Fri 01 Mar 2024 10:14:10 PM EET [Status=0] ./tput/teeThroughputX.sh -eemumu -ggtt -ggttgg -flt -common -makeonly ENDED(6) AT Fri 01 Mar 2024 10:16:26 PM EET [Status=0] (2) Step 2 - run tests on worker nodes (~1h30) ./tput/allTees.sh -hip STARTED AT Sun 03 Mar 2024 01:29:52 PM EET ./tput/teeThroughputX.sh -mix -hrd -makej -eemumu -ggtt -ggttg -ggttgg -gqttq -ggttggg -makeclean ENDED(1) AT Sun 03 Mar 2024 02:22:12 PM EET [Status=2] ./tput/teeThroughputX.sh -flt -hrd -makej -eemumu -ggtt -ggttgg -inlonly -makeclean ENDED(2) AT Sun 03 Mar 2024 02:39:13 PM EET [Status=0] ./tput/teeThroughputX.sh -makej -eemumu -ggtt -ggttg -gqttq -ggttgg -ggttggg -flt -bridge -makeclean ENDED(3) AT Sun 03 Mar 2024 02:58:20 PM EET [Status=2] ./tput/teeThroughputX.sh -eemumu -ggtt -ggttgg -flt -rmbhst ENDED(4) AT Sun 03 Mar 2024 03:02:06 PM EET [Status=0] SKIP './tput/teeThroughputX.sh -eemumu -ggtt -ggttgg -flt -common ' ENDED(5) AT Sun 03 Mar 2024 03:02:06 PM EET [Status=0] ./tput/teeThroughputX.sh -eemumu -ggtt -ggttgg -flt -common ENDED(6) AT Sun 03 Mar 2024 03:05:50 PM EET [Status=0] ./tput/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0_bridge.txt:Backtrace for this error: ./tput/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0_bridge.txt:ERROR! Fortran calculation (F77/CUDA) crashed ./tput/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd1.txt:Backtrace for this error: ./tput/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd1.txt:ERROR! Fortran calculation (F77/CUDA) crashed ./tput/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd1.txt:Backtrace for this error: ./tput/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd1.txt:ERROR! Fortran calculation (F77/CUDA) crashed ./tput/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0.txt:Backtrace for this error: ./tput/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0.txt:ERROR! Fortran calculation (F77/CUDA) crashed ./tput/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd0.txt:Backtrace for this error: ./tput/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd0.txt:ERROR! Fortran calculation (F77/CUDA) crashed ./tput/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd1.txt:Backtrace for this error: ./tput/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd1.txt:ERROR! Fortran calculation (F77/CUDA) crashed ./tput/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0.txt:Backtrace for this error: ./tput/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0.txt:ERROR! Fortran calculation (F77/CUDA) crashed ./tput/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0_bridge.txt:Backtrace for this error: ./tput/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0_bridge.txt:ERROR! Fortran calculation (F77/CUDA) crashed

(1) all tests but ggttggg STARTED AT Sun 03 Mar 2024 04:32:31 PM EET ENDED AT Sun 03 Mar 2024 05:04:29 PM EET (2) ggttggg tests only STARTED AT Sun 03 Mar 2024 04:53:00 PM EET ENDED AT Sun 03 Mar 2024 07:28:33 PM EET Status=0 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_d_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_f_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_eemumu_mad/log_eemumu_mad_m_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_d_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_f_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggttggg_mad/log_ggttggg_mad_m_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_d_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_f_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggttgg_mad/log_ggttgg_mad_m_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_d_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_f_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggttg_mad/log_ggttg_mad_m_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_d_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_f_inl0_hrd0.txt 16 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_ggtt_mad/log_ggtt_mad_m_inl0_hrd0.txt 12 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_d_inl0_hrd0.txt 12 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_f_inl0_hrd0.txt 12 /users/valassia/GPU2023/madgraph4gpu2/epochX/cudacpp/tmad/logs_gqttq_mad/log_gqttq_mad_m_inl0_hrd0.txt

git checkout f4d951c tput/logs_* tmad/logs_*

…no GPU, but a Gold AVX512 CPU (Eventually the outputs of this test will be split between CPU and GPU...)

git checkout f4d951c tput/logs_* tmad/logs_*

./CODEGEN/generateAndCompare.sh susy_gg_tttt -c 'import model MSSM_SLHA2; generate g g > t t~ t t~' CUDACPP_RUNTEST_DUMPEVENTS=1 ./runTest.exe cp ../../test/ref/dump_CPUTest.Sigma_MSSM_SLHA2_gg_ttxttx.txt ../../../CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/test/ref/

./CODEGEN/generateAndCompare.sh susy_gg_gogo -c 'import model MSSM_SLHA2; generate g g > go go' HRDCOD=1 make -j CUDACPP_RUNTEST_DUMPEVENTS=1 ./runTest.exe cp ../../test/ref/dump_CPUTest.Sigma_MSSM_SLHA2_gg_gogo.txt ../../../CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/test/ref/ NB: this process builds only in HRDCOD=1... for HRDCOD=0 the build fails as follows ccache g++ -O3 -std=c++17 -I. -I../../src -Wall -Wshadow -Wextra -ffast-math -fopenmp -march=skylake-avx512 -mprefer-vector-width=256 -DMGONGPU_FPTYPE_DOUBLE -DMGONGPU_FPTYPE2_DOUBLE -fPIC -c CPPProcess.cc -o CPPProcess.o CPPProcess.cc: In member function ‘virtual void mg5amcCpu::CPPProcess::initProc(const string&)’: CPPProcess.cc:554:38: error: ‘class mg5amcCpu::Parameters_MSSM_SLHA2’ has no member named ‘mdl_bsmIndepParam’; did you mean ‘nBsmIndepParam’? 554 | memcpy( bsmIndepParam, m_pars->mdl_bsmIndepParam, Parameters_MSSM_SLHA2::nBsmIndepParam * sizeof( double ) ); | ^~~~~~~~~~~~~~~~~ | nBsmIndepParam

…susy_gq_ttq ./CODEGEN/generateAndCompare.sh susy_gq_ttq -c 'import model MSSM_SLHA2; define q = u c d s u~ c~ d~ s~; generate g q > t t~ q' CUDACPP_RUNTEST_DUMPEVENTS=1 ./runTest.exe cp ../../test/ref/dump_CPUTest.Sigma_MSSM_SLHA2_*txt ../../../CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/test/ref/

…fixing madgraph5#821) into susy2

…_parameters (fix my wrong comment about std::complex string replace - thanks to Olivier for pointing this out)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix susy gg_tt builds and add it to the repo #625

fix susy gg_tt builds and add it to the repo #625

Commits on Apr 3, 2023

Commits on Feb 13, 2024

Commits on Feb 14, 2024

Commits on Feb 27, 2024

Commits on Feb 28, 2024

Commits on Feb 29, 2024

Commits on Mar 1, 2024

Commits on Mar 3, 2024

Commits on Mar 14, 2024