WIP (into master) channelid fixes/reimplementation and merge of master_june24 into master

.github/workflows/c-cpp.yml

@@ -2,9 +2,9 @@ name: C/C++ CI
 
 on:
   push:
-    branches: [ master ]
+    branches: [ master, master_june24 ]
   pull_request:
-    branches: [ master ]
+    branches: [ master, master_june24 ]
 
 jobs:
   debug_builds:

.github/workflows/testsuite_allprocesses.yml

@@ -16,7 +16,7 @@ on:
 
   # Trigger the all-processes workflow for pull requests to master
   pull_request:
-    branches: [ master ]
+    branches: [ master, master_june24 ]
 
   # Trigger the all-processes workflow when new changes to the workflow are pushed
   # (NB: this is now disabled to avoid triggering two jobs when pushing to a branch for which a PR is opened)

.github/workflows/testsuite_oneprocess.sh

@@ -20,9 +20,11 @@ function ECHO() { echo " "; }
 echo "Executing $0 $*"; ECHO
 topdir=$(cd $(dirname $0)/../..; pwd)
 
-# Disable OpenMP in the CI tests
-# To do this, set OMPFLAGS externally to an empty string (#758)
-export OMPFLAGS=
+# Enable OpenMP in the CI tests? (#758)
+###export USEOPENMP=1
+
+# Debug channelid in MatrixElementKernelBase?
+export MG5AMC_CHANNELID_DEBUG=1
 
 # Bypass known issues?
 BYPASS_KNOWN_ISSUES=0 # do not bypass known issues (fail)

.github/workflows/testsuite_oneprocess.yml

@@ -34,8 +34,7 @@ on:
         required: true
         type: choice
         # FIXME? Can the list of supported processes be specified only once in oneprocess.yml or allprocesses.yml?
-        options: [gg_tt.mad, gg_ttg.mad, gg_ttgg.mad, gg_ttggg.mad, ee_mumu.mad, nobm_pp_ttW.mad]
-        ###options: [ee_mumu.mad, gg_tt.mad, gg_ttg.mad, gg_ttgg.mad, gg_ttggg.mad, gg_tt01g.mad, gq_ttq.mad, pp_tt012j.mad]
+        options: [gg_tt.mad, gg_ttg.mad, gg_ttgg.mad, gg_ttggg.mad, ee_mumu.mad, nobm_pp_ttW.mad, susy_gg_tt.mad]
 
 #----------------------------------------------------------------------------------------------------------------------------------
 

epochX/cudacpp/CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/MG5aMC_patches/PROD/patch.P1

@@ -1,163 +1,3 @@
-diff --git b/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/auto_dsig1.f a/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/auto_dsig1.f
-index 4fbb8e6ba..d5accb9fb 100644
---- b/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/auto_dsig1.f
-+++ a/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/auto_dsig1.f
-@@ -484,23 +484,142 @@ C
-       INTEGER VECSIZE_USED
-
-       INTEGER IVEC
--
--
-+      INTEGER IEXT
-+
-+      INTEGER                    ISUM_HEL
-+      LOGICAL                    MULTI_CHANNEL
-+      COMMON/TO_MATRIX/ISUM_HEL, MULTI_CHANNEL
-+
-+      LOGICAL FIRST_CHID
-+      SAVE FIRST_CHID
-+      DATA FIRST_CHID/.TRUE./
-+      
-+#ifdef MG5AMC_MEEXPORTER_CUDACPP
-+      INCLUDE 'coupl.inc' ! for ALL_G
-+      INCLUDE 'fbridge.inc'
-+      INCLUDE 'fbridge_common.inc'
-+      INCLUDE 'genps.inc'
-+      INCLUDE 'run.inc'
-+      DOUBLE PRECISION OUT2(VECSIZE_MEMMAX)
-+      INTEGER SELECTED_HEL2(VECSIZE_MEMMAX)
-+      INTEGER SELECTED_COL2(VECSIZE_MEMMAX)
-+      DOUBLE PRECISION CBYF1
-+      INTEGER*4 NGOODHEL, NTOTHEL
-+
-+      INTEGER*4 NWARNINGS
-+      SAVE NWARNINGS
-+      DATA NWARNINGS/0/
-+      
-+      LOGICAL FIRST
-+      SAVE FIRST
-+      DATA FIRST/.TRUE./
-+      
-+      IF( FBRIDGE_MODE .LE. 0 ) THEN ! (FortranOnly=0 or BothQuiet=-1 or BothDebug=-2)
-+#endif
-+        call counters_smatrix1multi_start( -1, VECSIZE_USED ) ! fortranMEs=-1
- !$OMP PARALLEL
- !$OMP DO
--      DO IVEC=1, VECSIZE_USED
--        CALL SMATRIX1(P_MULTI(0,1,IVEC),
--     &	                         hel_rand(IVEC),
--     &                           col_rand(IVEC),
--     &				 channel,
--     &                           IVEC,
--     &				 out(IVEC),
--     &				 selected_hel(IVEC),
--     &				 selected_col(IVEC)
--     &				 )
--      ENDDO
-+        DO IVEC=1, VECSIZE_USED
-+          CALL SMATRIX1(P_MULTI(0,1,IVEC),
-+     &      hel_rand(IVEC),
-+     &      col_rand(IVEC),
-+     &      channel,
-+     &      IVEC,
-+     &      out(IVEC),
-+     &      selected_hel(IVEC),
-+     &      selected_col(IVEC)
-+     &      )
-+        ENDDO
- !$OMP END DO
- !$OMP END PARALLEL
-+        call counters_smatrix1multi_stop( -1 ) ! fortranMEs=-1
-+#ifdef MG5AMC_MEEXPORTER_CUDACPP
-+      ENDIF
-+
-+      IF( FBRIDGE_MODE .EQ. 1 .OR. FBRIDGE_MODE .LT. 0 ) THEN ! (CppOnly=1 or BothQuiet=-1 or BothDebug=-2)
-+        IF( LIMHEL.NE.0 ) THEN
-+          WRITE(6,*) 'ERROR! The cudacpp bridge only supports LIMHEL=0'
-+          STOP
-+        ENDIF
-+        IF ( FIRST ) THEN ! exclude first pass (helicity filtering) from timers (#461)
-+          call counters_smatrix1multi_start( 1, VECSIZE_USED ) ! cudacppHEL=1
-+          CALL FBRIDGESEQUENCE_NOMULTICHANNEL( FBRIDGE_PBRIDGE, ! multi channel disabled for helicity filtering
-+     &      P_MULTI, ALL_G, HEL_RAND, COL_RAND, OUT2,
-+     &      SELECTED_HEL2, SELECTED_COL2, .TRUE.) ! quit after computing helicities
-+          FIRST = .FALSE.
-+c         ! This is a workaround for https://github.com/oliviermattelaer/mg5amc_test/issues/22 (see PR #486)
-+          IF( FBRIDGE_MODE .EQ. 1 ) THEN ! (CppOnly=1 : SMATRIX1 is not called at all)
-+            CALL RESET_CUMULATIVE_VARIABLE() ! mimic 'avoid bias of the initialization' within SMATRIX1
-+          ENDIF
-+          CALL FBRIDGEGETNGOODHEL(FBRIDGE_PBRIDGE,NGOODHEL,NTOTHEL)
-+          IF( NTOTHEL .NE. NCOMB ) THEN
-+            WRITE(6,*) 'ERROR! Cudacpp/Fortran mismatch',
-+     &        ' in total number of helicities', NTOTHEL, NCOMB
-+            STOP
-+          ENDIF
-+          WRITE (6,*) 'NGOODHEL =', NGOODHEL
-+          WRITE (6,*) 'NCOMB =', NCOMB
-+          call counters_smatrix1multi_stop( 1 ) ! cudacppHEL=1
-+        ENDIF
-+        call counters_smatrix1multi_start( 0, VECSIZE_USED ) ! cudacppMEs=0
-+        IF ( .NOT. MULTI_CHANNEL ) THEN
-+          CALL FBRIDGESEQUENCE_NOMULTICHANNEL( FBRIDGE_PBRIDGE, ! multi channel disabled
-+     &      P_MULTI, ALL_G, HEL_RAND, COL_RAND, OUT2,
-+     &      SELECTED_HEL2, SELECTED_COL2, .FALSE.) ! do not quit after computing helicities
-+        ELSE
-+          IF( SDE_STRAT.NE.1 ) THEN
-+            WRITE(6,*) 'ERROR! The cudacpp bridge requires SDE=1' ! multi channel single-diagram enhancement strategy
-+            STOP
-+          ENDIF
-+          CALL FBRIDGESEQUENCE(FBRIDGE_PBRIDGE, P_MULTI, ALL_G, ! multi channel enabled
-+     &      HEL_RAND, COL_RAND, CHANNEL, OUT2,
-+     &      SELECTED_HEL2, SELECTED_COL2, .FALSE.) ! do not quit after computing helicities
-+        ENDIF
-+        call counters_smatrix1multi_stop( 0 ) ! cudacppMEs=0
-+      ENDIF
-+
-+      IF( FBRIDGE_MODE .LT. 0 ) THEN ! (BothQuiet=-1 or BothDebug=-2)
-+        DO IVEC=1, VECSIZE_USED
-+          CBYF1 = OUT2(IVEC)/OUT(IVEC) - 1
-+          FBRIDGE_NCBYF1 = FBRIDGE_NCBYF1 + 1
-+          FBRIDGE_CBYF1SUM = FBRIDGE_CBYF1SUM + CBYF1
-+          FBRIDGE_CBYF1SUM2 = FBRIDGE_CBYF1SUM2 + CBYF1 * CBYF1
-+          IF( CBYF1 .GT. FBRIDGE_CBYF1MAX ) FBRIDGE_CBYF1MAX = CBYF1
-+          IF( CBYF1 .LT. FBRIDGE_CBYF1MIN ) FBRIDGE_CBYF1MIN = CBYF1
-+          IF( FBRIDGE_MODE .EQ. -2 ) THEN ! (BothDebug=-2)
-+            WRITE (*,'(I4,2E16.8,F23.11,I3,I3,I4,I4)')
-+     &        IVEC, OUT(IVEC), OUT2(IVEC), 1+CBYF1,
-+     &        SELECTED_HEL(IVEC), SELECTED_HEL2(IVEC),
-+     &        SELECTED_COL(IVEC), SELECTED_COL2(IVEC)
-+          ENDIF
-+          IF( ABS(CBYF1).GT.5E-5 .AND. NWARNINGS.LT.20 ) THEN
-+            NWARNINGS = NWARNINGS + 1
-+            WRITE (*,'(A,I4,A,I4,2E16.8,F23.11)')
-+     &        'WARNING! (', NWARNINGS, '/20) Deviation more than 5E-5',
-+     &        IVEC, OUT(IVEC), OUT2(IVEC), 1+CBYF1
-+          ENDIF
-+        END DO
-+      ENDIF
-+
-+      IF( FBRIDGE_MODE .EQ. 1 .OR. FBRIDGE_MODE .LT. 0 ) THEN ! (CppOnly=1 or BothQuiet=-1 or BothDebug=-2)
-+        DO IVEC=1, VECSIZE_USED
-+          OUT(IVEC) = OUT2(IVEC) ! use the cudacpp ME instead of the fortran ME!
-+          SELECTED_HEL(IVEC) = SELECTED_HEL2(IVEC) ! use the cudacpp helicity instead of the fortran helicity!
-+          SELECTED_COL(IVEC) = SELECTED_COL2(IVEC) ! use the cudacpp color instead of the fortran color!
-+        END DO
-+      ENDIF
-+#endif
-+
-+      IF ( FIRST_CHID ) THEN
-+        IF ( MULTI_CHANNEL ) THEN
-+          WRITE (*,*) 'MULTI_CHANNEL = TRUE'
-+        ELSE
-+          WRITE (*,*) 'MULTI_CHANNEL = FALSE'
-+        ENDIF
-+        WRITE (*,*) 'CHANNEL_ID =', CHANNEL
-+        FIRST_CHID = .FALSE.
-+      ENDIF
-+
-       RETURN
-       END
-
 diff --git b/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/driver.f a/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/driver.f
 index 1124a9164..27a6e4674 100644
 --- b/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/driver.f
@@ -286,7 +126,7 @@ index 1124a9164..27a6e4674 100644
        open(unit=lun,file=tempname,status='old',ERR=20)
        fopened=.true.
 diff --git b/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/matrix1.f a/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/matrix1.f
-index e73e654d4..3072054f2 100644
+index 8b73188ef..710fd1297 100644
 --- b/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/matrix1.f
 +++ a/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/matrix1.f
 @@ -72,7 +72,10 @@ C
@@ -301,7 +141,7 @@ index e73e654d4..3072054f2 100644
  C     
  C     This is just to temporarily store the reference grid for
  C      helicity of the DiscreteSampler so as to obtain its number of
-@@ -217,6 +220,17 @@ C     ----------
+@@ -230,6 +233,17 @@ C            update.
            ENDIF
            IF(NTRY(IMIRROR).EQ.MAXTRIES)THEN
              ISHEL(IMIRROR)=MIN(ISUM_HEL,NGOOD(IMIRROR))

epochX/cudacpp/CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/MG5aMC_patches/PROD/patch.common

@@ -13,7 +13,7 @@ index a59181c70..af7e0efbc 100644
        PARAMETER(MAXTRIES=25)
  C     To pass the helicity configuration chosen by the DiscreteSampler to 
 diff --git b/epochX/cudacpp/gg_tt.mad/SubProcesses/makefile a/epochX/cudacpp/gg_tt.mad/SubProcesses/makefile
-index 348c283be..64b33c4f1 100644
+index 348c283be..49e6800ff 100644
 --- b/epochX/cudacpp/gg_tt.mad/SubProcesses/makefile
 +++ a/epochX/cudacpp/gg_tt.mad/SubProcesses/makefile
 @@ -1,6 +1,37 @@
@@ -76,7 +76,7 @@ index 348c283be..64b33c4f1 100644
 
  LIBS = $(LIBDIR)libbias.$(libext) $(LIBDIR)libdhelas.$(libext) $(LIBDIR)libdsample.$(libext) $(LIBDIR)libgeneric.$(libext) $(LIBDIR)libpdf.$(libext) $(LIBDIR)libgammaUPC.$(libext) $(LIBDIR)libmodel.$(libext) $(LIBDIR)libcernlib.$(libext) $(MADLOOP_LIB) $(LOOP_LIBS)
 
-@@ -43,41 +87,145 @@ ifeq ($(strip $(MATRIX_HEL)),)
+@@ -43,41 +87,148 @@ ifeq ($(strip $(MATRIX_HEL)),)
  endif
 
 
@@ -117,7 +117,11 @@ index 348c283be..64b33c4f1 100644
 +else
 +all: $(PROG)_fortran $(CUDACPP_BUILDDIR)/$(PROG)_cpp
 +endif
-+
+
+-$(LIBDIR)libmodel.$(libext): ../../Cards/param_card.dat
+-	cd ../../Source/MODEL; make
++# Disable OpenMP by default: enable OpenMP only if USEOPENMP=1 (#758)
++ifeq ($(USEOPENMP),1)
 +ifneq ($(shell $(CXX) --version | egrep '^Intel'),)
 +override OMPFLAGS = -fopenmp
 +LINKLIBS += -liomp5 # see #578
@@ -130,9 +134,8 @@ index 348c283be..64b33c4f1 100644
 +else
 +override OMPFLAGS = -fopenmp
 +endif
-
--$(LIBDIR)libmodel.$(libext): ../../Cards/param_card.dat
--	cd ../../Source/MODEL; make
++endif
++
 +$(PROG)_fortran: $(PROCESS) $(DSIG) auto_dsig.o $(LIBS) $(MATRIX) counters.o ompnumthreads.o
 +	$(FC) -o $(PROG)_fortran $(PROCESS) $(DSIG) auto_dsig.o $(MATRIX) $(LINKLIBS) $(BIASDEPENDENCIES) $(OMPFLAGS) counters.o ompnumthreads.o $(LDFLAGS)
 
@@ -238,7 +241,7 @@ index 348c283be..64b33c4f1 100644
 
  # Dependencies
 
-@@ -97,5 +245,80 @@ unwgt.o: genps.inc nexternal.inc symswap.inc cluster.inc run.inc message.inc \
+@@ -97,5 +248,80 @@ unwgt.o: genps.inc nexternal.inc symswap.inc cluster.inc run.inc message.inc \
  	 run_config.inc
  initcluster.o: message.inc
 

epochX/cudacpp/CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/launch_plugin.py

@@ -67,6 +67,8 @@ def reset_simd(self, old_value, new_value, name):
         if name == "vector_size" and new_value <= int(old_value):
             # code can handle the new size -> do not recompile
             return
+
+        # ok need to force recompilation of the cpp part
         Sourcedir = pjoin(os.path.dirname(os.path.dirname(self.path)), 'Source')
         subprocess.call(['make', 'cleanall'], cwd=Sourcedir, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
 
@@ -103,7 +105,7 @@ def default_setup(self):
     def write_one_include_file(self, output_dir, incname, output_file=None):
         """write one include file at the time"""
         if incname == "vector.inc":
-            if 'vector_size' not in self.user_set: return
+            if 'vector_size' not in self.user_set and 'wrap_size' not in self.user_set: return
             if output_file is None: vectorinc=pjoin(output_dir,incname)
             else: vectorinc=output_file
             with open(vectorinc+'.new','w') as fileout: