Merge remote-tracking branch 'upstream/master' (including Makefile PRs madgraph5#798 and madgraph5#841) into ewdim6

Fix conflicts:
	epochX/cudacpp/tmad/allTees.sh
	epochX/cudacpp/tput/throughputX.sh

Author: valassi

.github/workflows/c-cpp.yml

@@ -23,17 +23,18 @@ jobs:
       matrix:
         folder: [ epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum , epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg ]
         precision: [ d , f , m ]
+        backend: [ cppnone, cppauto ]
       fail-fast: false
     steps:
     - uses: actions/checkout@v2
     - name: github PR info
       run: date; echo github.event.pull_request.head.sha='${{ github.event.pull_request.head.sha }}'
     - name: make info
-      run: make FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk info
+      run: make BACKEND=${{ matrix.backend }} FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk info
     - name: make
-      run: make FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }}
-    - name: make check
-      run: make FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk check
+      run: make BACKEND=${{ matrix.backend }} FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }}
+    - name: make test
+      run: make BACKEND=${{ matrix.backend }} FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk test
   CPU_MAC:
     runs-on: macos-latest
     env:
@@ -42,27 +43,28 @@ jobs:
       matrix:
         folder: [ epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum, epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg ]
         precision: [ d , f , m ]
+        backend: [ cppnone, cppsse4 ]
       fail-fast: false
     steps:
     - uses: actions/checkout@v2
     - name: github PR info
       run: date; echo github.event.pull_request.head.sha='${{ github.event.pull_request.head.sha }}'
     - name: make info
-      run: make AVX=none OMPFLAGS= FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk info
+      run: make BACKEND=${{ matrix.backend }} OMPFLAGS= FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk info
     - name: make
-      run: make AVX=none OMPFLAGS= FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }}
-    - name: make check
-      run: make AVX=none OMPFLAGS= FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk check
+      run: make BACKEND=${{ matrix.backend }} OMPFLAGS= FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }}
+    - name: make test
+      run: make BACKEND=${{ matrix.backend }} OMPFLAGS= FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk test
   GPU:
     runs-on: self-hosted
     env:
       CUDA_HOME: /usr/local/cuda/
       FC: gfortran
-      REQUIRE_CUDA: 1
     strategy:
       matrix:
         folder: [ epochX/cudacpp/ee_mumu.mad/SubProcesses/P1_epem_mupmum , epochX/cudacpp/gg_ttgg.mad/SubProcesses/P1_gg_ttxgg ]
         precision: [ d , f , m ]
+        backend: [ cppauto, cuda ]
       fail-fast: false
     steps:
     - uses: actions/checkout@v2
@@ -71,8 +73,8 @@ jobs:
     - name: github PR info
       run: date; echo github.event.pull_request.head.sha='${{ github.event.pull_request.head.sha }}'
     - name: make info
-      run: make FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk info
+      run: make BACKEND=${{ matrix.backend }} FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk info
     - name: make
-      run: make FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }}
-    - name: make check
-      run: make FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk check
+      run: make BACKEND=${{ matrix.backend }} FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }}
+    - name: make test
+      run: make BACKEND=${{ matrix.backend }} FPTYPE=${{ matrix.precision }} -C ${{ matrix.folder }} -f cudacpp.mk test

.github/workflows/testsuite_oneprocess.sh

@@ -1,8 +1,8 @@
 #!/bin/bash
-# Copyright (C) 2020-2023 CERN and UCLouvain.
+# Copyright (C) 2020-2024 CERN and UCLouvain.
 # Licensed under the GNU Lesser General Public License (version 3 or later).
 # Created by: A. Valassi (Oct 2023) for the MG5aMC CUDACPP plugin.
-# Further modified by: A. Valassi (2023) for the MG5aMC CUDACPP plugin.
+# Further modified by: A. Valassi (2023-2024) for the MG5aMC CUDACPP plugin.
 
 # Verbose script
 ###set -x
@@ -43,7 +43,6 @@ function codegen() {
     git checkout HEAD ${proc}/CODEGEN*.txt
     if [ "${proc%.mad}" != "${proc}" ]; then
       git checkout HEAD ${proc}/Cards/me5_configuration.txt
-      ###sed -i 's/DEFAULT_F2PY_COMPILER=f2py.*/DEFAULT_F2PY_COMPILER=f2py3/' ${proc}/Source/make_opts
       git checkout HEAD ${proc}/Source/make_opts
     fi
     echo "git diff (start)"
@@ -149,7 +148,8 @@ function build() {
       gtestlibs=1
       make -f cudacpp.mk gtestlibs
     fi
-    make -j avxall
+    # NB: 'make bldall' internally checks if 'which nvcc' and 'which hipcc' succeed before attempting to build cuda and hip
+    make -j bldall
     popd >& /dev/null
   done
 }
@@ -211,25 +211,36 @@ function tput_test() {
     # FIXME3: handle all d/f/m, inl0/1, hrd0/1 etc...
     unamep=$(uname -p)
     unames=$(uname -s)
-    for simd in none sse4 avx2 512y 512z; do
-      # Skip tests for unsupported simd modes as done in tput tests (prevent illegal instruction crashes #791)
+    for backend in cuda hip cppnone cppsse4 cppavx2 cpp512y cpp512z; do
+      # Skip GPU tests for NVidia and AMD unless nvcc and hipcc, respectively, are in PATH
+      if ! nvcc --version &> /dev/null; then
+        if [ "${backend}" == "cuda" ]; then echo; echo "(SKIP ${backend} because nvcc is missing on this node)"; continue; fi
+      elif ! hipcc --version &> /dev/null; then
+        if [ "${backend}" == "hip" ]; then echo; echo "(SKIP ${backend} because hipcc is missing on this node)"; continue; fi
+      fi
+      # Skip C++ tests for unsupported simd modes as done in tput tests (prevent illegal instruction crashes #791)
       if [ "${unamep}" != "x86_64" ]; then
-        if [ "${simd}" == "avx2" ]; then echo; echo "(SKIP ${simd} which is not supported on ${unamep})"; continue; fi
-        if [ "${simd}" == "512y" ]; then echo; echo "(SKIP ${simd} which is not supported on ${unamep})"; continue; fi
-        if [ "${simd}" == "512z" ]; then echo; echo "(SKIP ${simd} which is not supported on ${unamep})"; continue; fi
+        if [ "${backend}" == "cppavx2" ]; then echo; echo "(SKIP ${backend} which is not supported on ${unamep})"; continue; fi
+        if [ "${backend}" == "cpp512y" ]; then echo; echo "(SKIP ${backend} which is not supported on ${unamep})"; continue; fi
+        if [ "${backend}" == "cpp512z" ]; then echo; echo "(SKIP ${backend} which is not supported on ${unamep})"; continue; fi
       elif [ "${unames}" == "Darwin" ]; then
-        if [ "${simd}" == "512y" ]; then echo; echo "(SKIP ${simd} which is not supported on ${unames})"; continue; fi
-        if [ "${simd}" == "512z" ]; then echo; echo "(SKIP ${simd} which is not supported on ${unames})"; continue; fi
+        if [ "${backend}" == "cpp512y" ]; then echo; echo "(SKIP ${backend} which is not supported on ${unames})"; continue; fi
+        if [ "${backend}" == "cpp512z" ]; then echo; echo "(SKIP ${backend} which is not supported on ${unames})"; continue; fi
       elif [ "$(grep -m1 -c avx512vl /proc/cpuinfo)" != "1" ]; then
-        if [ "${simd}" == "512y" ]; then echo; echo "(SKIP ${simd} which is not supported - no avx512vl in /proc/cpuinfo)"; continue; fi
-        if [ "${simd}" == "512z" ]; then echo; echo "(SKIP ${simd} which is not supported - no avx512vl in /proc/cpuinfo)"; continue; fi
+        if [ "${backend}" == "cpp512y" ]; then echo; echo "(SKIP ${backend} which is not supported - no avx512vl in /proc/cpuinfo)"; continue; fi
+        if [ "${backend}" == "cpp512z" ]; then echo; echo "(SKIP ${backend} which is not supported - no avx512vl in /proc/cpuinfo)"; continue; fi
       fi
-      if ls -d build.${simd}* > /dev/null 2>&1; then
-        bdirs="$(ls -d build.${simd}*)"
+      if ls -d build.${backend}* > /dev/null 2>&1; then
+        bdirs="$(ls -d build.${backend}*)"
         for bdir in ${bdirs}; do
           runExe ${bdir}/runTest.exe
-          runExe ${bdir}/check.exe -p 1 32 1
-          runExe ${bdir}/gcheck.exe -p 1 32 1
+          if [ -f ${bdir}/check.exe ]; then
+            runExe ${bdir}/check.exe -p 1 32 1
+          elif [ -f ${bdir}/gcheck.exe ]; then
+            runExe ${bdir}/gcheck.exe -p 1 32 1
+          else
+            echo "ERROR! Neither ${bdir}/check.exe nor ${bdir}/gcheck.exe was found?"; exit 1
+          fi
         done
       fi
     done

.github/workflows/testsuite_oneprocess.yml

@@ -98,7 +98,7 @@ jobs:
       id: cache-restore
       # See https://github.com/actions/cache
       # See https://github.com/actions/cache/blob/main/tips-and-workarounds.md#update-a-cache
-      uses: actions/cache/restore@v3
+      uses: actions/cache/restore@v4
       with:
         path: |
           CCACHE_DIR
@@ -123,7 +123,7 @@ jobs:
     - name: update_cache # update caches after the builds but before the tests (which may fail even if builds succeed)
       id: cache-update
       # See https://github.com/actions/cache
-      uses: actions/cache/save@v3
+      uses: actions/cache/save@v4
       with:
         path: |
           CCACHE_DIR

MG5aMC/mg5amcnlo

@@ -1,5 +1,5 @@
 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 0a3dfa449..8f4aaecd6 100644
+index 4fbb8e6ba..f9e2335de 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 +484,140 @@ C
@@ -284,7 +284,7 @@ index 71fbf2b25..0f1d199fc 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 817af778b..0c2ce6ec4 100644
+index e73e654d4..27fbe7302 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