resolve merge conflict

.gitignore

@@ -18,6 +18,9 @@ octave-workspace                # Octave crashes
 *.swp                           # Vim
 */*.swp
 */*/*.swp
+*.swo                           # Vim
+*/*.swo
+*/*/*.swo
 *.dSYM                          # Mac
 */*.dSYM
 */*/*.dSYM
@@ -26,6 +29,11 @@ octave-workspace                # Octave crashes
 */*/*.optrpt
 *__genmod.*                     # Intel Fortran compiler
 */*__genmod.mod
+*.patch
+*/*.patch
+*.dbg                           # Flang
+*/*.dbg
+*/*/*.dbg
 
 common/make.defs
 scripts/small/runfgmpi
@@ -43,17 +51,10 @@ func.c                          # PRK C89 stencil generated code
 *.output                        # ALCF Cobalt scheduler
 *.error                         # ALCF Cobalt scheduler
 
-FORTRAN/Stencil/stencil
-FORTRAN/Stencil/stencil-coarray
-FORTRAN/Stencil/stencil-omp
-FORTRAN/Stencil/stencil-pretty
-FORTRAN/Synch_p2p/p2p
-FORTRAN/Synch_p2p/p2p-coarray
-FORTRAN/Synch_p2p/p2p-omp
-FORTRAN/Transpose/transpose
-FORTRAN/Transpose/transpose-coarray
-FORTRAN/Transpose/transpose-omp
-FORTRAN/Transpose/transpose-pretty
+*.log
+*.log2
+*.log3
+
 MPI1/AMR/amr
 MPI1/Branch/branch
 MPI1/DGEMM/dgemm
@@ -99,11 +100,26 @@ SERIAL/Sparse/sparse
 SERIAL/Stencil/stencil
 SERIAL/Synch_p2p/p2p
 SERIAL/Transpose/transpose
+C1z/nstream
+C1z/nstream-openmp
+C1z/nstream-target
+C1z/nstream-taskloop
+C1z/nstream-mpi
+C1z/nstream-memkind
+C1z/nstream-memkind-openmp
+C1z/nstream-mmap
+C1z/nstream-mmap-openmp
 C1z/p2p
+C1z/p2p-avx
+C1z/p2p-sse
 C1z/p2p-innerloop
 C1z/p2p-innerloop-openmp
+C1z/p2p-hyperplane
+C1z/p2p-hyperplane-openmp
 C1z/p2p-tasks-openmp
 C1z/p2p-simd-openmp
+C1z/p2p-avx
+C1z/p2p-sse
 C1z/stencil
 C1z/stencil-cilk
 C1z/stencil-openmp
@@ -116,30 +132,54 @@ C1z/transpose-openmp
 C1z/transpose-target
 C1z/transpose-taskloop
 C1z/transpose-ispc
+Cxx11/boost
+Cxx11/compute
+Cxx11/triSYCL
+Cxx11/occa
+Cxx11/pstl
+Cxx11/parallelstl
+Cxx11/range-v3
 Cxx11/dgemm-vector
+Cxx11/dgemm-cblas
+Cxx11/dgemm-cublas
 Cxx11/p2p-openmp-target
 Cxx11/p2p-tasks-openmp
+Cxx11/p2p-tasks-tbb
 Cxx11/p2p-vector
 Cxx11/p2p-vector-doacross-openmp
 Cxx11/p2p-vector-raja
 Cxx11/p2p-vector-tbb
 Cxx11/p2p-innerloop-openmp
 Cxx11/p2p-doacross-vector-openmp
+Cxx11/p2p-doacross-openmp
 Cxx11/p2p-innerloop-opencl
 Cxx11/p2p-innerloop-vector
-Cxx11/p2p-innerloop-vector-openmp
+Cxx11/p2p-hyperplane-vector
+Cxx11/p2p-hyperplane-openmp
+Cxx11/p2p-hyperplane-vector-openmp
 Cxx11/p2p-innerloop-vector-tbb
+Cxx11/p2p-hyperplane-vector-stl
+Cxx11/p2p-hyperplane-vector-pstl
+Cxx11/p2p-hyperplane-vector-tbb
+Cxx11/p2p-hyperplane-sycl
 Cxx11/nstream-kokkos
 Cxx11/nstream-opencl
 Cxx11/nstream-valarray
 Cxx11/nstream-vector
 Cxx11/nstream-vector-openmp
 Cxx11/nstream-vector-pstl
 Cxx11/nstream-vector-raja
+Cxx11/nstream-raja
 Cxx11/nstream-vector-rangefor
 Cxx11/nstream-vector-stl
 Cxx11/nstream-vector-taskloop
 Cxx11/nstream-vector-tbb
+Cxx11/nstream-valarray-boost-compute
+Cxx11/nstream-vector-boost-compute
+Cxx11/nstream-cublas
+Cxx11/nstream-cuda
+Cxx11/nstream-openmp-target
+Cxx11/nstream-sycl
 Cxx11/sparse-vector
 Cxx11/stencil-opencl
 Cxx11/stencil-openmp-target
@@ -149,11 +189,17 @@ Cxx11/stencil-vector-cilk
 Cxx11/stencil-vector-stl
 Cxx11/stencil-vector-pstl
 Cxx11/stencil-vector-raja
+Cxx11/stencil-openmp
+Cxx11/stencil-raja
 Cxx11/stencil-vector-rangefor
 Cxx11/stencil-vector-tbb
 Cxx11/stencil-vector-taskloop
 Cxx11/stencil-kokkos
+Cxx11/stencil-cuda
+Cxx11/stencil-sycl
+Cxx11/transpose-openmp
 Cxx11/transpose-opencl
+Cxx11/transpose-sycl
 Cxx11/transpose-openmp-target
 Cxx11/transpose-valarray
 Cxx11/transpose-vector
@@ -164,10 +210,17 @@ Cxx11/transpose-vector-cilk
 Cxx11/transpose-vector-stl
 Cxx11/transpose-vector-pstl
 Cxx11/transpose-vector-raja
+Cxx11/transpose-raja
 Cxx11/transpose-vector-rangefor
 Cxx11/transpose-vector-tbb
 Cxx11/transpose-vector-taskloop
+Cxx11/transpose-vector-async
+Cxx11/transpose-vector-thread
 Cxx11/transpose-kokkos
+Cxx11/transpose-device-thrust
+Cxx11/transpose-host-thrust
+Cxx11/transpose-cublas
+Cxx11/transpose-cuda
 Cxx11/grid1.cl
 Cxx11/grid2.cl
 Cxx11/grid3.cl
@@ -186,6 +239,25 @@ Cxx11/star6.cl
 Cxx11/star7.cl
 Cxx11/star8.cl
 Cxx11/star9.cl
+Cxx11/star10.cl
+Cxx11/hipSYCL
+Cxx11/cpp-proposals-pub
+Cxx11/ornl-mdspan
+Cxx11/boost.tgz
+Cxx11/boost.tbz
+Cxx11/OpenCL-CLHPP
+Cxx11/GSL
+FORTRAN/dgemm-taskloop-openmp
+FORTRAN/dgemm-pretty
+FORTRAN/dgemm-openmp
+FORTRAN/dgemm-openmp-target
+FORTRAN/dgemm
+FORTRAN/nstream
+FORTRAN/nstream-openmp
+FORTRAN/nstream-pretty
+FORTRAN/nstream-taskloop-openmp
+FORTRAN/nstream-openmp-target
+FORTRAN/nstream-ornlacc
 FORTRAN/p2p
 FORTRAN/p2p-innerloop
 FORTRAN/p2p-coarray
@@ -194,19 +266,25 @@ FORTRAN/p2p-tasks-openmp
 FORTRAN/p2p-doacross-openmp
 FORTRAN/p2p-innerloop-openmp
 FORTRAN/p2p-datapar-openmp
+FORTRAN/p2p-innerloop-ornlacc
+FORTRAN/p2p-ornlacc
 FORTRAN/stencil
 FORTRAN/stencil-coarray
 FORTRAN/stencil-openmp
 FORTRAN/stencil-openmp-target
 FORTRAN/stencil-pretty
+FORTRAN/stencil-ornlacc
 FORTRAN/stencil-taskloop-openmp
+FORTRAN/stencil-ornlacc
 FORTRAN/transpose
 FORTRAN/transpose-coarray
 FORTRAN/transpose-openmp
 FORTRAN/transpose-openmp-target
 FORTRAN/transpose-pretty
+FORTRAN/transpose-ornlacc
 FORTRAN/transpose-taskloop-openmp
 FORTRAN/transpose-tasks-openmp
-RUST/p2p
-RUST/stencil
-RUST/transpose
+FORTRAN/transpose-ornlacc
+RUST/p2p/Cargo.lock
+RUST/stencil/Cargo.lock
+RUST/transpose/Cargo.lock

.travis.yml

@@ -134,6 +134,10 @@ matrix:
   # Mac issue with thread_t (see https://github.com/humairakamal/fgmpi/pull/1)
   - os: osx
     env: PRK_TARGET=allfgmpi
+  # SOS@OFI has not worked on MacOS in a while :-(
+  - os: osx
+    compiler: clang
+    env: PRK_TARGET=allshmem
   allow_failures:
   # Travis trusty breaks this
   - os: linux
@@ -144,11 +148,10 @@ matrix:
   - os: linux
     env: PRK_TARGET=allfgmpi
   - os: linux
+    compiler: clang
     env: PRK_TARGET=allmpi
   - os: linux
     env: PRK_TARGET=allcharm++
-  # Sadly, Python is XFAIL because Travis CI's Python 3.4.3 can't find cannot process_time.
-  - env: PRK_TARGET=allpython
 addons:
   apt:
     sources:

AMPI/Stencil/stencil.c

@@ -88,7 +88,7 @@ HISTORY: - Written by Rob Van der Wijngaart, November 2006.
 #endif
 
 /* define shorthand for indexing multi-dimensional arrays with offsets           */
-#define INDEXIN(i,j)  (i+RADIUS+(j+RADIUS)*(width+2*RADIUS))
+#define INDEXIN(i,j)  (i+RADIUS+(long)(j+RADIUS)*(long)(width+2*RADIUS))
 /* need to add offset of RADIUS to j to account for ghost points                 */
 #define IN(i,j)       in[INDEXIN(i-istart,j-jstart)]
 #define INDEXOUT(i,j) (i+(j)*(width))

C1z/Makefile

@@ -10,6 +10,10 @@ ifdef VERBOSE
   CFLAGS += -DVERBOSE
 endif
 
+ifdef PRK_USE_MMAP
+  CFLAGS += -DPRK_USE_MMAP
+endif
+
 ifeq ($(findstring musl,$(CC)),musl)
   CFLAGS += -DUSE_C11_THREADS
 endif
@@ -22,7 +26,7 @@ ORNLACCFLAGS = $(ORNLACCFLAG)
 CILKFLAGS = $(CILKFLAG)
 ISPCFLAGS = $(ISPCFLAG)
 
-.PHONY: all clean serial thread openmp target taskloop cilk ispc
+.PHONY: all clean serial thread openmp target taskloop ispc # cilk
 
 EXTRA=
 ifeq ($(shell uname -s),Darwin)
@@ -38,26 +42,60 @@ endif
 ifneq ($(CILKFLAG),)
   EXTRA += cilk
 endif
+ifeq ($(findstring xlc,$(CC)),xlc)
+  EXTRA = target
+  CFLAGS += -DXLC
+endif
 
 all: serial thread openmp taskloop $(EXTRA)
 
-serial: p2p p2p-innerloop stencil transpose
+serial: nstream p2p p2p-hyperplane stencil transpose
 
 thread: transpose-thread
 
-openmp: p2p-simd-openmp p2p-tasks-openmp p2p-innerloop-openmp stencil-openmp transpose-openmp
+openmp: nstream-openmp p2p-simd-openmp p2p-tasks-openmp stencil-openmp transpose-openmp p2p-hyperplane-openmp
+
+mpi: nstream-mpi
+
+memkind: nstream-memkind nstream-memkind-openmp
 
-target: stencil-target transpose-target
+mmap: nstream-mmap nstream-mmap-openmp
 
-taskloop: stencil-taskloop transpose-taskloop
+target: nstream-target stencil-target transpose-target
+
+taskloop: nstream-taskloop stencil-taskloop transpose-taskloop
 
 cilk: stencil-cilk transpose-cilk
 
 ispc: transpose-ispc
 
-p2p-innerloop: p2p-innerloop-openmp.c prk_util.h
+p2p-hyperplane: p2p-hyperplane-openmp.c prk_util.h
+	$(CC) $(CFLAGS) $< $(EXTRA_CLIBS) -o $@
+
+nstream: nstream-openmp.c prk_util.h
+	$(CC) $(CFLAGS) $< $(EXTRA_CLIBS) -o $@
+
+stencil: stencil-openmp.c prk_util.h
+	$(CC) $(CFLAGS) $< $(EXTRA_CLIBS) -o $@
+
+transpose: transpose-openmp.c prk_util.h
 	$(CC) $(CFLAGS) $< $(EXTRA_CLIBS) -o $@
 
+%-mpi: %-mpi.c prk_util.h
+	$(MPICC) $(CFLAGS) $< $(EXTRA_CLIBS) -o $@
+
+%-memkind: %-memkind.c prk_util.h
+	$(CC) $(CFLAGS) $< $(EXTRA_CLIBS) $(MEMKINDFLAGS) -o $@
+
+%-memkind-openmp: %-memkind.c prk_util.h
+	$(CC) $(CFLAGS) $(OMPFLAGS) $< $(EXTRA_CLIBS) $(MEMKINDFLAGS) -o $@
+
+%-mmap: %-mmap.c prk_util.h
+	$(CC) $(CFLAGS) $< $(EXTRA_CLIBS) -o $@
+
+%-mmap-openmp: %-mmap.c prk_util.h
+	$(CC) $(CFLAGS) $(OMPFLAGS) $< $(EXTRA_CLIBS) -o $@
+
 %-target: %-target.c prk_util.h
 	$(CC) $(CFLAGS) $< $(OMPFLAGS) $(TARGETFLAGS) $(EXTRA_CLIBS) -o $@
 
@@ -92,13 +130,16 @@ clean:
 	-rm -f *.optrpt
 	-rm -f *.dwarf
 	-rm -rf *.dSYM # Mac
-	-rm -f p2p p2p-innerloop stencil transpose
+	-rm -f nstream p2p p2p-hyperplane stencil transpose
 	-rm -f *-openmp
+	-rm -f *-mpi
 	-rm -f *-target
 	-rm -f *-taskloop
 	-rm -f *-cilk
 	-rm -f *-thread
 	-rm -f *-ispc
+	-rm -f nstream-mmap nstream-memkind
+	-rm -f nstream-mmap-openmp nstream-memkind-openmp
 
 cleancl:
 	-rm -f star[123456789].cl

C1z/generate-c-stencil.py

@@ -76,13 +76,13 @@ def instance(src,model,pattern,r):
 def main():
     for model in ['seq','openmp','target','cilk','taskloop']:
       src = open('stencil_'+model+'.h','w')
-      if (model=='target'):
-          src.write('OMP( declare target )\n')
+      #if (model=='target'):
+      #    src.write('OMP( declare target )\n')
       for pattern in ['star','grid']:
         for r in range(1,10):
           instance(src,model,pattern,r)
-      if (model=='target'):
-          src.write('OMP( end declare target )\n')
+      #if (model=='target'):
+      #    src.write('OMP( end declare target )\n')
       src.close()
 
 if __name__ == '__main__':