Add workflow to test examples are compilable

.github/utils/CheckExamples/CMakeLists.txt

@@ -0,0 +1,50 @@
+# Checks that all "example codes" in the `Checks/` folders are compilable to
+# avoid committing incorrect examples. It's done here, separately and in bulk,
+# so that no overhead is added by default when submitting a new check. You'll
+# hopefully only have to know this exists if your example fails to compile
+
+cmake_minimum_required(VERSION 3.25)
+project("Test that the check examples compile")
+
+enable_language(Fortran)
+# Build them as libraries, not executables
+set (CMAKE_TRY_COMPILE_TARGET_TYPE STATIC_LIBRARY)
+
+set (BASE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../../../Checks)
+message(STATUS "Looking for files in '${BASE_DIR}'")
+
+# The non-recursive glob only includes files in the first level directory of the
+# Checks which is where the examples are usually located. This way we avoid
+# benchmark programs which may be harder to compile and will be checked anyways
+# by the benchmark running script
+file(GLOB check_files
+  ${BASE_DIR}/*/*.c
+  ${BASE_DIR}/*/*.cpp
+  ${BASE_DIR}/*/*.f
+  ${BASE_DIR}/*/*.F
+  ${BASE_DIR}/*/*.f90
+  ${BASE_DIR}/*/*.F90
+)
+
+# PWR068: Has a module file that has to be compiled in order, and a compilation error on purpose
+# PWD003, PWD006: GFortran catches the data transfer error
+set (excluded_checks_regex "PWR068|PWD003|PWD006")
+
+list(FILTER check_files EXCLUDE REGEX ${excluded_checks_regex})
+
+foreach(check_file ${check_files})
+  message(CHECK_START "Trying to compile ${check_file}")
+  set (extra_flags "-fopenmp -fopenacc")
+  try_compile(
+    TRY_RESULT ${CMAKE_BINARY_DIR}
+    SOURCES ${check_file}
+    CMAKE_FLAGS -DCOMPILE_DEFINITIONS:STRING=${extra_flags}
+    OUTPUT_VARIABLE TRY_RESULT_OUTPUT
+  )
+
+  if (TRY_RESULT)
+    message(CHECK_PASS "Good")
+  else()
+    message(CHECK_FAIL "Bad:\n ${TRY_RESULT_OUTPUT}")
+  endif()
+endforeach()

.github/workflows/compile-check-examples.yml

@@ -0,0 +1,19 @@
+name: Compile Check examples
+
+on:
+  push:
+    paths:
+      - '.github/workflows/**'
+      - '.github/utils/CheckExamples/**'
+      - 'Checks/**'
+
+jobs:
+  # Let us run this job on Ubuntu only, so we can check at the moment that
+  # examples compile at least on GCC && GFortran
+  compile-check-examples:
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Check out repository
+        uses: actions/checkout@v4
+      - name: Check that all check examples compile
+        run: cmake -B build -DCMAKE_C_COMPILER=gcc -DCMAKE_Fortran_COMPILER=gfortran -S .github/utils/CheckExamples/

Checks/PWD003/example-omp.c

@@ -4,7 +4,7 @@
 void example(int *a, int *b, int *sum, int size) {
   // Array bounds should be specified
   #pragma omp target map(to: a, b) map(from: sum)
-  #pragma omp parallel for default(none) shared(a, b, sum)
+  #pragma omp parallel for default(none) shared(a, b, sum, size)
   for (int i = 0; i < size; i++) {
     sum[i] = a[i] + b[i];
   }

Checks/PWD003/example-omp.f90

@@ -10,7 +10,7 @@ subroutine example(a, b, sum, size)
 
   ! Array bounds should be specified
   !$omp target map(to: a, b) map(from: sum)
-  !$omp parallel do default(none) shared(a, b, sum)
+  !$omp parallel do default(none) shared(a, b, sum, size)
   do i = 1, size
     sum(i) = a(i) + b(i)
   end do

Checks/PWD009/README.md

@@ -25,7 +25,7 @@ on non-overlapping positions of the array:
 void example(int m, double *A, double *B, double *C) {
   double temp;
 
-  #pragma omp parallel for default(none) private(i, temp, C) shared(A, B, m)
+  #pragma omp parallel for default(none) private(temp, C) shared(A, B, m)
   for (int i = 0; i < m; i++) {
     temp = A[i] * B[i];
     C[i] = C[i] + temp;
@@ -39,7 +39,7 @@ To fix this, `C` should be moved to a `shared` clause:
 void example(int m, double *A, double *B, double *C) {
   double temp;
 
-  #pragma omp parallel for default(none) private(i, temp) shared(A, B, C, m)
+  #pragma omp parallel for default(none) private(temp) shared(A, B, C, m)
   for (int i = 0; i < m; i++) {
     temp = A[i] * B[i];
     C[i] = C[i] + temp;

Checks/PWD009/example-omp.c

@@ -4,7 +4,7 @@ void example(int m, double *A, double *B, double *C) {
   double temp;
 
   // "C" should be shared
-  #pragma omp parallel for default(none) private(i, temp, C) shared(A, B, m)
+  #pragma omp parallel for default(none) private(temp, C) shared(A, B, m)
   for (int i = 0; i < m; i++) {
     temp = A[i] * B[i];
     C[i] = C[i] + temp;

Checks/PWR006/README.md

@@ -32,7 +32,7 @@ void example() {
   int B[SIZE] = {5, 4, 3, 2, 1};
   int sum[SIZE];
 
-  #pragma omp parallel for shared(sum) firstprivate(A, B) private(i)
+  #pragma omp parallel for shared(sum) firstprivate(A, B)
   for (int i = 0; i < SIZE; i++) {
     sum[i] = A[i] + B[i];
   }
@@ -51,7 +51,7 @@ void example() {
   int B[SIZE] = {5, 4, 3, 2, 1};
   int sum[SIZE];
 
-  #pragma omp parallel for shared(sum, A, B) private(i)
+  #pragma omp parallel for shared(sum, A, B)
   for (int i = 0; i < SIZE; i++) {
     sum[i] = A[i] + B[i];
   }

Checks/PWR006/example.c

@@ -7,7 +7,7 @@ void example() {
   int B[SIZE] = {5, 4, 3, 2, 1};
   int sum[SIZE];
 
-  #pragma omp parallel for shared(sum) firstprivate(A, B) private(i)
+  #pragma omp parallel for shared(sum) firstprivate(A, B)
   for (int i = 0; i < SIZE; i++) {
     sum[i] = A[i] + B[i];
   }

Checks/PWR013/README.md

@@ -27,7 +27,7 @@ used:
 ```c
 void example(double *A, double *B, double *C) {
   #pragma omp target teams distribute parallel for schedule(auto) shared(A, B) \
-          private(i) map(to: A[0:100], B[0:100]) map(tofrom: C[0:100])
+          map(to: A[0:100], B[0:100]) map(tofrom: C[0:100])
   for (int i = 0; i < 100; i++) {
     C[i] += A[i];
   }
@@ -39,7 +39,7 @@ This can be easily corrected by removing references to B from all the clauses:
 ```c
 void example(double *A, double *B, double *C) {
   #pragma omp target teams distribute parallel for schedule(auto) shared(A) \
-          private(i) map(to: A[0:100]) map(tofrom: C[0:100])
+          map(to: A[0:100]) map(tofrom: C[0:100])
   for (int i = 0; i < 100; i++) {
     C[i] += A[i];
   }

Checks/PWR013/example-omp.c

@@ -2,7 +2,7 @@
 
 void example(double *A, double *B, double *C) {
   #pragma omp target teams distribute parallel for schedule(auto) shared(A, B) \
-      private(i) map(to: A[0:100], B[0:100]) map(tofrom: C[0:100])
+      map(to: A[0:100], B[0:100]) map(tofrom: C[0:100])
   for (int i = 0; i < 100; i++) {
     C[i] += A[i];
   }

Checks/PWR015/example-omp.c

@@ -3,7 +3,7 @@
 void example() {
   int A[100], B[100], sum[100];
   #pragma omp target map(to: A[0:100], B[0:100]) map(from: sum[0:100])
-  #pragma omp parallel for private(i)
+  #pragma omp parallel for
   for (int i = 0; i < 50; i++) {
     sum[i] = A[i] + B[i];
   }