[OpenMP] [test] Fix target_thread_limit.cpp to not assume 4 or more cores

[mstorsjo]

Summary

Previously, the test ran a section with

#pragma omp target thread_limit(4)

and expected it to execute exactly 4 times, even though it would
in practice execute min(cores, 4) times.

Increment a counter and check that it executed 1-4 times.

[jdoerfert]

This does not impact the test in the way you describe. set_num_threads, and thread_limit, are both upper bounds only. We can still pick 1 thread and it is conformant.
Adding the call is fine, but we should instead increment a number, not check for printed lines. The final result should be 1 <= result <= 4.

[mstorsjo]

Check the number of runs with a counter and FileCheck pattern, instead of expecting a certain number of printout lines.

[protze.joachim]

openmp/runtime/test/target/target_thread_limit.cpp

23–24

Fix the data race on count++

73–76

same here

[mstorsjo]

openmp/runtime/test/target/target_thread_limit.cpp

23–24

Thanks! I meant to ask about whether this needed some extra atomic accesses or similar, but forgot to ask when posting the patch.

[mstorsjo]

Fixed the data races

openmp/runtime/test/target/target_thread_limit.cpp

int main(void) {
 
   int tl = 4;
   printf("\nmain: thread_limit = %d", omp_get_thread_limit());
   // OMP51: main: thread_limit = {{[0-9]+}}
 
 #pragma omp target thread_limit(tl)
   {
     printf("\ntarget: thread_limit = %d", omp_get_thread_limit());
+    int count = 0;
 // OMP51: target: thread_limit = 4
 // check whether thread_limit is honoured
-#pragma omp parallel
-    { printf("\ntarget: parallel"); }
-// OMP51: target: parallel
-// OMP51: target: parallel
-// OMP51: target: parallel
-// OMP51: target: parallel
-// OMP51-NOT: target: parallel
+#pragma omp parallel reduction(+:count)
+    { count++; }
+    printf("\ntarget: parallel: count = %d", count);
+// OMP51: target: parallel: count = {{(1|2|3|4)$}}
 
 // check whether num_threads is honoured
 #pragma omp parallel num_threads(2)
     { printf("\ntarget: parallel num_threads(2)"); }
 // OMP51: target: parallel num_threads(2)
 // OMP51: target: parallel num_threads(2)
 // OMP51-NOT: target: parallel num_threads(2)
 
#pragma omp parallel
     printf("\ntarget: thread_limit = %d", omp_get_thread_limit());
     // OMP51: target: thread_limit = 4
   }
 
 // checking consecutive target regions with different thread_limits
 #pragma omp target thread_limit(3)
   {
     printf("\nsecond target: thread_limit = %d", omp_get_thread_limit());
+    int count = 0;
 // OMP51: second target: thread_limit = 3
-#pragma omp parallel
-    { printf("\nsecond target: parallel"); }
-    // OMP51: second target: parallel
-    // OMP51: second target: parallel
-    // OMP51: second target: parallel
-    // OMP51-NOT: second target: parallel
+#pragma omp parallel reduction(+:count)
+    { count++; }
+    printf("\nsecond target: parallel: count = %d", count);
+    // OMP51: second target: parallel: count = {{(1|2|3)$}}
   }
 
   // confirm that thread_limit's effects are limited to target region
   printf("\nmain: thread_limit = %d", omp_get_thread_limit());
   // OMP51: main: thread_limit = {{[0-9]+}}
 #pragma omp parallel num_threads(10)
   { printf("\nmain: parallel num_threads(10)"); }
   // OMP51: main: parallel num_threads(10)