The libgomp.oacc-fortran/privatized-ref-1.f95 also has the same issue, and has
been fixed in the same way.
Kwok
On 16/09/2020 11:33 am, Kwok Cheung Yeung wrote:
Hello
The libgomp.oacc-c++/privatized-ref-[23].C testcases request 64 workers in a
parallel section, but Nvidia only supports a maximum of 32 workers, and GCN a
maximum of 16. The worker numbers are overridden by the compiler with a warning
message printed, which causes test failures on Nvidia. On GCN, the warning
message is not printed by default (enabled by -foffload=-Wopenacc-dims).
This patch fixes this by requesting 16 workers, which is acceptable on both
Nvidia and AMD GCN. Committed on the OG10 branch as obvious.
Kwok
commit abe747ac4e1e7c6a14b0877b88a7555b11e6bcb1
Author: Kwok Cheung Yeung <k...@codesourcery.com>
Date: Thu Sep 17 03:16:04 2020 -0700
Fix warning messages in libgomp.oacc-fortran/privatized-ref-1.f95 testcase
2020-09-17 Kwok Cheung Yeung <k...@codesourcery.com>
libgomp/
* testsuite/libgomp.oacc-fortran/privatized-ref-1.f95 (workers,
vectors):
Reduce number of workers to 16.
diff --git a/libgomp/ChangeLog.omp b/libgomp/ChangeLog.omp
index 890a4e2..fbe8bb4 100644
--- a/libgomp/ChangeLog.omp
+++ b/libgomp/ChangeLog.omp
@@ -1,3 +1,8 @@
+2020-09-17 Kwok Cheung Yeung <k...@codesourcery.com>
+
+ * testsuite/libgomp.oacc-fortran/privatized-ref-1.f95 (workers,
vectors):
+ Reduce number of workers to 16.
+
2020-09-16 Kwok Cheung Yeung <k...@codesourcery.com>
* testsuite/libgomp.oacc-fortran/privatized-ref-2.f90: XFAIL on nvptx.
diff --git a/libgomp/testsuite/libgomp.oacc-fortran/privatized-ref-1.f95
b/libgomp/testsuite/libgomp.oacc-fortran/privatized-ref-1.f95
index f16f69c..e4b8520 100644
--- a/libgomp/testsuite/libgomp.oacc-fortran/privatized-ref-1.f95
+++ b/libgomp/testsuite/libgomp.oacc-fortran/privatized-ref-1.f95
@@ -34,7 +34,7 @@ contains
integer :: i, j
real, intent(out) :: res(:)
- !$acc parallel copyout(res) num_gangs(64) num_workers(64)
+ !$acc parallel copyout(res) num_gangs(64) num_workers(16)
!$acc loop gang
do i=0,255
@@ -54,7 +54,7 @@ contains
integer :: i, j
real, intent(out) :: res(:)
- !$acc parallel copyout(res) num_gangs(64) num_workers(64)
+ !$acc parallel copyout(res) num_gangs(64) num_workers(16)
!$acc loop gang worker
do i=0,255
