From: Pan Xiuli <xiuli....@intel.com>
If last sub group has the max sub group size, it will return 0 instead
of max sub group size. Fix the bug.
V2: Remove useless check for threadn, threadn should always bigger than 1.
Signed-off-by: Pan Xiuli <xiuli....@intel.com>
---
backend/src/libocl/tmpl/ocl_simd.tmpl.cl | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/backend/src/libocl/tmpl/ocl_simd.tmpl.cl
b/backend/src/libocl/tmpl/ocl_simd.tmpl.cl
index 9c09b21..8e22015 100644
--- a/backend/src/libocl/tmpl/ocl_simd.tmpl.cl
+++ b/backend/src/libocl/tmpl/ocl_simd.tmpl.cl
@@ -30,8 +30,8 @@ uint get_sub_group_size(void)
{
uint threadn = get_num_sub_groups();
uint threadid = get_sub_group_id();
- if((threadid == (threadn - 1)) && (threadn > 1))
- return (get_local_size(0)*get_local_size(1)*get_local_size(2)) %
get_max_sub_group_size();
+ if (threadid == (threadn - 1))
+ return (get_local_size(0)*get_local_size(1)*get_local_size(2) -1) %
get_max_sub_group_size() + 1;
else
return get_max_sub_group_size();
}
--
2.7.4
_______________________________________________
Beignet mailing list
Beignet@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/beignet
