portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-11-10 07:55:35 +00:00
Code Releases Activity

Force reqd subgroup size fo SYCL

Browse Source
This commit is contained in:
Peter Boyle 2021-06-22 17:56:10 +00:00
parent 6cd9224dd7
commit 403bff1a47
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
Grid/threads/Accelerator.h
View File

@ -257,11 +257,14 @@ accelerator_inline int acceleratorSIMTlane(int Nsimd) {
unsigned long nt=acceleratorThreads(); \
unsigned long unum1 = num1; \
unsigned long unum2 = num2; \
if(nt < 8)nt=8; \
cl::sycl::range<3> local {nt,1,nsimd}; \
cl::sycl::range<3> global{unum1,unum2,nsimd}; \
cgh.parallel_for<class dslash>( \
cl::sycl::nd_range<3>(global,local), \
[=] (cl::sycl::nd_item<3> item) /*mutable*/ { \
[=] (cl::sycl::nd_item<3> item) /*mutable*/ \
[[intel::reqd_sub_group_size(8)]] \
{ \
auto iter1 = item.get_global_id(0); \
auto iter2 = item.get_global_id(1); \
auto lane = item.get_global_id(2); \