diff options
| author | Cedric Nugteren <web@cedricnugteren.nl> | 2016-08-20 12:50:31 +0200 |
|---|---|---|
| committer | Cedric Nugteren <web@cedricnugteren.nl> | 2016-08-20 12:50:31 +0200 |
| commit | 6eca53ee235e8105a2822f725beb23d4ab7a7bd6 (patch) | |
| tree | 03c49539feb9b34d149714a49a2f35292cf7f29a /src/kernels/level3/convert_symmetric.opencl | |
| parent | 1ec21421d7dedd715badbc7cf8f555d3b76a022b (diff) | |
| parent | 57f1aa76857cf0566e05b43b9b2a98a3a6139c8b (diff) | |
Merge branch 'master' of https://github.com/dvasschemacq/CLBlast into dvasschemacq-master
Conflicts:
src/kernels/level1/xaxpy.opencl
src/kernels/level2/xgemv.opencl
src/kernels/level2/xgemv_fast.opencl
src/kernels/level2/xger.opencl
src/kernels/level2/xher.opencl
src/kernels/level2/xher2.opencl
src/kernels/level3/xgemm_part2.opencl
Diffstat (limited to 'src/kernels/level3/convert_symmetric.opencl')
| -rw-r--r-- | src/kernels/level3/convert_symmetric.opencl | 28 |
1 files changed, 14 insertions, 14 deletions
diff --git a/src/kernels/level3/convert_symmetric.opencl b/src/kernels/level3/convert_symmetric.opencl index c6ce93ca..8ae53b37 100644 --- a/src/kernels/level3/convert_symmetric.opencl +++ b/src/kernels/level3/convert_symmetric.opencl @@ -20,13 +20,13 @@ R"( // Kernel to populate a squared symmetric matrix, given that the triangle which holds the data is // stored as the lower-triangle of the input matrix. This uses the padding kernel's parameters. -__attribute__((reqd_work_group_size(PAD_DIMX, PAD_DIMY, 1))) -__kernel void SymmLowerToSquared(const int src_dim, - const int src_ld, const int src_offset, - __global const real* restrict src, - const int dest_dim, - const int dest_ld, const int dest_offset, - __global real* dest) { +__kernel __attribute__((reqd_work_group_size(PAD_DIMX, PAD_DIMY, 1))) +void SymmLowerToSquared(const int src_dim, + const int src_ld, const int src_offset, + __global const real* restrict src, + const int dest_dim, + const int dest_ld, const int dest_offset, + __global real* dest) { // Loops over the work per thread in both dimensions #pragma unroll @@ -53,13 +53,13 @@ __kernel void SymmLowerToSquared(const int src_dim, } // Same as above, but now the matrix' data is stored in the upper-triangle -__attribute__((reqd_work_group_size(PAD_DIMX, PAD_DIMY, 1))) -__kernel void SymmUpperToSquared(const int src_dim, - const int src_ld, const int src_offset, - __global const real* restrict src, - const int dest_dim, - const int dest_ld, const int dest_offset, - __global real* dest) { +__kernel __attribute__((reqd_work_group_size(PAD_DIMX, PAD_DIMY, 1))) +void SymmUpperToSquared(const int src_dim, + const int src_ld, const int src_offset, + __global const real* restrict src, + const int dest_dim, + const int dest_ld, const int dest_offset, + __global real* dest) { // Loops over the work per thread in both dimensions #pragma unroll |