intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2025-12-25 13:33:02 +08:00
Code Issues Packages Projects Releases Wiki Activity

Apply (2/4)x4x1 layout when generating local ids for kernel with images only

Browse Source 
- For SIMD8 apply 2x4x1 layout
- For SIMD16/SIMD32 apply 4x4x1 layout

Change-Id: I31bceb49387011c66da5f96ad2a71125b96d4cda
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
This commit is contained in:
Mateusz Jablonski
2018-08-08 12:49:36 +02:00
committed by sys_ocldev
parent cf971158c6
commit 47f3dad619
9 changed files with 289 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
runtime/command_queue/local_id_gen.h
View File

@@ -79,6 +79,8 @@ void generateLocalIDsSimd(void *b, const std::array<uint16_t, 3> &localWorkgroup
const std::array<uint8_t, 3> &dimensionsOrder);
void generateLocalIDs(void *buffer, uint16_t simd, const std::array<uint16_t, 3> &localWorkgroupSize,
const std::array<uint8_t, 3> &dimensionsOrder);
const std::array<uint8_t, 3> &dimensionsOrder, bool hasKernelOnlyImages);
void generateLocalIDsWith4x4Layout(void *b, const std::array<uint16_t, 3> &localWorkgroupSize, uint16_t simd);
bool isCompatibleWith4x4Layout(const std::array<uint16_t, 3> &localWorkgroupSize, const std::array<uint8_t, 3> &dimensionsOrder, uint16_t simd);
} // namespace OCLRT