Hello,
I am new to openCL, I got two blogs of openCL :
https://developer.qualcomm.com/blog/matrix-multiply-adreno-gpus-part-1-o...
https://developer.qualcomm.com/blog/matrix-multiply-adreno-gpus-part-2-h...
I tested as the bolg said, but can not rearch the performance .
In my test, 1024 matrix, it costs about 52ms (VS 23ms in blog).
I tested on qualcomm 835.
my codes are copied from the bolg except the global/local range arguments.
Could anyone share me the complete codes of the blogs ?
my mail is [email protected]
Thanks very much!
Hi,
We may release the full source code in the near future. Stay tuned.
Thanks.
Hongqiang Wang
Senior Staff Engineer/Manager
GPU compute, Graphics Research Team
Qualcomm Technologies, Inc.