bhack

Results 1417 comments of bhack

Intel Beignet spatial convolution OpenCL compile failure

Intel extensions in beignet are in https://cgit.freedesktop.org/beignet/tree/include/CL/cl_intel.h

Intel Beignet spatial convolution OpenCL compile failure

I think that the problem could be on libdrm and kernel version.. What versions of both are you using?

Intel Beignet spatial convolution OpenCL compile failure

Mhh.. Can you add a print of fixed_local_sz[i] inside the loop and before modulo at https://cgit.freedesktop.org/beignet/tree/src/cl_api.c#n3031

Intel Beignet spatial convolution OpenCL compile failure

Have you tried to debug/print that loop?

Intel Beignet spatial convolution OpenCL compile failure

I don't know if this [Beignet Workgroup guide](https://www.freedesktop.org/wiki/Software/Beignet/optimization-guide/) is still valid.

Intel Beignet spatial convolution OpenCL compile failure

It is important to check if `realGroupSize *= fixed_local_sz[i];` it is cumulated correctly. If you have compiled with debug symbols you can check also with gdb break points.

Intel Beignet spatial convolution OpenCL compile failure

@gongzg How can enter in https://cgit.freedesktop.org/beignet/tree/src/cl_api.c#n3036 if local_work_size is not NULL?

Intel Beignet spatial convolution OpenCL compile failure

Ok so probably this message was generated by autotuning code. Where is "Verification was not successful, fallback to basic kernel" in code?

Initial implementation of Inductor FX Graph Remote Cache

Can you expand a little bit more on this?

What versions of Python and TensorFlow are required for using K-FAC?

We had a proposal for a TF2 porting at https://github.com/tensorflow/addons/issues/2388. It is under an TF ecosystem review.