How can I use a new gemm-kernel in tensorflow or other machine learning framework?

google / gemmlowp

Low-precision matrix multiplication

Apache License 2.0

1.77k stars 451 forks source link

How can I use a new gemm-kernel in tensorflow or other machine learning framework? #185

Open knjwhn opened 5 years ago

knjwhn commented 5 years ago

Hello everyone. I've modified some content on the file called kernel_neon.h , I imitate the original kernel(12x8x2) and wrote a new kernel(8x8x8), and I've run the benchmark, It seems that the results are not very different on my arm64 environment. so i want to use in the machine learning framewok such as tensorflow or mxnet ,etc. I've tried so many methods but failed at last. So is this feasible? anyone can help? Thanks a lot.

ArtyZe commented 5 years ago

@knjwhn Hello， so which compile have you used, I met some problem "error: impossible constraint in 'asm'“， when I remove all variables in output, it will be ok. I guess maybe it's the proble of compile