Plan for NHWC optimization for NEON

ermig1979 / Simd

C++ image processing and machine learning library with using of SIMD: SSE, AVX, AVX-512, AMX for x86/x64, NEON for ARM.

http://ermig1979.github.io/Simd

MIT License

2.06k stars 413 forks source link

Plan for NHWC optimization for NEON #86

Closed biaoxiaoduan closed 5 years ago

biaoxiaoduan commented 5 years ago

I noticed the NHWC implementation of convolution, and it's extremely fast on AVX2. But it seems Neon version is not implemented, I didn't find source file named SimdNeonConvolution.cpp. Do you have any plan for that?

ermig1979 commented 5 years ago

Yes. I have such plans. Now I'm waiting for ARM dev board (it is arriving 1-2 weeks) to start NEON optimizations.

biaoxiaoduan commented 5 years ago

That's great! I am looking foward to that.