Dose TVM only support mkldnn dense operation?

Questions
#1

Hi,

I follow the build step in TVM and BLAS libraries to use the mkl-dnn. Then I run the script tutorials/get_started/relay_quick_start.py and print the module ir. It seems the convolution operation not use mkldnn. After that, I grep the src/include/python directory. I only find dense_mkldnn op.

My question is that TVM currently does not support convolution using the mkldnn library or dose I have missed some important information？

Git branch is master, commit id is 348144cb8b0485adca37aead0dfef9269cd2300d

#2

To use native libraries, you have to pass them to the libs flag like so: llvm -libs=mkdnn. Also make sure you have set USE_MKLDNN=On in build/config.cmake.

Right now we only support using mkldnn for matrix multiplication.

#3

@tkonolige Thanks for replying! I read the source code of tvm and find BYOC is exactly what’s I need. TVM use dnnl (a new mkldnn name) instead of mkldnn.

Howerver, is there a network benchmark to use dnnl compiler? I can only find snippet network in tests/python/relay/test_external_codegen.py. Is there a resnet or inception serials network to use the BYOC-based TVM?

#4

We do support network with DNNL in BYOC, but we mainly used it for BYOC demonstration and didn’t push its performance.