dsteger opened a new pull request #17: URL: https://github.com/apache/tvm-vta/pull/17
Update compute instruction VTA_MEM_ID_ACC to load using the load_pad_2d method instead of the load_2d. This resolves failing unit testing "test_padded_load()". You will also notice that the indexs are calculated without pragmas to use LUTs over DSPs. We save 2 DSPs and 1 LUT by not using the pragma. I've only tested this commit with the ultra96 using unit test ""test_vta_insn.py". ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
