[GitHub] [incubator-tvm] merrymercy commented on pull request #6040: [TOPI] Improve schedule for injective on x86
merrymercy commented on pull request #6040: URL: https://github.com/apache/incubator-tvm/pull/6040#issuecomment-662727589 I close this PR because after more test I found the current python version in tvm master works well. This pr is trying to fix the c++ version, which is only used in the old codebase. 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: us...@infra.apache.org
[GitHub] [incubator-tvm] merrymercy commented on pull request #6040: [TOPI] Improve schedule for injective on x86
merrymercy commented on pull request #6040: URL: https://github.com/apache/incubator-tvm/pull/6040#issuecomment-660172510 Currently, both c++ and python versions work well. This PR is ready to be merged from my side. I have no idea how to solve the mess of duplicated python and c++ topi functions. 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: us...@infra.apache.org
[GitHub] [incubator-tvm] merrymercy commented on pull request #6040: [TOPI] Improve schedule for injective on x86
merrymercy commented on pull request #6040: URL: https://github.com/apache/incubator-tvm/pull/6040#issuecomment-660172683 Currently, both c++ and python versions work well. This PR is ready to be merged from my side. cc @icemelon9 I have no idea how to solve the mess of duplicated python and c++ topi functions. 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: us...@infra.apache.org
[GitHub] [incubator-tvm] merrymercy commented on pull request #6040: [TOPI] Improve schedule for injective on x86
merrymercy commented on pull request #6040: URL: https://github.com/apache/incubator-tvm/pull/6040#issuecomment-659141089 I added a python version. It handles parallel and vectorization generally for most input shapes. I tested in with BERT (seq_len=128, batch=1) on c5.9xlarge. It reduces the latency from 19.0ms to 18.70ms. My questions: 1. Why do we have both c++ version and python version? Should we just delete the python version? 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: us...@infra.apache.org
[GitHub] [incubator-tvm] merrymercy commented on pull request #6040: [TOPI] Improve schedule for injective on x86
merrymercy commented on pull request #6040: URL: https://github.com/apache/incubator-tvm/pull/6040#issuecomment-659138546 Why do we have both c++ version and python version? Should we just delete the python version? 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: us...@infra.apache.org
[GitHub] [incubator-tvm] merrymercy commented on pull request #6040: [TOPI] Improve schedule for injective on x86
merrymercy commented on pull request #6040: URL: https://github.com/apache/incubator-tvm/pull/6040#issuecomment-657384814 I pushed some changes to my branch. But why does not GitHub sync the changes to this PR? 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: us...@infra.apache.org