Skip to content
This repository has been archived by the owner on Jan 10, 2023. It is now read-only.

AVX512 support for Myelin #185

Merged
merged 2 commits into from
May 17, 2018
Merged

AVX512 support for Myelin #185

merged 2 commits into from
May 17, 2018

Conversation

ringgaard
Copy link
Contributor

I have added support for generating AVX512 code to Myelin.

  • Expression generator for AVX512
  • Support extended registers and opmask registers in macro assembler
  • AVX512 support for SoftMax, PoolingGather, ScatterAdd, Norm, and MatMul kernels

This PR also has a few "bonus" features:

  • Fix bug in hoisting
  • Add iterator support for inputs with trivial outer shape
  • Add missing __init.py__ file for myelin package
  • Learning rate decay for optimizers

@ringgaard ringgaard self-assigned this May 16, 2018
@ringgaard ringgaard requested a review from rahul1980 May 16, 2018 19:05
@ringgaard ringgaard merged commit 941939a into google:caspar May 17, 2018
@ringgaard ringgaard deleted the myelin512 branch May 17, 2018 07:17
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants