tkonolige opened a new pull request #10310: URL: https://github.com/apache/tvm/pull/10310
Add an intrinsic that performs a dot product of 8 4-element vectors at once. Also conditionally inline fused operators into the main convolution loop depending on convolutions size. Small convolution = no inlining. Performance improves by ~20% on mobilenet on raspberry pi 4 and ~30% improvement on performance for the individual convolutions. | model | before | after | speedup| | --- | --- | --- | --- | |mobilenet_v2 | 47.7ms | 39.6ms | 20.5%| |inception_v3 | 282.8ms | 209.1ms | 35.2%| |resnet_50 | 363.5ms | 321.3ms | 13.1%| @masahi @mbrookhart @leandron -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
