giuseros opened a new pull request #6110: URL: https://github.com/apache/incubator-tvm/pull/6110
This improves reduction by ~10x using x86 schedules. While for `float32` reduction is not common, it is widely used in quantized operations (`int8`, `uint8`) in order to compute the offset contributions. We might revisit this later to tailor the schedule for Arm CPUs. Change-Id: I9cd85deac6a57666b82ff7250d827652a4000d82 ---------------------------------------------------------------- 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]
