masahi edited a comment on pull request #7669: URL: https://github.com/apache/tvm/pull/7669#issuecomment-800610821
Looks like one reasonable way to implement `ceil(log2(x)` is ` 32 - clz(x) + (x & (x-1) ? 1 : 0)` for 32 bit integers. We need to be careful with 32 bit vs 64 bit and signed vs unsigned. We need to add intrinsic lowering of `tvm.tir.clz` for llvm and spirv. I'll do that next week. ---------------------------------------------------------------- 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]
