JackWeiw commented on PR #16679: URL: https://github.com/apache/tvm/pull/16679#issuecomment-1982225958
> I did a bit of testing, and these simplifications should not occur. For the test case `TestCase(tvm.te.min(0, fld(x + 3, 4) * 4 - 1), 0)`, substituting `0` for `x`, the `min(0, fld(x+3, 4)*4 - 1)` would evaluate to `min(0, -1)`. However, this simplification would cause it to instead evaluate to `0`. > > Similarly for `TestCase(tvm.te.max(1, fld(x + 3, 4) * 4), fld(x + 3, 4) * 4)`, substituting zero for `x` results in `max(1, 0)`, but the simplification rule would result in `0`. Sir, x here refers to col size of a input matrix, it should never be zero, would it be make sense if i add condition that x > 0？Or would please give me some advise on how to make [script here](https://gist.github.com/JackWeiw/3e7bd0f1ed62225505916e87738d7751) the index (when read a col of matrix into memory and the len of col is dynamic) more simplify? -- 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]
