Yingwei Zheng
Yingwei Zheng
@opusforlife2 How is this issue going now? If you are not working on the issue, I want to give it a try.
@emjas92 @SameenAhnaf I think the first video/audio enqueued should remain paused. Users can click "Start playing" if it needs to play immediately. The current implementation enqueues the video/audio to the...
Could you please provide proofs about FMF propagation for all your recent issues? Thank you :)
> Should this be done for SDPatternMatch as well? I don't think MIPatternMatch has an equivalent though SDPatternMatch always copies APInt :( We should refactor `ISD::isConstantSplatVector` first. https://github.com/llvm/llvm-project/blob/4d69855e9d380ecb7c1f7a64c7b37258fe36f525/llvm/include/llvm/CodeGen/SDPatternMatch.h#L667-L688 https://github.com/llvm/llvm-project/blob/4d69855e9d380ecb7c1f7a64c7b37258fe36f525/llvm/include/llvm/CodeGen/SDPatternMatch.h#L695-L707
> Failed Tests (3): LLVM :: CodeGen/AMDGPU/amdgpu-simplify-libcall-pow.ll LLVM :: CodeGen/AMDGPU/amdgpu-simplify-libcall-pown.ll LLVM :: CodeGen/AMDGPU/simplify-libcalls.ll
> generalization may be a little tricky for this one. What is the generalized pattern?
> but we cannot transform it in LLVM? It is just not implemented. But I am not willing to do this in LLVM. We should fix it in `hashbrown` or...
An example from `tokio-rs`: https://godbolt.org/z/Gr55dfMc8
@nikic See the IR diff in https://github.com/dtcxzyw/llvm-opt-benchmark/pull/289/files: bench/diesel-rs/optimized/1hskzwx2vflsavf7.ll: https://godbolt.org/z/8e6qd9Gjo bench/image-rs/optimized/2mngkegtim1o10y3.ll: enables factorization: `(%2505 + 1) * 392 + 15 --> %2505 * 392 + 407`
> @dtcxzyw Most of the callers that eventually arrived in `hashbrown::raw::RawTableInner::free_buckets` had `std::collections::HashMap::some_fn` higher on the stack, right? (I don't know what your tooling is like as to whether it...