[llvm] [AMDGPU] Stop rounding up LDS block size for gfx950 (PR #208046)

Changpeng Fang via llvm-commits llvm-commits at lists.llvm.org
Tue Jul 28 06:56:20 PDT 2026


================

----------------
changpeng wrote:

Yes, the granularity is specified by hardware, and  should not be round up to power-of-2. Thanks for catching this. 

https://github.com/llvm/llvm-project/pull/208046


More information about the llvm-commits mailing list