Open seven-mile opened 3 weeks ago
Original CodeGen treat vec3 as vec4 to get aligned memory access. It would be valuable to document as a testcase how these accesses look like in CIR and LLVM.
Originally posted by @jopperm in https://github.com/llvm/clangir/pull/674#pullrequestreview-2107348018
Original CodeGen treat vec3 as vec4 to get aligned memory access. It would be valuable to document as a testcase how these accesses look like in CIR and LLVM.
Originally posted by @jopperm in https://github.com/llvm/clangir/pull/674#pullrequestreview-2107348018