| Issue |
180910
|
| Summary |
[RISCV] Bitcast long bitwidth int to vector type could be optimized.
|
| Labels |
backend:RISC-V
|
| Assignees |
|
| Reporter |
MouseSplinter
|
When I investigate the [codegenprepare issue](https://github.com/llvm/llvm-project/issues/179950), I find a codegen optimization opportunity.
https://rust.godbolt.org/z/zf94qjzMf
When bitcast i256 to <16 x i16> or <8 x i32>, it will fall back to scalar type.
![]()
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
