llvmir dialect¶
LLVM IR target punch-through operations. These ops preserve target specific intent in Loom IR while still lowering through structured LLVMIR builders instead of raw textual emission.
| Property | Value |
|---|---|
| Bytecode dialect ID | 0x11 |
| Registered by default | no |
| Operations | 3 |
Operations¶
| Operation | Summary |
|---|---|
llvmir.inline_asm |
Structured LLVM inline assembly call. |
llvmir.intrinsic |
Structured call to a supported LLVM intrinsic. |
llvmir.target |
LLVMIR target-family record. |