[Mlir-commits] [mlir] [mlir][spirv] Add SPV_EXT_float8 support (PR #179246)

[Davide Grohmann]

davidegrohmann wrote:

> https://github.com/llvm/llvm-project/blob/266a45128a843c4232133507edd7d7b9d3da0ed6/llvm/lib/Support/APFloat.cpp#L61-L70

I came up with that conclusion by reading the table in the paper and the descriptions of the types given in here:
https://github.com/llvm/llvm-project/blob/main/mlir/include/mlir/IR/BuiltinTypes.td#L103
and
https://github.com/llvm/llvm-project/blob/main/mlir/include/mlir/IR/BuiltinTypes.td#L145
(which refer the same paper btw.)

If you wonder why not Float8E5M2FNUZ it is because the representation of NaN given at
https://github.com/llvm/llvm-project/blob/main/mlir/include/mlir/IR/BuiltinTypes.td#L168
does not match the table above. 


https://github.com/llvm/llvm-project/pull/179246