[PATCH] D145542: [RISCV] Don't parse the decimal minimum value for fli.s/fli.d/fli.h.

Tue Mar 14 20:37:04 PDT 2023

joshua-arch1 added a comment.

In D145542#4195327 <https://reviews.llvm.org/D145542#4195327>, @craig.topper wrote:

> In D145542#4195310 <https://reviews.llvm.org/D145542#4195310>, @joshua-arch1 wrote:
>
>> @craig.topper Since we parse the instructions using double precision and use separate lookup tables for fli.h/fli.d now, I think there are no reason not to parse minimum v
>
>
>
> In D145542#4195310 <https://reviews.llvm.org/D145542#4195310>, @joshua-arch1 wrote:
>
>> @craig.topper Since we parse the instructions using double precision and use separate lookup tables for fli.h/fli.d now, I think there are no reason not to parse minimum values for FLI.
>
> The min value for double is very long. I don’t think anyone would use it. I think we should ask on the riscv-isa-manual GitHub before supporting it.

I think we can at least test min value for half. Also, the min value for double is 2^(-1022) which is a certain number. There will be some circumstances where someone will use its specific value.

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D145542/new/

https://reviews.llvm.org/D145542