On 2026-02-05 21:21, Chao Liu wrote: > Hi Max, > > I've been testing the OCP FP8 implementation by writing > a simple test suite in tests/fp/ that covers various boundary cases for E4M3, > E5M2, E2M1, and BFloat16 formats. During testing, I found some issues in the > float_class_inf handling in partsN(uncanon_sat). >
Hi Liu, Ooops looks like my random test cases miss the cases that input is InF from v2. Thanks for pointing out this issue. Richard has provided a v4 for the softfloat part, which is better and clearer than v3. I intend to address this issue based on that one. And I'll seperate the riscv isa part to another v4 based on the softfloat v4. Thanks, rnax
