[Bug Report] Incorrect upward rounding result of fsub.s instruction

[zhangkanqi]

Hi,

When executing the fsub.s ft5, ft5, ft2, rup instruction in Spike, I found that the result is incorrect, where ft5 is initialized as 0x284d52df and ft2 is initialized as 0x7b418a35.
I believe the correct final result in ft5 should be 0xfb418a34 (as Spike does), however, the result calculated in RSD is 0xfb418a35.

[lpha-z]

Thanks for your bug report.
The current implementation of the floating-point arithmetic unit does not support directional rounding (so really it does not fully support RV32F).
We plan to add support for directional rounding in the near future.

[zhangkanqi]

I see. Thanks a lot.

[lpha-z]

We reopen this report because we believe it will help others who have faced similar problems.
When we finish supporting the directional rounding, we will close this report.