[libre-riscv-dev] partitioned compare and mux

Michael Nolan mtnolan2640 at gmail.com
Fri Feb 7 15:24:24 GMT 2020

Luke Kenneth Casson Leighton <lkcl at lkcl.net> writes:

> i have a feeling that partitioned-not-equal could prooobably do with
> being an actual operator in eq_gt_ge.py because i suspect that the
> external negation of in-A, just to make the partitioned-eq go the
> other way round, would be more efficiently done by inverting the
> single-bit *result* of the comparisons rather than inverting *all* the
> bits of the input.

Partsig does exactly this right now, instead of negating A (which
doesn't give the correct result anyway), it negates the output of (A==B)
from the partitioned comparison module. While I don't have a problem
with adding the negation of the output bits to the partitioned
comparison, it shouldn't make a difference, size or speed-wise, whether
it happens in eq_gt_ge.py or partsig.py

> if you can do that i think we can throw a stake in the ground, call
> #171 done, and add it to the list of things to submit an rfp for.
> http://bugs.libre-riscv.org/show_bug.cgi?id=171


More information about the libre-riscv-dev mailing list