I think we can use softfloat for everything now. Years ago, softfloat wasn't BSD. So, we somewhat rolled our own to get the Arm instructions working, but it wasn't comprehensive. Since softfloat was added in the past few months, no one has yet migrated the old code. As far as I know, this would be a safe thing to do.
Cheers, Jason On Fri, May 14, 2021 at 4:15 AM Gabe Black via gem5-dev <gem5-dev@gem5.org> wrote: > Hi folks. I've taken a little time recently to go macro hunting, and while > doing that I noticed what I think are three different software floating > point implementations in gem5. There is ext/fputils which has support for > 80 bit floats, arch/arm/insts/fplib.hh and .cc which I'm assuming ARM uses, > and ext/softfloat which RISCV uses. > > I can't think of a good reason for us to need three different solutions to > the same problem. Is there a good reason for this, or was each just not > aware of the others? Can we fold one (or two) of these into a preferred > implementation and get rid of the extra(s)? > > Gabe > _______________________________________________ > gem5-dev mailing list -- gem5-dev@gem5.org > To unsubscribe send an email to gem5-dev-le...@gem5.org > %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
_______________________________________________ gem5-dev mailing list -- gem5-dev@gem5.org To unsubscribe send an email to gem5-dev-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
