https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126240
Aldy Hernandez <aldyh at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rguenth at gcc dot gnu.org
Blocks| |126010
--- Comment #1 from Aldy Hernandez <aldyh at gcc dot gnu.org> ---
As a proof of concept, I hacked an incredibly ugly, not fit for public viewing
patch, that implements inequalities with frange, and all 56 failures disappear.
So I think that if we had frange inequalities, that’s all (??) of what we need
for DOM/ranger floating point parity. At least, when it comes to LAPACK.
I think this is as far as I’ll take it, cause more poking at this is
diminishing returns. I have filed PRs for everything we’re obviously missing,
and as Richi mentioned, perhaps time is better spent removing the DOM threader,
and seeing what falls (once testsuite regressions are analyzed and accounted
for).
Thoughts?
Referenced Bugs:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126010
[Bug 126010] [meta-bug] Remove DOM
