On 08/05/2018 12:35 PM, Marc Glisse wrote: > On Sun, 5 Aug 2018, Bernd Edlinger wrote: > >>> merging log a + log b => log a*b and >> >> Maybe a*b could overflow, while adding the logarithms would not? > > Well, that's a discussion that happens every time a new transformation > is added to -funsafe-math-optimizations (I assume this one is under that > umbrella?). We already may get extra overflow with -fassociative-math > for instance. Sure, the overflow (or underflow!) is more likely for log > than for addition, but this still seems like the kind of transformation > that -ffast-math should enable, in my opinion. Now we can see if I am in > the minority... > One of the requirements for adding something to -ffast-math is that it doesn't break spec (we've never been specific about the version, though I think spec2017 and spec2006 are the only ones that matter anymore and the latter's importance diminishes daily).
I don't know if the original submitter has access to the spec suite, so we'll likely need to do that testing. I can't commit to any timeframe for that right now, but I think that's the biggest hurdle we need to jump. jeff
