This is the mail archive of the
mailing list for the GCC project.
Re: DFA for PPro, P2, P3
> > The thing you may describe well using DFA and not using old scheme are the
> > decoders - ie PentiumPro has 3 decoders, where decoder 0 is able to decode
> > more than others so you need to order the triples of instructions issued
> > same cycle accordingly.
> Yes, there's multiple ways we can approach this -- either by having a cpu
> unit for each decoder or using the VLIW packing capabilities. It's unclear
> to me which will best best.
I modeled each decoder as unit (well, pentium has just two of them) and then
used presence set to ensure that pentium u unit is fed before pentium v unit.
This works quite naturally for me. The nonpairable instructions simply
allocate both units. I guess pentiumpro can use same scheme, just with
three units. Is there some better alternative?
This reminds me, that I should post the patch, give me some time to dig it
out, I've just returned from a trip.