Jukka Larja (roskakori2006.delete@this.gmail.com) on October 30, 2020 11:29 am wrote:
> Mark Roulo (nothanks.delete@this.xxx.com) on October 30, 2020 10:21 am wrote:
> > Jukka Larja (roskakori2006.delete@this.gmail.com) on October 30, 2020 9:52 am wrote:
> > > me (me.delete@this.me.com) on October 30, 2020 9:47 am wrote:
> > > > > And I presume you are talking about AVX2. Both PS4 and Xbox One support AVX,
> > > > > and that alone obviously didn't gather any significant support (even though
> > > > > consoles have much more benefit from CPU optimizations compared to PC).
> > > > >
> > > > > -JLarja
> > > >
> > > > IIRC Jaguar's AVX isn't any faster than using SSE so devs mostly ignored it I guess.
> > >
> > > Is any AVX faster than SSE?
> > >
> > > -JLarja
> >
> > For my employer’s loads, AVX2 runs about 1.7x faster than SSE for compute bound loads.
>
> Yes, but that's AVX2, which is not available on Jaguar.
>
> -JLarja


Floating-point AVX operations were already 256-bit on Sandy Bridge and they had a decent speed increase over SSE, approaching double speed for favorable cases, like GEMM.

AVX2 of Haswell extended to 256-bit also the integer operations, simultaneous with the introduction of FMA, which could provide another doubling of the floating-point throughput.


So even with simple AVX, without AVX2 & FMA, like in Sandy Bridge/Ivy Bridge, the floating-point speed should be better, except for small CPUs (like presumably in Jaguar), which might cheat and implement AVX just for compatibility, without actually doubling the floating-point throughput.