Michael S (already5chosen.delete@this.yahoo.com) on August 10, 2014 5:50 am wrote:
> Wilco (Wilco.Dijkstra.delete@this.ntlworld.com) on August 10, 2014 5:00 am wrote:
> > Michael S (already5chosen.delete@this.yahoo.com) on August 10, 2014 4:27 am wrote:
> > > juanrga (nospam.delete@this.juanrga.com) on August 10, 2014 3:40 am wrote:
> > > > anon (no.delete@this.thank.you) on August 9, 2014 10:17 am wrote:
> > > > > juanrga (nospam.delete@this.juanrga.com) on August 9, 2014 6:43 am wrote:
> > > > > > Aaron Spink (aaronspink.delete@this.notearthlink.net) on August 9, 2014 3:44 am wrote:
> > > > > > > juanrga (nospam.delete@this.juanrga.com) on August 8, 2014 10:49 am wrote:
> > > > > > >
> > > > > > > > Take a modern A57 core. According to AMD the A57 Opteron is faster than jaguar based Opteron but
> > > > > > > > consumes less power. The ARM core performance is ~40% faster, and consumes roughly one half.
> > > > > > > >
> > > > > > > Faster at what? Its a bunch of market point pointing to nothingness.
> > > > > >
> > > > > > SPECint
> > > > > >
> > > > >
> > > > > Could you provide a link for this?
> > > >
> > > > http://www.anandtech.com/show/7724/it-begins-amd-announces-its-first-arm-based-server-soc-64bit8core-opteron-a1100
> > >
> > >
> > > SPECInt "estimate" is very problematic. It can mean that they reallly ran the suite on production system,
> > > but for whatever reason do not want to submit the result to SPEC.org. Or it could mean that they really
> > > estimated it and didn't measure. Obviously, in the second case the number has lower credibility.
> > >
> > > Anyway, SPECint_rate2006=80 is an approximate equivalent of Intel Pentium G3220 - $64 CPU.
> > > That's not a very good place to be for new, unproven and relatively low volume product.
> > > http://www.spec.org/cpu2006/results/res2014q2/cpu2006-20140407-29307.html
> >
> > G3220 has a TDP of 53W, not exactly microserver territory.
> >
> > Also there is no point in comparing with Intel SPEC results due to the Intel compiler breaking
> > some of the SPEC benchmarks - the correct result is likely about 55. If you want to compare
> > CPU performance using SPEC, you have to use the same compiler and settings.
> >
> > Wilco
>
> Since there is no official definition of "microserver" and since different companies manufacture
> very different devices under that name I can't say if 53W is microserver territory or not.

Well these are big power hungry cores, so I wouldn't call them microservers. The whole point is replacing big cores with small cores in tasks that can use lots of cores and improve perf/W that way.

> From technical perspective , something like Core i3-4350T looks like smarter forward-looking buy: higher throughput
> (HT, cache), lower TDP, non-crippled ISA (AES, AVX2). And not that much more expensive than G3220.

Still when you adjust SPEC scores for icc then Seattle has 60-70% perf/W advantage.

Wilco