Wes Felter (wmf.delete@this.felter.org) on November 19, 2013 1:06 pm wrote:
> Michael S (already5chosen.delete@this.yahoo.com) on November 18, 2013 3:57 pm wrote:
>
> > I can imagine brand new core only in one scenario - if VPU-less variant of the same core is used
> > for Intel's variant of Niagara. But, then again, why would Intel want to build its own Nigara
> > when even the latest iteration of SparcT is seriously less Niagarish then the previous one.
>
> Isn't Avoton Intel's Niagara? I guess it doesn't have the threads.

Avoton does not have the threads (8 vs 128 in Sparc T5) and most likely it does not have enough of parallelism in the memory subsystem to support too many threads.

If Avoton was useful for throughput computers then by now either Intel or one of their partners would already publish benchmark results for at least one of the common server throughput tests.
In that regard Avoton is very similar to various ARM microserver offerings - they all claim better performance per watt than "normal" Xeons, but struggle to find established benchmarks that would confirm the claim.