On 05.06.2014 22:53, Dan McDonald wrote:
On Jun 5, 2014, at 3:50 PM, Alexander Motin <[email protected]> wrote:
I'd like to note that I've seen quite a lot of examples already, when
congestion barely measurable on 24 Westmere cores just explodes on 40
IvyBridge. This looks like one of them.
Must be nice to have that kind of HW. :)
Yea, not bad. It doesn't allow to relax, showing problems in software
that perfectly worked before. :) It was not so hard to reach million
disk IOPS from 12-core high-end desktop machine. On 24-core server I
spent few months trying to repeat that. On this new 40-core iron I
haven't succeeded yet.
I wonder if it's the 40 cores (20 cores with HT, right?), or the faster clocks
that are tickling things?
It is number of cores (and possibly architecture) that changes
everything (yes, it is 2 packages of 10 cores, plus HTT). The frequency
is more or less comparable.
--
Alexander Motin
_______________________________________________
developer mailing list
[email protected]
http://lists.open-zfs.org/mailman/listinfo/developer
