[Beowulf] memory bandwidth scaling

[Miguel Costa]

> Well, that's the rub.  If we compare:
>
> DUAL INTEL XEON 6C E5-2620V3 (2.4GHz/8GT/s/15MB)            $870
> DUAL INTEL XEON 12C E5-2670V3 (2.3GHz/9.6GT/s/30MB) CPU  [+$2400.00]
>
> So for the marginal cost we're close to being able to buy a whole other
node.

Well, if you use only 6 cores of the 12 core models, you do get access to
double the cache per core (the 30MB is last level cache, shared among
cores), that could be quite significant for memory bound codes.

And for codes that are not memory bound, you still have the option of using
all cores, which would be equivalent to the "other node" option, cpuwise.

But which is better? Yes, it depends :)

my 2 cents

Miguel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.beowulf.org/pipermail/beowulf/attachments/20151004/ee55b85d/attachment.html>