About Multi-cores and Multi-tasking

[Gilboa Davara]

On Wed, 2010-04-21 at 13:57 +0300, Oleg Goldshmidt wrote:
> On Wed, Apr 21, 2010 at 12:53 PM, Gilboa Davara <gilboad at gmail.com> wrote:
> 
> > In my experience (running a kernel based packet inspection software), HT
> > on Xeon 55xx yields around 15-20% performance benefit. (YMMV, of-course)
> 
> YMMV is the operative term here.
> 
> Shachar is right: the main problem with HT is increased cache
> contention for unrelated processes (causing "cache thrashing"). This
> may degrade performance for memory-intensive applications. This is the
> principal reason why many people switched HT off.

Of-course.
Again, I suggested that he'll benchmark his application with and w/o HT
before making a decision as the HT implementation in Core 7 based CPUs
has improved dramatically.

> Can you even buy a computer with just one CPU/core today?

As weird as it sounds, yep.
Intel has ATOM, embedded Core (Mostly under the Celeron M moniker). AMD
has Sempron CPUs, etc.

Per my example, I could just as well compare a dual Xeon E5502 (2 x 2
core) to a single underclocked Core i7 CPU or and Phenom X4 to an
Opteron 23xx.
But I preferred the old CPUs as I actually benchmarked the two options.
(Using my own application, dual socket, single core Opteron was ~20-30%
faster)

- Gilboa