[mythtv-users] What is a "dual core processor"?
drescherjm at gmail.com
Thu Dec 4 20:27:14 UTC 2008
On Thu, Dec 4, 2008 at 2:32 PM, ryan patterson <ryan.goat at gmail.com> wrote:
> On Thu, Dec 4, 2008 at 2:00 PM, Jeff Thurston <jthurston at skyline-ats.com> wrote:
>> Sorry for the top-post, I use Outlook :/
>> Hyperthreading has been described by many as "latency hiding".
>> I won't go into gory detail of what that means, but the simple answer is
>> "No." it is nothing close to dual-core. It was/is mostly a gimmick to get
>> around the poor performance of the early Pentium4 design.
>> For AV processing, hyperthreading could even hurt your performance.
> WTF! Don't apologize for top posting just don't do it. Does outlook
> actively stop you from scrolling to the bottom of the message before
> you type your reply?
> Furthermore hyperthreading is not a gimmick like you are saying.
> Hyperthreading is very useful in systems that spend a lot of time near
> 100% CPU usage. It is not useful for systems that sit idle most of
> the time and only want maximum performance in short bursts. In
> summary, for 99.9% of home users hyperthreading is not useful. But
> for 99.9% of the computational problems that get solved on computers
> hyperthreading is useful.
Hyperthreading was designed to make use of the excessively long
pipeline in PIV chips. With these chips the pipeline stalled
frequently because of branch mispredictions. During stalls the
processor basically is not executing new instructions so this leads to
inefficiency. So a clever trick would be to add a second set of
registers and make use of the free slots in the pipeline during these
stalls. In PIV this resulted in a -10% to 50% improvement when HT was
enabled. Core2 shortened the pipeline (and thus the cpu stalled less
frequently) so that HT would not be efficient. I am not sure what I7
did since it does not have the long pipeline of the PIV.
More information about the mythtv-users