question about Intel P4 versus Alpha's

Richard Walsh rbw at ahpcrc.org
Fri Jan 10 17:49:46 EST 2003


On Fri, 10 Jan 2003 09:53:21 Dominic Wu wrote:

> Is HT anything more than a thinly-veiled attempt at luring more software
> developers to develop multi-threaded applications so as to help Intel sell
> more CPU in the future?  (I.E. the new fangled software that is optimized
> for HT can really benefit from additional REAL processors instead of using
> just HT?)

I would say no as well.  In a single sentence, hyperthreading (its better
than "super" or "multi" threading ;-) ) delivers idle work slots from each 
functional unit inside all superscalar processors in an SMP system to 
any 'needs-to-be-run' thread of work from any process. 

There is some scheduling over head of course.

The following article does an excellent job of explaining this.

http://arstechnica.com/paedia/h/hyperthreading/hyperthreading-1.html

However, YMMV depending on job type/work mix.  In the worst case,
where two processes execute identical instructions at the same time,
there will be no benefit over multithreading. On the other hand,
several synthetic/nonsense threads designed/insinuated to use only 
idle slots in superscalar functional unit space (which spans processors) 
could perhaps be run simultaneously without disturbing the performance
of the above process.  

What you will see depends on where between these two extremes your 
hyperthreaded workload comes to rest.  More reports measuring the 
benefits for "HPC code" are needed to detect the quantity of "hype"
in hyperthreading.

;-)

rbw

_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list