lkml.org 
[lkml]   [2002]   [Oct]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Kernel reports 4 CPUS instead of 2...
Mark Mielke wrote:

>On Wed, Oct 16, 2002 at 01:47:32PM -0700, Samuel Flory wrote:
>
>
>> Try shutting off hyperthreading in the bios. Keep in mind
>>hyperthreading is net loss if you are running a single nonthreaded app.
>>Also you might want to check if there aren't io speed issues.
>>
>>
>
>Is this true? It seems to me that the 'on-demand execution units' would
>simply be devoted to the one task, resulting in zero loss.
>
>

In perfect world yes, but in reality there is overhead. I've tested
this on a quad xeon. A "make bzImage" is a bit faster with
hyperthreading off. Of course a make -j 8 bzImage is faster with
hyperthreading on. I haven't tried this on a dual xeon. (It could be a
scaling issue 4 vs 8 processors.)

>I see hyperthreading becoming a problem if two threads are scheduled to
>execute at the same time before the operating system, and if they each
>need access to the same execution units at the same time.
>
>
And if both threads need different items in L(whatever) cache it gets
even worse.

There are a few good overviews on the subject:
http://www.intel.com/technology/hyperthread/
http://arstechnica.com/paedia/h/hyperthreading/hyperthreading-1.html
http://www.anandtech.com/cpu/showdoc.html?i=1576&p=2

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:30    [W:0.084 / U:0.764 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site