Re: [OT?] New Computer

[Shachar Shemesh]

AFAIK, this has more to do with algorithmic functions than with 
application layout. Some distributed algorithms will work better on 
hyperthreading, some will work less well.

I do know that the Linux kernel's spinlock is a problem for 
hyperthreading, for example. I don't know about other constructs, and 
whether there is any alternative with other OSes.

            Shachar

Baruch Shpirer wrote:

For some applications I know of , it seems the implementation of the
application itself proves to show better performance in some
Less significant os'es . It isn't a real smp os support issue but more
of programming skills
-----Original Message-----
From: Shachar Shemesh [mailto:[EMAIL PROTECTED] 
Sent: Sunday, August 10, 2003 06:33 AM
To: Orna Agmon
Cc: Adir Abraham; Baruch Shpirer; 'Linux-IL'; 'Micha Feigin'; 'Oded
Arbel'
Subject: Re: [OT?] New Computer

Orna Agmon wrote:

 

On Sat, 9 Aug 2003, Adir Abraham wrote:

   

Yes. Hyperthreading works fine in Linux (2.4.x) and it works as if you
     

 

had two processors. Actually, Linux doesn't mind about it, because it 
really looks for Linux as if you had two processors. It starts from 
the BIOS
  

     

Not exactly, from the performance point of view. It depends on what you
   

 

want to do with your machine. For some applications, hyper-threading 
might hurt the performance.

   

But is that a Linux specific problem? Will another OS under the same 
circumstances not suffer from performance penalties?

            Shachar

 



--
Shachar Shemesh
Open Source integration consultant
Home page & resume - http://www.shemesh.biz/


=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]