linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* (no subject)
@ 2001-03-06 23:55 Ying Chen
  2001-03-07  0:07 ` threads J . A . Magallon
  2001-03-07  0:40 ` your mail Don Dugger
  0 siblings, 2 replies; 3+ messages in thread
From: Ying Chen @ 2001-03-06 23:55 UTC (permalink / raw)
  To: linux-kernel

Hi,

I have two questions on Linux pthread related issues. Would anyone be able 
to help?

1. Does any one have some suggestions (pointers) on good kernel Linux thread 
libraries?
2. We ran multi-threaded application using Linux pthread library on 2-way 
SMP and UP intel platforms (with both 2.2 and 2.4 kernels). We see 
significant increase in context switching when moving from UP to SMP, and 
high CPU usage with no performance gain in turns of actual work being done 
when moving to SMP, despite the fact the benchmark we are running is 
CPU-bound. The kernel profiler indicates that the a lot of kernel CPU ticks 
went to scheduling and signaling overheads. Has anyone seen something like 
this before with pthread applications running on SMP platforms? Any 
suggestions or pointers on this subject?

Thanks a lot!

Ying



_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: threads
  2001-03-06 23:55 Ying Chen
@ 2001-03-07  0:07 ` J . A . Magallon
  2001-03-07  0:40 ` your mail Don Dugger
  1 sibling, 0 replies; 3+ messages in thread
From: J . A . Magallon @ 2001-03-07  0:07 UTC (permalink / raw)
  To: Ying Chen; +Cc: linux-kernel


On 03.07 Ying Chen wrote:
> 2. We ran multi-threaded application using Linux pthread library on 2-way 
> SMP and UP intel platforms (with both 2.2 and 2.4 kernels). We see 
> significant increase in context switching when moving from UP to SMP, and 
> high CPU usage with no performance gain in turns of actual work being done 
> when moving to SMP, despite the fact the benchmark we are running is 
> CPU-bound. The kernel profiler indicates that the a lot of kernel CPU ticks 
> went to scheduling and signaling overheads. Has anyone seen something like 
> this before with pthread applications running on SMP platforms? Any 
> suggestions or pointers on this subject?
> 

Too much contention ? How frequently do you create and destroy threads ?
How much frequently do they access shared-writable-data ?
How do you protect them ?

It seems like your system spents more time creating and killing threads
that doing real work.

-- 
J.A. Magallon                                                      $> cd pub
mailto:jamagallon@able.es                                          $> more beer

Linux werewolf 2.4.2-ac13 #3 SMP Wed Mar 7 00:09:17 CET 2001 i686


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: your mail
  2001-03-06 23:55 Ying Chen
  2001-03-07  0:07 ` threads J . A . Magallon
@ 2001-03-07  0:40 ` Don Dugger
  1 sibling, 0 replies; 3+ messages in thread
From: Don Dugger @ 2001-03-07  0:40 UTC (permalink / raw)
  To: Ying Chen; +Cc: linux-kernel

Ying-

I'm a little confused here.  It's very hard to compare a UP application
vs. the same app. converted to use threads.  Unless the app. is
structured such that multiple threads can run at the same time then
no, you won't see any improvement by going to SMP, in fact a true
single threaded app. will frequently slow down when run on an SMP
kernel.

Have you watched a CPU meter while your benchmark runs?  Even something
basic like `top' should give you a feel for whether or not your
using all of the CPU's.


On Tue, Mar 06, 2001 at 03:55:55PM -0800, Ying Chen wrote:
> Hi,
> 
> I have two questions on Linux pthread related issues. Would anyone be able 
> to help?
> 
> ...
>
> 2. We ran multi-threaded application using Linux pthread library on 2-way 
> SMP and UP intel platforms (with both 2.2 and 2.4 kernels). We see 
> significant increase in context switching when moving from UP to SMP, and 
> high CPU usage with no performance gain in turns of actual work being done 
> when moving to SMP, despite the fact the benchmark we are running is 
> CPU-bound. The kernel profiler indicates that the a lot of kernel CPU ticks 
> went to scheduling and signaling overheads. Has anyone seen something like 
> this before with pthread applications running on SMP platforms? Any 
> suggestions or pointers on this subject?
> 
> Thanks a lot!
> 
> Ying
> 
> 
> 
> _________________________________________________________________
> Get your FREE download of MSN Explorer at http://explorer.msn.com
> 
> -
> 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/

-- 
Don Dugger
"Censeo Toto nos in Kansa esse decisse." - D. Gale
n0ano@valinux.com
Ph: 303/938-9838

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2001-03-07  0:40 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-03-06 23:55 Ying Chen
2001-03-07  0:07 ` threads J . A . Magallon
2001-03-07  0:40 ` your mail Don Dugger

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).