Re: [RFC PATCH] v2 mm: balance_dirty_pages. reduce calls to global_page_state to reduce cache references

[Richard Kennedy <richard@rsk.demon.co.uk>]

On Mon, 2009-09-07 at 11:06 +0100, Richard Kennedy wrote:
> On Sun, 2009-09-06 at 11:55 +0800, Wu Fengguang wrote:
> > On Fri, Sep 04, 2009 at 07:05:30PM +0800, Richard Kennedy wrote:
> > > Reducing the number of times balance_dirty_pages calls global_page_state
> > > reduces the cache references and so improves write performance on a
> > > variety of workloads.
> > > 
> > > 'perf stats' of simple fio write tests shows the reduction in cache
> > > access.
> > > Where the test is fio 'write,mmap,600Mb,pre_read' on AMD AthlonX2 with
> > > 3Gb memory (dirty_threshold approx 600 Mb)
> > > running each test 10 times, dropping the fasted & slowest values then
> > > taking 
> > > the average & standard deviation
> > > 
> > > 		average (s.d.) in millions (10^6)
> > > 2.6.31-rc8	648.6 (14.6)
> > > +patch		620.1 (16.5)
> > > 
> > > Achieving this reduction is by dropping clip_bdi_dirty_limit as it  
> > > rereads the counters to apply the dirty_threshold and moving this check
> > > up into balance_dirty_pages where it has already read the counters.
> > > 
> > > Also by rearrange the for loop to only contain one copy of the limit
> > > tests allows the pdflush test after the loop to use the local copies of
> > > the counters rather than rereading them.
> > > 
> > > In the common case with no throttling it now calls global_page_state 5
> > > fewer times and bdi_stat 2 fewer.
> > > 
> > > This version includes the changes suggested by 
> > > Wu Fengguang <fengguang.wu@intel.com>
> > 
> > It seems that an redundant pages_written test can be reduced by
> > 
> > --- linux.orig/mm/page-writeback.c	2009-09-06 11:44:39.000000000 +0800
> > +++ linux/mm/page-writeback.c	2009-09-06 11:44:42.000000000 +0800
> > @@ -526,10 +526,6 @@ static void balance_dirty_pages(struct a
> >  		    (background_thresh + dirty_thresh) / 2)
> >  			break;
> >  
> > -		/* done enough? */
> > -		if (pages_written >= write_chunk)
> > -			break;
> > -
> >  		if (!bdi->dirty_exceeded)
> >  			bdi->dirty_exceeded = 1;
> >  
> > @@ -547,7 +543,7 @@ static void balance_dirty_pages(struct a
> >  			pages_written += write_chunk - wbc.nr_to_write;
> >  			/* don't wait if we've done enough */
> >  			if (pages_written >= write_chunk)
> > -				continue;
> > +				break;
> >  		}
> >  		congestion_wait(BLK_RW_ASYNC, HZ/10);
> >  	}
> > 
> > Otherwise the patch looks good to me. Thank you for the nice work!
> > 
> > Reviewed-by: Wu Fengguang <fengguang.wu@intel.com>
> > 
> Thank you.
> 
> I'll give your suggestion a try & run some tests. I think you're right
> it should be better. Not re-reading the global counters again should be
> of some benefit!
> regards
> Richard
> 
I've tried this change and ran a series of tests overnight but it turns
out to be worse. The existing patch does better overall.

e.g. fio 600mb
cache references average (s.d.) in millions (10^6)

2.6.31-rc8	648.6 (14.6)
+patch		620.1 (16.5)
patch+change	635.6 (10.2)

I'm not sure how to trace this to work out exactly why, but without the
change the patch is closer to the behavior of the existing code.

So I think we should go with it as is.

regards
Richard
 



--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>