Date: Mon, 13 Sep 2004 19:21:27 -0300
From: Marcelo Tosatti <marcelo.tosatti@cyclades.com>
To: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: William Lee Irwin III <wli@holomorphy.com>, Andrew Morton <akpm@osdl.org>,
       linux-kernel@vger.kernel.org
Subject: Re: [pagevec] resize pagevec to O(lg(NR_CPUS))
Message-ID: <20040913222127.GA23588@logos.cnet>
References: <20040909163929.GA4484@logos.cnet> <20040909155226.714dc704.akpm@osdl.org> <20040909230905.GO3106@holomorphy.com> <20040909162245.606403d3.akpm@osdl.org> <20040910000717.GR3106@holomorphy.com> <414133EB.8020802@yahoo.com.au> <20040910174915.GA4750@logos.cnet> <20040912045636.GA2660@holomorphy.com> <4143D07E.3030408@yahoo.com.au>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4143D07E.3030408@yahoo.com.au>
User-Agent: Mutt/1.5.5.1i
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 904
Lines: 23


> Sure. And when you fill it with pages, they'll use up 32KB of dcache
> by using a single 64B line per page. Now that you've blown the cache,
> when you go to move those pages to another list, you'll have to pull
> them out of L2 again one at a time.
>
> OK, so a 511 item pagevec is pretty unlikely. How about a 64 item one
> with 128 byte cachelines, and you're touching two cachelines per
> page operation? That's 16K.

Nick, 

Note that you dont read/write data to the actual pages most of the 
times pagevec's are used. The great majority is just page management code.  
So we dont really blow the caches like you said.

I agree we need more tests :) 

-
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/