From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pa0-f45.google.com (mail-pa0-f45.google.com [209.85.220.45]) by kanga.kvack.org (Postfix) with ESMTP id 0B5546B0113 for ; Mon, 11 Nov 2013 13:24:38 -0500 (EST) Received: by mail-pa0-f45.google.com with SMTP id kp14so5692496pab.4 for ; Mon, 11 Nov 2013 10:24:38 -0800 (PST) Received: from psmtp.com ([74.125.245.193]) by mx.google.com with SMTP id do4si9436591pbc.107.2013.11.11.10.24.36 for ; Mon, 11 Nov 2013 10:24:37 -0800 (PST) Message-ID: <1384194271.6940.40.camel@buesod1.americas.hpqcorp.net> Subject: Re: [PATCH] mm: cache largest vma From: Davidlohr Bueso Date: Mon, 11 Nov 2013 10:24:31 -0800 In-Reply-To: <20131111120116.GA21291@gmail.com> References: <1383337039.2653.18.camel@buesod1.americas.hpqcorp.net> <1383537862.2373.14.camel@buesod1.americas.hpqcorp.net> <20131104073640.GF13030@gmail.com> <1384143129.6940.32.camel@buesod1.americas.hpqcorp.net> <20131111120116.GA21291@gmail.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org List-ID: To: Ingo Molnar Cc: Linus Torvalds , Andrew Morton , Hugh Dickins , Michel Lespinasse , Mel Gorman , Rik van Riel , Guan Xuetao , "Chandramouleeswaran, Aswin" , Linux Kernel Mailing List , linux-mm On Mon, 2013-11-11 at 13:01 +0100, Ingo Molnar wrote: > * Davidlohr Bueso wrote: > > > Hi Ingo, > > > > On Mon, 2013-11-04 at 08:36 +0100, Ingo Molnar wrote: > > > * Davidlohr Bueso wrote: > > > > > > > I will look into doing the vma cache per thread instead of mm (I hadn't > > > > really looked at the problem like this) as well as Ingo's suggestion on > > > > the weighted LRU approach. However, having seen that we can cheaply and > > > > easily reach around ~70% hit rate in a lot of workloads, makes me wonder > > > > how good is good enough? > > > > > > So I think it all really depends on the hit/miss cost difference. It makes > > > little sense to add a more complex scheme if it washes out most of the > > > benefits! > > > > > > Also note the historic context: the _original_ mmap_cache, that I > > > implemented 16 years ago, was a front-line cache to a linear list walk > > > over all vmas (!). > > > > > > This is the relevant 2.1.37pre1 code in include/linux/mm.h: > > > > > > /* Look up the first VMA which satisfies addr < vm_end, NULL if none. */ > > > static inline struct vm_area_struct * find_vma(struct mm_struct * mm, unsigned long addr) > > > { > > > struct vm_area_struct *vma = NULL; > > > > > > if (mm) { > > > /* Check the cache first. */ > > > vma = mm->mmap_cache; > > > if(!vma || (vma->vm_end <= addr) || (vma->vm_start > addr)) { > > > vma = mm->mmap; > > > while(vma && vma->vm_end <= addr) > > > vma = vma->vm_next; > > > mm->mmap_cache = vma; > > > } > > > } > > > return vma; > > > } > > > > > > See that vma->vm_next iteration? It was awful - but back then most of us > > > had at most a couple of megs of RAM with just a few vmas. No RAM, no SMP, > > > no worries - the mm was really simple back then. > > > > > > Today we have the vma rbtree, which is self-balancing and a lot faster > > > than your typical linear list walk search ;-) > > > > > > So I'd _really_ suggest to first examine the assumptions behind the cache, > > > it being named 'cache' and it having a hit rate does in itself not > > > guarantee that it gives us any worthwile cost savings when put in front of > > > an rbtree ... > > > > So having mmap_cache around, in whatever form, is an important > > optimization for find_vma() - even to this day. It can save us at least > > 50% cycles that correspond to this function. [...] > > I'm glad it still helps! :-) > > > [...] I ran a variety of mmap_cache alternatives over two workloads that > > are heavy on page faults (as opposed to Java based ones I had tried > > previously, which really don't trigger enough for it to be worthwhile). > > So we now have a comparison of 5 different caching schemes -- note that > > the 4 element hash table is quite similar to two elements, with a hash > > function of (addr % hash_size). > > > > 1) Kernel build > > +------------------------+----------+------------------+---------+ > > | mmap_cache type | hit-rate | cycles (billion) | stddev | > > +------------------------+----------+------------------+---------+ > > | no mmap_cache | - | 15.85 | 0.10066 | > > | current mmap_cache | 72.32% | 11.03 | 0.01155 | > > | mmap_cache+largest VMA | 84.55% | 9.91 | 0.01414 | > > | 4 element hash table | 78.38% | 10.52 | 0.01155 | > > | per-thread mmap_cache | 78.84% | 10.69 | 0.01325 | > > +------------------------+----------+------------------+---------+ > > > > In this particular workload the proposed patch benefits the most and > > current alternatives, while they do help some, aren't really worth > > bothering with as the current implementation already does a nice enough > > job. > > Interesting. > > > 2) Oracle Data mining (4K pages) > > +------------------------+----------+------------------+---------+ > > | mmap_cache type | hit-rate | cycles (billion) | stddev | > > +------------------------+----------+------------------+---------+ > > | no mmap_cache | - | 63.35 | 0.20207 | > > | current mmap_cache | 65.66% | 19.55 | 0.35019 | > > | mmap_cache+largest VMA | 71.53% | 15.84 | 0.26764 | > > | 4 element hash table | 70.75% | 15.90 | 0.25586 | > > | per-thread mmap_cache | 86.42% | 11.57 | 0.29462 | > > +------------------------+----------+------------------+---------+ > > > > This workload sure makes the point of how much we can benefit of caching > > the vma, otherwise find_vma() can cost more than 220% extra cycles. We > > clearly win here by having a per-thread cache instead of per address > > space. I also tried the same workload with 2Mb hugepages and the results > > are much more closer to the kernel build, but with the per-thread vma > > still winning over the rest of the alternatives. > > That's also very interesting, and it's exactly the kind of data we need to > judge such matters. Kernel builds and DB loads are two very different, yet > important workloads, so if we improve both cases then the probability that > we improve all other workloads as well increases substantially. > > Do you have any data on the number of find_vma() calls performed in these > two cases, so that we can know the per function call average cost? > For the kernel build we get around 140 million calls to find_vma(), and for Oracle around 27 million. So the function ends up costing significantly more for the DB workload. Thanks, Davidlohr -- 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: email@kvack.org