Re: [PATCH 11/17] find: micro-optimize for_each_{set,clear}_bit()

From: Yury Norov <yury.norov@gmail.com>
To: Petr Mladek <pmladek@suse.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org,
	linux-arch@vger.kernel.org, linux-kselftest@vger.kernel.org,
	linux-mmc@vger.kernel.org, linux-perf-users@vger.kernel.org,
	kvm@vger.kernel.org,
	"James E.J. Bottomley" <James.Bottomley@hansenpartnership.com>,
	Alexander Lobakin <alobakin@pm.me>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Alexey Klimov <aklimov@redhat.com>,
	Andrea Merello <andrea.merello@gmail.com>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	Arnaldo Carvalho de Melo <acme@redhat.com>,
	Arnd Bergmann <arnd@arndb.de>, Ben Gardon <bgardon@google.com>,
	Benjamin Herrenschmidt <benh@kernel.crashing.org>,
	Brian Cain <bcain@codeaurora.org>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Christoph Lameter <cl@linux.com>,
	Daniel Bristot de Oliveira <bristot@redhat.com>,
	David Hildenbrand <david@redhat.com>,
	Dennis Zhou <dennis@kernel.org>,
	Geert Uytterhoeven <geert@linux-m68k.org>,
	Heiko Carstens <hca@linux.ibm.com>,
	Ian Rogers <irogers@google.com>, Ingo Molnar <mingo@redhat.com>,
	Jaegeuk Kim <jaegeuk@kernel.org>,
	Jakub Kicinski <kuba@kernel.org>, Jiri Olsa <jolsa@redhat.com>,
	Joe Perches <joe@perches.com>, Jonas Bonn <jonas@southpole.se>,
	Leo Yan <leo.yan@linaro.org>, Mark Rutland <mark.rutland@arm.com>,
	Namhyung Kim <namhyung@kernel.org>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	Paolo Bonzini <pbonzini@redhat.com>, Peter Xu <peterx@redhat.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Rasmus Villemoes <linux@rasmusvillemoes.dk>,
	Rich Felker <dalias@libc.org>,
	Samuel Mendoza-Jonas <sam@mendozajonas.com>,
	Sean Christopherson <seanjc@google.com>,
	Sergey Senozhatsky <senozhatsky@chromium.org>,
	Shuah Khan <shuah@kernel.org>,
	Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>,
	Steven Rostedt <rostedt@goodmis.org>, Tejun Heo <tj@kernel.org>,
	Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
	Ulf Hansson <ulf.hansson@linaro.org>,
	Will Deacon <will@kernel.org>,
	Wolfram Sang <wsa+renesas@sang-engineering.com>,
	Yoshinori Sato <ysato@users.sourceforge.jp>
Subject: Re: [PATCH 11/17] find: micro-optimize for_each_{set,clear}_bit()
Date: Mon, 30 Aug 2021 09:15:05 -0700	[thread overview]
Message-ID: <YS0ECS2Lb4rwqJ4b@yury-ThinkPad> (raw)
In-Reply-To: <20210830121249.2fgyvf47py2tz5s5@pathway.suse.cz>

On Mon, Aug 30, 2021 at 02:12:49PM +0200, Petr Mladek wrote:
> On Thu 2021-08-26 14:09:55, Yury Norov wrote:
> > On Thu, Aug 26, 2021 at 03:57:13PM +0200, Petr Mladek wrote:
> > > On Sat 2021-08-14 14:17:07, Yury Norov wrote:
> > > > The macros iterate thru all set/clear bits in a bitmap. They search a
> > > > first bit using find_first_bit(), and the rest bits using find_next_bit().
> > > > 
> > > > Since find_next_bit() is called shortly after find_first_bit(), we can
> > > > save few lines of I-cache by not using find_first_bit().
> > > 
> > > Is this only a speculation or does it fix a real performance problem?
> > > 
> > > The macro is used like:
> > > 
> > > 	for_each_set_bit(bit, addr, size) {
> > > 		fn(bit);
> > > 	}
> > > 
> > > IMHO, the micro-opimization does not help when fn() is non-trivial.
> >  
> > The effect is measurable:
> > 
> > Start testing for_each_bit()
> > for_each_set_bit:                15296 ns,   1000 iterations
> > for_each_set_bit_from:           15225 ns,   1000 iterations
> > 
> > Start testing for_each_bit() with cash flushing
> > for_each_set_bit:               547626 ns,   1000 iterations
> > for_each_set_bit_from:          497899 ns,   1000 iterations
> > 
> > Refer this:
> > 
> > https://www.mail-archive.com/dri-devel@lists.freedesktop.org/msg356151.html
> 
> I see. The results look convincing on the first look.
> 
> But I am still not sure. This patch is basically contradicting many
> other patches from this patchset:
> 
>   + 5th patch optimizes find_first_and_bit() and proves that it is
>     much faster:
> 
>     Before (#define find_first_and_bit(...) find_next_and_bit(..., 0):
>     Start testing find_bit() with random-filled bitmap
>     [  140.291468] find_first_and_bit:           46890919 ns,  32671 iterations
>     Start testing find_bit() with sparse bitmap
>     [  140.295028] find_first_and_bit:               7103 ns,      1 iterations
> 
>     After:
>     Start testing find_bit() with random-filled bitmap
>     [  162.574907] find_first_and_bit:           25045813 ns,  32846 iterations
>     Start testing find_bit() with sparse bitmap
>     [  162.578458] find_first_and_bit:               4900 ns,      1 iterations
> 
>        => saves 46% in random bitmap
> 	  saves 31% in sparse bitmap
> 
> 
>   + 6th, 7th, and 9th patch makes the code use find_first_bit()
>     because it is faster than find_next_bit(mask, size, 0);
> 
>   + Now, 11th (this) patch replaces find_first_bit() with
>     find_next_bit(mask, size, 0) because find_first_bit()
>     makes things slower. It is suspicious at minimum.
> 
> 
> By other words. The I-cache could safe 10% in one case.
> But find_first_bit() might safe 46% in random case.

Those are different cases. find_first_bit() is approximately twice
faster than find_next_bit, and much smaller. The conclusion is simple:
use 'first' version whenever possible if there's no other considerations.

In case of for_each_bit() macros, however, we have such a consideration.
In contrast to regular pattern, where user calls either first, or next
versions N times, here we call find_first_bit once, and then find_next_bit
N-1 times.

Because we know for sure that we'll call find_next_bit shortly, we can
benefit from locality under heavy pressure on I-cache, if replace 'first'
with 'next'. Consider it as a prefetch mechanism for the following calls
to find_next_bit().

> Does I-cache cost more than the faster code?

In this case cache miss is more expensive.

> Or was for_each_set_bit() tested only with a bitmap
> where find_first_bit() optimization did not help much?

I tried to ensure that the effect of I-cache is real and in this case
more important than code performance, so in the test I called 'first'
once and 'next' twice.

> How would for_each_set_bit() work with random bitmap?

It would work for all bitmaps.

> How does it work with larger bitmaps?

Percentage gain (but not absolute) will decrease proportionally to the
number of calls of find_next_bit() for big N.

Thanks,
Yury

> Best Regards,
> Petr