Date: Tue, 19 Feb 2008 06:58:06 +0100
From: Willy Tarreau <w@1wt.eu>
To: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Andi Kleen <andi@firstfloor.org>, Arjan van de Ven <arjan@infradead.org>,
       Roel Kluin <12o3l@tiscali.nl>, geoffrey.levand@am.sony.com,
       linuxppc-dev@ozlabs.org, cbe-oss-dev@ozlabs.org,
       lkml <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 1/3] Fix Unlikely(x) == y
Message-ID: <20080219055806.GA8404@1wt.eu>
References: <47B70A61.9030306@tiscali.nl> <20080216094226.1e8eede1@laptopd505.fenrus.org> <p731w7axrat.fsf@bingen.suse.de> <200802191333.53607.nickpiggin@yahoo.com.au>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <200802191333.53607.nickpiggin@yahoo.com.au>
User-Agent: Mutt/1.5.11
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2001
Lines: 41

On Tue, Feb 19, 2008 at 01:33:53PM +1100, Nick Piggin wrote:
> > Note in particular the last predictors; assuming branch ending
> > with goto, including call, causing early function return or
> > returning negative constant are not taken. Just these alone
> > are likely 95+% of the unlikelies in the kernel.
> 
> Yes, gcc should be able to do pretty good heuristics, considering
> the quite good numbers that cold CPU predictors can attain. However
> for really performance critical code (or really "never" executed
> code), then I think it is OK to have the hints and not have to rely
> on gcc heuristics.

in my experience, the real problem is that gcc does what *it* wants and not
what *you* want. I've been annoyed a lot by the way it coded some loops that
could really be blazingly fast, but which resulted in a ton of branches due
to its predictors. And using unlikely() there was a real mess, because instead
of just hinting the compiler with probabilities to write some linear code for
the *most* common case, it ended up with awful branches everywhere with code
sent far away and even duplicated for some branches.

Sometimes, for performance critical paths, I would like gcc to be dumb and
follow *my* code and not its hard-coded probabilities. For instance, in a
tree traversal, you really know how you want to build your loop. And these
days, it seems like the single method of getting it your way is doing asm,
which obviously is not portable :-(

Maybe one thing we would need would be the ability to assign probabilities
to each branch based on what we expect, so that gcc could build a better
tree keeping most frequently used code tight.

Hmm I've just noticed -fno-guess-branch-probability in the man, I never tried
it.

regards,
Willy

--
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/