[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87sikjhdfu.fsf@rasmusvillemoes.dk>
Date: Tue, 26 Aug 2014 10:17:57 +0200
From: Rasmus Villemoes <linux@...musvillemoes.dk>
To: David Miller <davem@...emloft.net>
Cc: dborkman@...hat.com, ffusco@...hat.com, tgraf@...g.ch,
linux-kernel@...r.kernel.org, matthew@....cx, wli@...omorphy.com,
akpm@...ux-foundation.org, nyc@...omorphy.com
Subject: Re: [PATCH/RFC] hash: Let gcc decide how to multiply
David Miller <davem@...emloft.net> writes:
> From: Daniel Borkmann <dborkman@...hat.com>
> Date: Mon, 25 Aug 2014 14:58:37 +0200
>
>> On 08/25/2014 02:13 PM, Rasmus Villemoes wrote:
>>> A 9+ years old comment in hash_64 says that gcc can't optimize
>>> multiplication by GOLDEN_RATIO_PRIME_64. Well, compilers get smarter
>>> and CPUs get faster all the time, so it is perhaps about time to
>>> revisit that assumption.
>>
>> Seems fine by me, but Cc'ing a couple of others (as those you have
>> Cc'ed haven't written that code :)). You might want to let your
>> changes go via Andrew's tree, too, perhaps ...
>
> You need to test the code generation for a cpu where the multiply
> instruction is more expensive than the shifts and adds.
Yes, I should, but I'm afraid I can't. I only have x86 hardware, and
even if I managed to set up some cross-compiler toolchain, I wouldn't
know what CPUs to target. So I was hoping someone else could test this
on appropriate platforms.
Thanks,
Rasmus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists