| Re: [eigen] [Fwd: eigen cwise] unexpected performance regression with eigen's cwise() |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] [Fwd: eigen cwise] unexpected performance regression with eigen's cwise()
- From: Rohit Garg <rpg.314@xxxxxxxxx>
- Date: Thu, 20 Aug 2009 10:10:37 +0530
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:content-type :content-transfer-encoding; bh=eZ9WTPx54mtenVsCcbim8Vdidb2rFjdNQm6noT5LEhg=; b=I8CEzZtSdnbiplA2zSk7U9kZ6tdZhYvaFAIS6KqcEelzPdOXjUcpRuV/QA9GIQFHO8 L0C9qvfjSLjh2OW85PR+5TRFBqS6VMLIF1r+VodBqdX0aM0MMu6SCP/v+jZeb9UulTPv ZcTgwzLbNpS4l+5L43pA/9JQ9hT6L9X6Ascn0=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type:content-transfer-encoding; b=k84eESHhd167+H/OQ9iJ5FPkH019xKrjqux5vkBvP5aoiesnUJa8dmr6jf1CFG6o7X Ew9aNTypAZa46rxuMgCDTWuOgrhqesm/JPq182T4RcMVgcn2AOvdUAqdN5jXzaVF7cKx JbBgGrBFYpGuhVDwJfhF51Z2so6GrJgh9JSqI=
I'll ask him to try gcc 4.4 as well. But I don't see how it is
accessing memory out of bounds. The loop in i goes from [0 , SIZE-2),
or [0,SIZE-3], So at max, he accesses [0, SIZE-3+2] which should be
fine.
On Wed, Aug 19, 2009 at 11:58 PM, Markus
Moll<markus.moll@xxxxxxxxxxxxxxxx> wrote:
> Hi
>
> On Wednesday 19 August 2009 20:01:26 Rohit Garg wrote:
>> A freind of mine who is using eigen for his code, hit this unexpected
>> speed bump. I have no idea why.
>
> Undefined behavior, as the program tries to access blah[SIZE] and
> blah[SIZE+1]. Assuming that your friend wrote j instead of i, does he have all
> optimizations (and SSE) enabled?
>
> Otherwise, the unstable branch works fine for me:
>
> your friend's "cwiseop":
> 14.66s, 14.91s, 14.65s
>
> Eigen:
> 7.67s, 7.67s, 7.74s
>
> (Core2 Quad 2.4 GHz, gcc 4.4.0, fedora 11, eigen unstable)
>
> Markus
> --
> PGP key on www.esat.kuleuven.be/~mmoll/public_key.pgp
> Fingerprint is
> 90C4 B47D 1A00 5AC1 9147 3197 EDA7 1E0E 99E4 9EDB
>
>
>
--
Rohit Garg
http://rpg-314.blogspot.com/
Senior Undergraduate
Department of Physics
Indian Institute of Technology
Bombay