| Re: [eigen] Mapping array of scalars into quaternions |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] Mapping array of scalars into quaternions
- From: Benoit Jacob <jacob.benoit.1@xxxxxxxxx>
- Date: Thu, 12 Nov 2009 11:36:26 -0500
- 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; bh=rDFCCk4hjsEWCUMJdBG3YghkpW8jQjUk7EgwZOEpWI8=; b=HQh6WE7zwYCuuXk0eCMTvsplkKfukxVYBQOzw3yDMMXH2xdV7vByyNMnanUVT/CWft eOvfwq2utMhtdKdb8h++ngUvoAZtd6HFpadpHu9XkE+migARhEuXBUFtgVKY77xHn1F8 gSLwGj8LfaVSQTS2/YVkTj3B0einpGDCZpthw=
- 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; b=Pc8FoqR0DmfLpi0fqyYFwFITUuedc+sfv2JaRIMKtNzru6eGC4pmWGKGsITMtRj+Ld ul6NWVem+IvchgmrdEXgLc4p3v+uk7nVJRdlWRbvXchhb62mEowZ0qrLgeIDJgVyORMq 9ggdFdrWncZW4ZuhXZ68J9IrDjGM60wjHxi/w=
2009/11/12 Mathieu Gautier <mathieu.gautier@xxxxxx>:
>
>> and does inlining the cross product result in a significant
>> performance improvement?
>
> I have not tested on my former code yet. I think that the difference will be
> very small. I am just comparing the generated code with the one generated
> with my previous math library.
ok. if there is no significant performance improvement, why bother
with EIGEN_STRONG_INLINE at all? In the long term, it is safest / most
comfortable to just let the compiler decide when to inline, possibly
using the "inline" keyword to give it a hint that it most probably
should.
The typical use case for EIGEN_STRONG_INLINE is with trivial
functions. Indeed Eigen uses very, very deep chains of trivial
function calls, in which case inlining becomes crucial, and compilers
may fail to understand that as that isn't a very usual situation for
them.
That's why i'm asking for some justification for EIGEN_STRONG_INLINE here.
I'll look at your patch as soon as possible today.
Benoit
>
> --
> Mathieu Gautier
>
>
>