Re: [eigen] EIGEN_INITIALIZE_MATRICES_BY_NAN? |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] EIGEN_INITIALIZE_MATRICES_BY_NAN?
- From: Gael Guennebaud <gael.guennebaud@xxxxxxxxx>
- Date: Thu, 19 Apr 2012 22:34:04 +0200
- Cc: Rhys Ulerich <rhys.ulerich@xxxxxxxxx>
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc:content-type:content-transfer-encoding; bh=zZEd+T4w6bU4ZhxJqj0IRYPY3U4MmA4Ry8JJNzjBsgo=; b=U7euLm9HVATeYYPdoTeHmKxdDwyOLsE7mFs8kkbsKUCqpMXt5dOZlDx7tPH+Ilu5Qh KSFU4JpVzZOpSkl0eI5oV/vM1Nu+IZMxKQLFJS1pJWF3TsnFOiVXkWilzN5EjFxF6tij ed1ZNuimJ+efePheVLtQuDlHuxCcxU5JEJE313syQA201Ez4/uTWlxHmlpLDSPwt6OXU BwZJ/qqWKGAV3zJQS5Jo7LjFaE/tYaLFDVkZNYgpFGS5qtmlBgr3AasJTnISdD2Bl339 xrXukNBfaWzqeKb6y/RdPNxrpwJwXidARoHl0UVKWgbY9F+ACPueCMRlJkxRoMghJiZk tzbw==
Is valgrind not enough to detect use of uninitialized values?
In my opinion we already have too many compilation options that
changes the behavior of Eigen, and I'd like to keep their number to a
minimum.
gael
On Thu, Apr 19, 2012 at 9:37 PM, Radu B. Rusu <radu.b.rusu@xxxxxxxxx> wrote:
> +1. That looks like a useful feature.
>
> Cheers,
> Radu.
>
>
> On 04/19/2012 12:35 PM, Rhys Ulerich wrote:
>>
>> Any thoughts on having an option like
>> EIGEN_INITIALIZE_MATRICES_BY_ZERO called
>> EIGEN_INITIALIZE_MATRICES_BY_NAN which does what it suggests? It'd be
>> handy for tracking down algorithmic problems due to uninitialized data
>> being used. Only hiccups I can see would be initializing the real-
>> and imaginary portion when NumTraits<Scalar>::IsComplex holds and/or
>> using numeric_limits<NumTraits<Scalar>::Real>::has_quiet_NaN to
>> determine when it could be performed at all.
>>
>> - Rhys
>>
>>
>
>