| Re: [eigen] Machine precision<> too coarse |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] Machine precision<> too coarse
- From: Benoit Jacob <jacob.benoit.1@xxxxxxxxx>
- Date: Tue, 3 Feb 2009 19:22:26 +0100
- 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=YxZ2eJ7zRJaNU1cflY/lAuYB3B8Z9OQoE1V+/jAYIV0=; b=FdQNQQmd+8u6afG9QrYloLLaBPrzfrgIy7sZq/L4ULCRsAB+wMAWvXi0NI/iWQzL5m 1IdKe9dAl7ejWd2pwTP8SvI6w/9J+QYJHz2RsBjdK5jrm1hQHlNubH1iDUslliyA395U 0uJF3fgOogr3ZkZ9Vg0heuPIEjrOYO+4bA0Xs=
- 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=xZnjDudWvWsPPz1TqGFpXPURP1yZAxRVadpAeVcQToXmCJLPB8B+V3kxu1zy32h8Az Tm9VJbRRfLglYQGOzTl+bnisy4dXoBLnXzIRr28RiB4xYk7M3a7Qwa+avBEfAUjTsNxx G55BkrmkXwjqp9pzB2YVBH9QvYvIRmZT6CxLU=
2009/2/3 Benoit Jacob <jacob.benoit.1@xxxxxxxxx>:
> And this is being very very sensitive to the context. For example, as
> I said, the partial pivoting LU gives 100x more imprecision than the
> full pivoting LU,
One more thing: this 100x difference was with double. With float, (to
my relief) the difference is much smaller, more like x2.
This also suggests that a general formula depending on machine epsilon
would look very complicated. Simple formulas like the one you pasted
above don't account for this kind of phenomenon.
Benoit