| RE: [eigen] 3.3-beta2 released! |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: <eigen@xxxxxxxxxxxxxxxxxxx>
- Subject: RE: [eigen] 3.3-beta2 released!
- From: <Daniel.Vollmer@xxxxxx>
- Date: Thu, 28 Jul 2016 09:06:02 +0000
- Accept-language: de-DE, en-US
- Thread-index: AQHR54kGyXM7+kpBoUO9NUInulp8IqAsOHm3gABRzqeAACSNAIAA3pwm
- Thread-topic: [eigen] 3.3-beta2 released!
Hi Gael,
> It is quite surprising that unaligned vectorization is slowing down
> the execution, though. What's your CPU? What if you completely
> disable explicit vectorization? (-DEIGEN_DONT_VECTORIZE)
The CPU I tested on is an Intel(R) Core(TM) i7-3615QM CPU @ 2.30GHz
(which supports AVX, but I think clang only targets SSE2 or maybe
SSE4.2 by default) in a Macbook Pro.
When I compile without Eigen's vectorization, I don't get any change
in performance, probably because for the currently implemented
equations, our matrix sizes are not helpful (5x1 and 5x5 or 6x1 and
and 6x6 as well as 3x1 for geometric information, all doubles).
Best regards
Daniel Vollmer
---
Deutsches Zentrum für Luft- und Raumfahrt e.V. (DLR)
German Aerospace Center
Institute of Aerodynamics and Flow Technology | Lilienthalplatz 7 | 38108 Braunschweig | Germany
Daniel Vollmer | AS C²A²S²E
www.DLR.de