|Re: [eigen] alignment question|
[ Thread Index |
| More lists.tuxfamily.org/eigen Archives
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] alignment question
- From: Sylvain Pointeau <sylvain.pointeau@xxxxxxxxx>
- Date: Wed, 28 Jan 2015 22:38:07 +0100
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; bh=9P13vDlgVYF3C241P4Mr+rM52o0ywQS1HOgOsGm0aQ8=; b=t7ogLvzQOQEhocstRaj4diSPbrLGDKEZjuSl1ikH883ekNV6GyTKjYqsMBVBdWlmPS p2lQJlyMj2ICHsAwopzMJgEGueTt5Or5WgdgcSDL56RlG7j7tiLnlrKE0N9Ey2hjLOUp Mf14VwaX5h9ewZupBpx9JMxe8r88TUaBftnvFr00LnwI4wVie0ALBkNPXbvQrPDlGauJ Km3a7C+kCq+DtBeyRNQ1ifA0/4QRd9MzNZRRfMtbjPqnnBatNsHha614cIYpYnMRAnIA +ksCf5DM0CxsN35Gea7OFGzK7+nokhVat1HPJUNRZlTF+h3LO6iSVIaHMtLNco2WqkLo ljKQ==
I just tried to make a reduced example with no luck, then you are probably right, I might had a different bug that I solved by creating temp objects.
if the issue reappears, I will try first the compiler options -mstackrealign or -mstack-alignment=16 to see.
@Gael, I would like to move away from Vector3 to enable the vectorization, I will have a lot of calculs, better to have everything as vector4 so I don't have any conversion later on. On a negative side, I need to be careful if w() is 1 (Point) or 0 (Vector)...