getfem-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Getfem-users] High-level assembly


From: Torquil Macdonald Sørensen
Subject: [Getfem-users] High-level assembly
Date: Tue, 05 May 2015 17:35:50 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0

Hi!

I'm experimenting with integration on faces using the high-level assembly. I was not expecting that the following two expressions would give different results when integrating on a single face in a FEM_QK(2,1) mesh:

Case 1: "Test_u*Grad_u.Normal"

Case 2: "Test_u*Normal.Grad_u"

For all the details, please see the attached test program which shows that these expressions give different results in the case of integrating along face 0 in convex 0, in a 2x2 FEM_QK(2,1) mesh. I expect Case 1 to be the correct behaviour, because its results coincide with the result of selecting the x-component of Grad_u explicitly (since the outward unit normal for the face in question is [1,0]):

Case 3: "Test_u*Grad_u(1)"

So unless I have misunderstood, something wrong happens in Case 2. I'm using a very recent SVN-version of Getfem++ and the latest GCC 5 snapshot.

Best regards,
Torquil Sørensen

Attachment: bug.cpp
Description: Text Data


reply via email to

[Prev in Thread] Current Thread [Next in Thread]