quick question about matmul
From: Benjamin (snowkid_23_at_yahoo.com)
Date: 10/30/03
- Next message: John Harper: "Re: Something strange with g77 ..."
- Previous message: Dan Nagle: "Re: unsigned byte intrinsic type"
- Next in thread: Daniel Grimwood: "Re: quick question about matmul"
- Reply: Daniel Grimwood: "Re: quick question about matmul"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 30 Oct 2003 12:51:43 -0800
More of a verification rather than a question:
I have 3 4x4 matrices that I want to multiply together:
Conceptually:
Z = L2*P*L1 ! Z, L2, P, L1 are 4x4
Using MATMUL, I implement it like this:
REAL(8), DIMENSION(4,4) :: Z,L1,P,L2
...
Z = MATMUL(MATMUL(L2,P),L1)
Is this indeed doing what I want it to do? It gives numerically
reasonable results, so I just wanted to see if anyone has comments
about doing it this way vs. some other method unknown to me.
The reason I have any doubt is that I'm using a completely different
technique to arrive at the result Z (a different mathematical
technique)
and it's Z is the transpose of the result I receive using the MATMUL
method.
Regards,
-Ben
- Next message: John Harper: "Re: Something strange with g77 ..."
- Previous message: Dan Nagle: "Re: unsigned byte intrinsic type"
- Next in thread: Daniel Grimwood: "Re: quick question about matmul"
- Reply: Daniel Grimwood: "Re: quick question about matmul"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]