19 users logged in
Proud partner of GDC 2025
Before posting, review our community guidelines.
Support GameDev.net with a monthly GDNet+ subscription!
How to get the transpose of 3x3 matrix?
June 27, 2005 05:43 AM
Hi I've got a 3x3 matrix and Im suppose to get the transpose for other calculations. Except I'm not really sure what the tranpose of a matrix is. So what would be the output of a transpose of a matrix? eg if ihad the values in the matrix like so: m11, m12, m13 m21, m22, m23 m31, m32, m33 Thanks
140
June 27, 2005 05:49 AM
Transpose matrix is simply:
m11, m21, m31
m12, m22, m32
m13, m23, m33
switch rows with columns, and that`s it...
m11, m21, m31
m12, m22, m32
m13, m23, m33
switch rows with columns, and that`s it...
Bulma
124
June 27, 2005 11:46 AM
in code:
public static int [] [] MatrixTranspose ( int [] [] A ){ int [] [] TransposedMatrix = new int [ A.length ] [ A[0].length ]; for ( int a = 0 ; a < A[0].length ; a++ ){ for ( int b = 0 ; b < A.length ; b++ ){ int temp = A [ a ] ;<br> A [ a ] = A [ a ];<br> A [ a ] = temp;<br> }<br> }<br> <span class=cpp-keyword>return</span> TransposedMatrix;<br> }<br><br></pre></div><!–ENDSCRIPT–><br>Hope this helps
(0110101101000110)The Murphy Philosophy: Smile . . . tomorrow will be worse.
500
June 28, 2005 02:38 AM
Here is the function I use to transpose matrices in my math lib.
inline void matrix::transpose(void){ matrix T = *this; M12=T.M21; M13=T.M31; M21=T.M12; M23=T.M32; M31=T.M13; M32=T.M23; }
767
June 28, 2005 03:20 AM
here's mine.
3x3 transpose is super-useful in SSE for computing simultaneous dot-products and having the result set up for another parallel operation (like rsqrtps for doing parallel vector normalizes). So this is a good one to remember!
(P.S. If anyone has a faster way, I'd love to hear it. This uses 5 registers, and so destroys one of the inputs. Still, it's a great problem for people that are into this sort of thing).
// input xyz in xmm5,7,6// output in xmm0,1,7// ?? == don't care. // xmm5 : ?? z0 y0 x0 // xmm7 : ?? z1 y1 x1 // xmm6 : ?? z2 y2 x2movaps xmm0, xmm7 // xmm0 : ?? z1 y1 x1movaps xmm1, xmm5 // xmm1 : ?? z0 y0 x0unpcklps xmm0, xmm5 // xmm0 : y1 y0 x1 x0unpckhps xmm7, xmm5 // xmm7 : ?? ?? z1 z0movhlps xmm1, xmm0 // xmm1 : ?? z1 y1 y0shufps xmm7, xmm6, 11100100b // xmm7 : ?? z2 z1 z0movlhps xmm0, xmm6 // xmm0 : ?? x2 x1 x0shufps xmm1, xmm6, 01010100b // xmm1 : ?? y2 y1 y03x3 transpose is super-useful in SSE for computing simultaneous dot-products and having the result set up for another parallel operation (like rsqrtps for doing parallel vector normalizes). So this is a good one to remember!
(P.S. If anyone has a faster way, I'd love to hear it. This uses 5 registers, and so destroys one of the inputs. Still, it's a great problem for people that are into this sort of thing).
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
Recommended Tutorials
Advertisement