Vectorize the covariance calculation

[kemchenj]

I confirm I have read the contributing guidelines and signed the Contributor License Agreement.

[mvaligursky]

nice, but even cleaner would be

    vec3 M0 = scale.x * rot[0];
    vec3 M1 = scale.y * rot[1];
    vec3 M2 = scale.z * rot[2];

    covA = vec3(
        dot(M0, M0),
        dot(M0, M1),
        dot(M0, M2)
    );

    covB = vec3(
        dot(M1, M1),
        dot(M1, M2),
        dot(M2, M2)
    );

[kemchenj]

Thanks for the suggestion! I've implemented the change and it looks much cleaner now. Appreciate your feedback!

[slimbuck]

Hi @kemchenj , did you mean to close this PR? Looks like a useful simplification.

[mvaligursky]

I merged this and it looks like this now (not correct)

[kemchenj]

Hi @kemchenj , did you mean to close this PR? Looks like a useful simplification.

Hi, yes, I did intend to close the PR because I realized the modification here was not correct. I was planning to fix it before reopening the PR, but it seems it has already been merged...

[mvaligursky]

Hi.. I'm happy to accept a new PR that fixes it :)
Thanks!

[kemchenj]

Hi.. I'm happy to accept a new PR that fixes it :) Thanks!

I opened a new PR to fix this #6100 🫣