Conversation
"Geometric algebra for computer science," by Dorst et al. As well as "Linear and geometric algebra" by Macdonald
5
1
4
…but would the performance cost vs traditional linear algebra be too prohibitive?
1
@bjzaba_ there is actually no performance cost if you implement it properly (ie. static types, no dynamic arrays).
1
How about space issues? Aren't GA data structures meant to be pretty heavy in that respect?
1
@bjzaba_ There is no need to treat everything as a multivector. Just have seperate types for vector/rotor/etc..
2
1
@bjzaba_ I use 4D rotors. Also slowly replacing all the quaternions with 3D rotors.
1
Replying to
Awesome! It'd be great to see a post on it if you can ever manage it.