I'm a small-time YouTuber, and I use the Blender VSE to edit / render my vid's. I'm currently running this on my laptop, but—especially if I apply transformations or multiple over-drop layers—rendering is slooooow! Would getting a box with a supported GPU help me get shorter render times? (And, is there a list of supported / recommended GPUs somewhere?) To be clear, everything I'm doing is 2D and via the VSE. I have a background in computing (so I'm not frightened by matrix transformations or technical terminology), but I don't understand what Blender's capabilities are vis-à-vis using a GPU to do 2D compositing.
While we're at it, it seems that if I, say, chain ten transform layers it's substantially slower than applying one … Why is this‽ I.e., I'd imagine that you'd multiply all of the transform matrices together once, cache that, and then apply it to each frame (in which case N transforms ought to take no longer than 1. To be more technically precise, I would expect that applying the same N transformations to M frames would take O(N + M) but it, empirically, seems to take O(N * M).).
//cheers
~K.D.P.Ross
