Productivity and waste

OpenGL OpenGL: Advanced Coding
1

Greeting all…

I have a job interview in two weeks time and want to get my engine (cough) up to par within that time to make a good impression.

For dynamic geometry I’m currently using NON-indexed vertex arrays. This works fine. But I can’t help but think I’m wasting alot of memory/speed this way. I’m currently trying to optimise for GeForce so VAR is the way to go.
But would using an INDEXED array introduce a slowdown? I ask because, as far as I know, the index array has to be kept in system memory.
I’m sure it would speed things up on generic hardware, but (in VAR) I’m curious as to whether the CPU is able to send the index data to the HW fast enough to not be a bottleneck.

in short… Is it worth the effort?

cheers.

2

If you have vertex reuse, indexed data is a huge win for VAR.

Make sure to reuse vertices quickly – if you space out the uses of a vertex by too much, the vertex cache hit rate will be decreased.

  • Matt
3

Thanks for the quick reply Matt.
I’ll give it a go and post the results here later.

Cheers.