Hi, I know what CVAs are, but can somebody tell me what VARs are and does anybody have a good SIMPLE tutorial on VARs. I know that they are a form of vertex arrays, but I would like some more detail and how to implement them, and what their advantage is over CVAs. Thanks

VAR is only geforce class gfx cards I think and yeah they are similar to CVA but faster.

Similar in what way?

With ordinary (and CVA’s) vertex arrays. You hold your vertex data in system memory. Each time you render, the driver first copies all the data into AGP memory, then tells the card to render it.

With VAR, you allocate AGP (or video) memory for your vertex data straight away, to remove the copy process before the render. This is why it speeds it up. You can use Fences on top of this, so you can find out when you can start updating the array again, without having to wait until the GPU is totally finished.

You can technically use VAR with system memory, but this will slow down the GPU loads when it tries to pull all the vertex data across the main system bus.


Maybe it’s done differently, but my radeon has a VAR extension, I do believe.

Thanks, but you still havent told me where I can get some tutorials for VARs, I have a geforce256 so I should be able to use VARs…

you’ll find some demo with source on VAR here :

good luck!

Is your Radeon T&L?

I presume it is, cos it doesn’t make sense to have VAR on a non T&L board.

This is good though! Especially if it operates exactly the same as Nvidia one!