I've read about pros and cons when dealing with many small vertex buffers rather than one large one. Most of the things I've read were concerning optimizing rendering speed. But in my case, I don't think that is an issue. What concerns me is the amount of memory that may become wasted due to padding issues or other such things. Is this something I should be worried about? That's all I need to know, but here's more specifics: I have somewhere between 1 to 16 seperate buffers (layers) per terrain section. Each layer is 33x33 vertices, but rarely will a single layer contain all of the vertices. The vertices are also very small and simple (32 bytes). So the size of a single layer is always less than 34kb, most often much less. I doubt I'm going to get any kind of speed boost by storing all layers of a terrain section in the same buffers, because I render layers based on textures. Rarely would two layers of one section be rendered in sequence, because they all use different textures. That means the stream source is going to keep changing, regardless. Thanks for any information.

Thanks for the suggestions, much appreciated.

Quote:Original post by jollyjeffers
Much of this is down to the classic 80:20 rule.... "80% of your time is spent in 20% of your code". Invest your time where it'll make a difference.

I don't have a clue how much memory may actually be used to align the buffers, but it seems more logical to me to keep the buffers seperated. So I'm not trying to optimize, I just want to stay logical without being wasteful. I'll have to invest the same amount of time, either way I choose to write the routine. But I definitely don't want to go back and rewrite it later.

80% of my time is spent procrastinating on GameDev
Thanks again [smile]

Quote:If you know the size required by your buffers (e.g. sizeof(vertex)*count) then you can subtract this from the values PIX sends out to determine how much wasted space there is.

At the risk of sounding dumb, what is PIX? [smile]

Quote:I could be wrong, but having lots of small entries, each wasting a little bit in alignment could be worse than 1 large buffer wasting a bit more.

Why would a large buffer waste more space? You mean if I had to leave empty buffer space in order to combine them all? If so, that's not the case. I could cram all of the layers into a single buffer with zero wasted memory. But I don't know how much space the card will waste. I don't really want to bother with it if the amount of space required to align the data is really small (like less than 1000 bytes). Can this information be requested from the device? I'll look and see.

Thanks man

PIX is the performance analysis tool for DirectX. It was originally an Xbox tool but it's available on Win32 now too, if you've got a reasonably recent version of the SDK. You can find it in your "DirectX Utilities" folder. You'll need to read the docs to learn how to use it, but it's a handy thing to have around.

-Mezz

Ahhh, I see. I rarely use the Start Menu, so I've never seen it before [smile]

Thanks :D