Skip to main content
gamedev.net
🔒 Locked

geometry bottleneck

Started by glNoname3f Aug 8, 2002 at 2:57 PM 1 replies 550+ views
Original Post
glNoname3f
glNoname3f
I have little problem rendering one thing, from interleaved array. I use some glDrawElements with trianglestrips. Very slow... There was many shared wertices, so I have decided to use glLockArraysEXT function. Surprisingly there is no performance gain. No single fps! It makes me crazy. Please help me. Please... Please.. Please.
weasalmongler
weasalmongler
How many triangles are u sending to the card?
Are u doing a lot of texture binds as they are expensive?
What spec PC do u have?

We need more info before we can answer.

- Weasalmongler
glNoname3f
glNoname3f
quote:
Original post by weasalmongler
How many triangles are u sending to the card?
Are u doing a lot of texture binds as they are expensive?
What spec PC do u have?

We need more info before we can answer.

- Weasalmongler


My PC: cpu AMD K6 400, g. card Matrox G200 8 MB, RAM 128
No T&L, cpu makes geometry stuff.
It''s slow, so I hoped vertex sharing would help.
My card should fill about 1000Ktriangles/sec (benchmarks say so).
I tried with diffrent number of triangles per glDrawElements call and without texturing. Framerate is below 13fps when my object has 12000 triangles or more.
Every glDrawElements shares one row of vertices with another one call. It''s a waste of time to recompute them.
I would be grateful if someone told me how to cache those vertices... please.

Topic Locked

This topic has been locked by a moderator. New replies are not allowed.

Sign in to reply to this topic.