PDA

View Full Version : Performance and Vertex Array Range



budi
08-20-2002, 03:52 PM
Hi,
I have several questions to ask you:
1. What is the highest triangle rendering rate that you have been able to achieve using your OWN OpenGL code (with Geforce4)?
I have only been able to achieve 11.5 Million triangles per sec with Vertex Array Range (with video memory). The machine is a dual Athlon 1.2 GHz, 2 GB memory and
Geforce4 TI4600. No triangle strips.
This seems low for a Geforce4 right?
2. I can't seem to allocate AGP memory. I tried the Nvidia demo AGP memory program that returns how many AGP memory available in the system, and it returns 0.
Is there a special patch that I have to install?
3. Have anyone been able to use Vertex Array Range under Linux (Red Hat 7.2)? glXAllocateMemoryNV always return NULL for me. I tried allocating AGP memory and video memory. It doesn't matter.
We have installed the new nvidia driver (version 2960).

Thank you,

Budi

jwatte
08-20-2002, 07:10 PM
2) is directly related to 1).

You probably need to get decent AGP drivers for your motherboard. If it's a VIA chip set, try getting the latest 4-in-1 from off the VIA web site (www.via.com.tw) and install that. If some other vendor, then go to their site.

I've even had the case where a friend bought an Athlon machine from Micron, and it came loaded with the wrong AGP drivers. He thought I was crazy when I told him his brand new machine was a piece of crap and needed an upgrade :-)

zed
08-20-2002, 07:38 PM
20.5mil now on a gf2mx http://www.opengl.org/discussion_boards/ubb/smile.gif it just wont go any higher scotty

>> I can't seem to allocate AGP memory. I tried the Nvidia demo AGP memory program that returns how many AGP memory available in the system, and it returns 0.<<

i used to have the same problem with my celeron + motherboard the solution was as jwatte suggested (if u have a via motherboard) get the new 4-1 divers