Beyond3D Forum

phed · 09-Oct-2003, 07:20

Hello, are there any benchmarkes out there that directly tries to measure the eg. polygons/sec & pixels/sec under different renderingmodes?
Also that tries different rendering techniques, and preferably with different factors (like vertexarray() with a lot of polygons and many times with small polygons).

Two reasons: To understand what codingpractices are good, and to understand what cards are good at the stuff I want them to be good at. (Eg. I favour dynamic models rather than on-board memory models).

I made some of these, and some of the numbers surprised me; eg. how the number of polygons increased by 33% when the framerate was lowered.
But I would hope that someone with more background in 3d-programming than me would be more suited than me for this.

phed · 10-Oct-2003, 06:35

Hay, I found this one: http://developer.nvidia.com/object/BenMark5.html

It supposedly gives 24MTris/s on a GF2GTS. My 2GHz w/GF3Ti200 ticks in at ... 24MTris, and it is NOT busbound.

Why? What is your number?

zsouthboy · 10-Oct-2003, 06:55

I would think that different rendering techniques, as you say, like vertexarray versus calling drawprimitive(i think that's right) many times would be dependant on the driver. If the same polys are passed to either, it should be the same. In the real world, though...

EDIT: Ran that bench, got 77mtris/s on my Soft9700

09-Oct-2003, 07:20	#1
phed Junior Member Join Date: Aug 2003 Posts: 66	A proper benchmark Hello, are there any benchmarkes out there that directly tries to measure the eg. polygons/sec & pixels/sec under different renderingmodes? Also that tries different rendering techniques, and preferably with different factors (like vertexarray() with a lot of polygons and many times with small polygons). Two reasons: To understand what codingpractices are good, and to understand what cards are good at the stuff I want them to be good at. (Eg. I favour dynamic models rather than on-board memory models). I made some of these, and some of the numbers surprised me; eg. how the number of polygons increased by 33% when the framerate was lowered. But I would hope that someone with more background in 3d-programming than me would be more suited than me for this. __________________ weblog

10-Oct-2003, 06:35	#2
phed Junior Member Join Date: Aug 2003 Posts: 66	Hay, I found this one: http://developer.nvidia.com/object/BenMark5.html It supposedly gives 24MTris/s on a GF2GTS. My 2GHz w/GF3Ti200 ticks in at ... 24MTris, and it is NOT busbound. Why? What is your number? __________________ weblog

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Stand-alone benchmark and games with same engine	Reverend	Beyond3D News	9	24-Mar-2005 17:50
Game benchmarks vs benchmark applications	Patric Ojala	3D Architectures & Chips	0	16-Oct-2003 13:38
Gun Metal benchmark	Reverend	PC Games	5	24-Jun-2003 08:44
Benchmark Life Cycles	Chris	3D Architectures & Chips	11	04-Jun-2003 22:38
UT2003 Demo Benchmarking Information	Brent	PC Games	19	07-Oct-2002 20:52

10-Oct-2003, 06:55	#3
zsouthboy Member Join Date: Aug 2003 Location: Derry, NH Posts: 563	I would think that different rendering techniques, as you say, like vertexarray versus calling drawprimitive(i think that's right) many times would be dependant on the driver. If the same polys are passed to either, it should be the same. In the real world, though... EDIT: Ran that bench, got 77mtris/s on my Soft9700

Beyond3D Forum

Welcome, Unregistered.