(12/16/2001)

The Test Chamber is a simple room created in WorldCraft, intended for the testing of various mapping techniques and the effect they have on r-speeds. The idea is to have a controlled environment that can be replicated on other PCs, where variables can be isolated and tested one at a time.

The chamber is a simple cube, 384x384x384, centered exactly on the WorldCraft x-y-z grid. The cube has been hollowed with walls of 64 thickness, resulting in an interior room of 256x256x256. The walls have the CRETE4_WALL01B texture (128x128), the floor has CRETE_FLR03 (128x128), and the ceiling has CRETE3_CEIL03 (64x64). All are initially left with the default settings. An INFO_PLAYER_START inside the room completes the chamber.

Download a copy of the Test Chamber for yourself.

These experiments are being conducted on a Pentium 233MMX PC, with 96MB of RAM, and a Voodoo3 2000 Video Card with 16MB of RAM. (I told you my PC was ancient!) . I'm running in Open GL mode, at 800x600 resolution. I used Zoner's Half-Life Tools to compile.

Essentially r-speeds are a set of statistics that Half-Life can provide to tell you what your video card is up against, and how well it is dealing with it. The statistics are:

• FPS - Frames Per Second - or how many times per second the screen refreshes the image.

• MS - MilliSeconds - or the time it takes to refresh the screen once.

• WPOLY - World Polygons - or the number of brush based polygons the image is constructed of.

• EPOLY - Entity Polygons - or the number of entity based polygons the image is constructed of.

FPS and MS are measures of how well your particular computer is dealing with creating the images on the screen. Ideally FPS should be kept above 20, if it drops below about 10 you'll start to notice some choppiness in the action. Hardware plays a role in FPS and a good video card may give you acceptable frame rates where an another, less well equipped PC will have lousy ones.

WPOLY and EPOLY provide a better picture of how your level will play on PCs in general, and these are what most people are talking about when they discuss r-speeds. Opinions vary, but generally you want WPOLY to peak at less than 800 (less is always better), and from that deduct 80 WPOLY for every 1000 EPOLY you have. Sure, if you have good hardware much higher values will work fine on your PC, but if you intend to post your creation you need to think about folks with ancient equipment too. Believe me, the reviewers will ping you big time if you don’t!

1. Texture Scale 1 - how texture scale effects r-speeds when applied to large objects.

2. Texture Scale 2 - how texture scale effects r-speeds when applied to small objects.

3. Texture Alignment - Is it important to r-speeds?

4. Mitered Joints - do mitered brush joints make any difference?

5. Brush Contact - does leaving a tiny gap between brushes have an effect?

6. Func_Wall - can making one brush a Func_Wall reduce r-sppeds?

7. Epoly - what is the relationship between epoly an wpoly?

8. Texture Scale 3 - a new look at how Texture scale effects r-speeds.

9. Floors an Ceilings in a Round Room - Did it make a difference whether the floor/ceiling was inside the room or on top? Is Firebinder on drugs?

10. Texture Scale 4 - Do texture scales of less than 1.00 adversely effect FPS when applied to small objects?

Conclusions - for those who just want to cut to the chase.

Return to Top