nV News Deals Shop Archive Search Files Forum Feed Articles IRC Chat GeForce.com


Search Site
Ads by Google
Links To NVIDIA
Drivers
Products
Communities
Support
NVIDIA Blog
News Room
About NVIDIA
GeForce Technology
CUDA
DirectX 11
Optimus
PhysX
SLI
3D Vision
3D Vision Surround
Articles
GeForce GTX 580
GeForce GTX 570
GeForce GTX 560 Ti
GeForce GTX 480
GeForce GTX 465
GeForce GTX 460
GeForce GTS 450
GeForce GTX 295
GeForce GTX 280
GeForce GTX 260
GeForce GT 240
GeForce 9800 GTX
GeForce 9800 GX2
GeForce 9600 GT
GeForce 8800 Ultra
GeForce 8800 GTX
GeForce 8800 GTS
GeForce 8800 GT
GeForce 8600 GTS
GeForce 8500 GT
GeForce 7950 GX2
GeForce 7950 GT
GeForce 7900 GTX
GeForce 7900 GS
GeForce 7800 GTX
Watercooling Project
My Book 500GB
Raptor Hard Drive
Guide To Doom 3
EVGA Stuff
EVGA E-LEET
EVGA Precision
GPU Voltage Tuner
OC Scanner
SLI Enhancement
EVGA Bot
EVGA Gear
Reviews and Awards
Associates
Benchmark Reviews
Fraps
GeForce Italia
GPU Review
Hardware Pacers
LaptopVideo2Go
MVKTECH
News3D (NVITALIA)
OutoftheBoxMods
OSNN.net
Overclocker Cafe
PC Extreme
PC Gaming Standards
PhysX Links & Info
TestSeek
3DChip (German)
8Dimensional
Gainward FX PowerPack! Ultra/1800 XP Review - Page 5 of 9

HALO

I'd hoped to do some in-game benchmarking by following a set path and using FRAPS. However, I never received consistent results. I could retrace my steps very closely (even for just a 20 second romp with no enemy contact) and yet the resulting frames per second would vary by 15fps or more. So, I used the game's built-in timedemo. It's not perfect but it's better than nothing.

Halo: Combat Evolved

In hopes to appease those of you that aren't fans of the game's timedemo, I did at least include results from all rendering paths and at two different resolutions. Below are screenshots of the video and audio settings used.

Halo Video Settings
Halo Video Settings

Halo Audio Settings
Halo Audio Settings

The results below are what you'd expect. Some may think that the PS2.0 results are surprisingly on par with the PS1.4 results. However, this is actually expected because there doesn't appear to be a huge difference between PS1.4 and PS2.0 implementation within Halo as stated in the Halo PC technical and performance FAQ:

Having said this, for many simple effects, even if you are running PS2.0, the game will automatically use a 1.4 or 1.1 shader because the visual result is exactly the same.

So, that's why there's no difference between the PS2.0 and PS1.4 shader path scores. Except for some specific cases there really is no glaring visual difference between the two either. Note: I dropped my overclocking down to 540/1020 as 555/1050 brought about some artifacting but no crashing of the game.

Halo Benchmark Results @ 540/1020
Halo Benchmark Results @ 540/1020

As you can see, there is roughly a 50% performance hit between 1024x768 and 1600x1200 in all but the FF path where that delta is closer to 15%.

The following is from one of Mike's previous reviews. I include it here for easier reference for those of you that may not be too familiar with the rendering paths.

Microsoft has published a Halo PC technical and performance FAQ, which contains information on the games rendering features, performance tuning, and timedemo mode. Halo supports rendering paths that make use of DirectX 9, DirectX 8.1, DirectX 8, and even DirectX 7. Here an excerpt from the FAQ:

Halo supports 4 different rendering code paths:

Pixel Shader 2.0 (DirectX 9.0)

In this code path, you are making absolutely no compromises on the visual quality of the game. You are seeing everything as best as possible, as engineered by our team. All the effects are in their most demanding form (as complex of a calculation as necessary to generate the best visual result possible).

Having said this, for many simple effects, even if you are running PS2.0, the game will automatically use a 1.4 or 1.1 shader because the visual result is exactly the same.

Pixel Shader 1.4 (DirectX 8.0)

When running in PS1.4, you are compromising only a subset of effects. Specifically:

  • No bumped mirrored surfaces
  • Some video effects are two-pass

Pixel Shader 1.1 (DirectX 8.0)

PS1.1 is probably the most widespread pixel shader version currently. When running in the PS1.1 rendering code path, the visual compromises are (in addition to the PS1.4 compromises):

  • No model self-illumination (excluding some specific environmental models)
  • No animated lightmaps
  • Fog calculations are triangle based, not pixel based
  • No specular lights

Hardware T&L (Fixed Function - DirectX 7.0)

This is the most basic rendering code path for Halo. When running in that mode, you have to accept many visual compromises but are still getting a compelling Halo visual experience. The compromises are:

  • No shadows
  • Simple active camouflage effect
  • No glows or flares
  • Very basic fog, water and lighting

Note that you can always scale down your video card but you canít scale up. Check your video cardís documentation to determine witch version of DirectX / pixel shaders it supports. If, for example, your card supports PS2.0, you can choose to compromise many of Haloís effects by forcing it to run PS1.1. However, if your card supports PS1.1, you canít force it to PS2.0 Ė this is actually happening in hardware, not in software.

The Halo timedemo can be run in each of the four rendering modes by adding a command line variable to the shortcut to the Halo executable. For example:

E:\Games\Halo\halo.exe -use20 -nosound -vidmode 1280,1024,72 -timedemo

This string of command line variables will launch Halo in timedemo mode with sound disabled, use pixel shader version 2.0, at resolution of 1280x1024 and a 72Hz refresh rate. The shader option for pixel shader 1.4 is -use14 and for pixel shader 1.1 is -use 1.1.

SCREENSHOTS

Below are a few composite screenshots that show some interesting comparisons to the four rendering paths. There is an enormous difference between the Fixed Function (DX7) path and the other three paths.  In fact, the differences between PS1.1, PS1.4 and PS2.0 are negligible at best. Granted, Halo doesn't heavily implement PS2.0 effects but based on what we see here there's really no difference.

This first composite is from slightly different angles but they are close enough to show comparisons. You can see drastic differences in the bump mapping on the bridge, overall lighting effects and the water in the background.  1600x1200 PNG ~ 2.63MB

Halo Rendering Path Comparison 1 of 2
Halo Rendering Path Comparison


This next composite is all from the exact same angle. Lighting is also a major difference here. The DX7 part of the composite has completely different colors as a result and is just drab in appearance overall. Another big difference is that of the smoke/fog effects in the middle ground. Notice how the DX7 version renders the smoke as much more opaque whereas the other paths show its more accurate translucent appearance. The final point of interest here is that of the red light in the ceiling. In the game this light pulses so that could be why the DX9 path appears brighter since the screenshot could not be taken of a specific frame in each case. Despite this, that light did appear to be "cleaner" in the DX9 path than it did in the other paths.  1600x1200 PNG ~ 1.22MB

 

Halo Rendering Path Comparison 2 of 2
Halo Rendering Path Comparison

 

Next Page: Gameplay Testing - Unreal Tournament 2003

Last Updated on January, 2004


Table of Contents

Advertisement

nV News - Copyright © 1998-2014.
Search Products
Search
for


Ads by Casale
Tweaks
Metro: Last Light
PlanetSide 2
Miscellaneous Links
AutoDesk 123 Design
Build Your Gaming PC
FPS vs. Frame Time
Free Games And MMOs
GeForce SLI Technology
HPC For Dummies
PC Game Release Dates
Play Classic PC Games
Steam Hardware Survey
Video Game Designers
TechTerms Dictionary
GPU Applications
AMD GPU Clock Tool
AMD System Monitor
ATITool
aTuner
EVGA E-LEET
EVGA OC Scanner
EVGA Precision
EVGA Voltage Tuner
Gainward ExperTool
GPU-Shark
GPU Voltage Tuner
Fraps
FurMark
GLview
GPU Caps Viewer
GPU PerfStudio
GPU Shark
GPU-Z
MSI Afterburner
nHancer
NiBiTor
NVClock (Linux)
NVFlash
NVIDIA Inspector
NvTempLogger
NVTray
PowerStrip
RivaTuner
SLI Profile Tool
The Compressonator
3DCenter Filter Test
3DMark 11
3DMark Vantage
PhysX Applications
Cell Factor Revolution
Cryostatis Tech Demo
Cube Wall Demo
PhysX FluidMark
Fluid Physics
NV PhysX Tweaker
NVIDIA OPTIX 2
PhysX Downloads
PhysX at YouTube
Add-In Partners
AFOX
ASUS
AXLE
BFG Technologies
BIOSTAR
Chaintech
Colorful
ELSA
emTek
EVGA
GAINWARD
GALAXY
GIGABYTE
FORSA
FOXCONN
Inno3D
Jaton
Leadtek
Manli
MSI
Palit
PNY
Point of View
Prolink
SPARKLE
XFX
ZOGIS
ZOTAC