Looking Back
Let's recap what we've seen so far. We've already looked at 7 games - Call of Duty, Richard Burns Rally, SW: KOTOR, F.E.A.R, Quake 4, Serious Sam II and Splinter Cell: Chaos Theory. Out of these 7 games, a pair of GeForce 7600GT in SLI mode can only offer a performance increase in three games, Richard Burns Rally, F.E.A.R and Splinter Cell: Chaos Theory. Based on our previous findings, a pair of GeForce 6600 can only offer a performance increase with Call of Duty, Richard Burns Rally, F.E.A.R and Splinter Cell: Chaos Theory. We also hardly see any increase with a pair of GeForce 7800GTX, except for Richard Burns Rally and F.E.A.R. In contrast, we saw an increase in Call of Duty, Richard Burns Rally, F.E.A.R, Quake 4 and Serious Sam II with both a pair of Radeon X1600XT and a pair of Radeon X1900s. The table below clearly illustrate this.| 6600 SLI with Forceware 81.98 | 7800GTX SLI with Forceware 81.98 | 7600GT SLI with Forceware 84.21 | ||||
| Call of Duty | ||||||
| Default | AA AF | Default | AA AF | Default | AA AF | |
| 1024 | Slower | Faster | Slower | Slower | Slower | Same |
| 1280 | Same | Faster | Slower | Same | Slower | Same |
| 1600 | Same | Faster | Slower | Same | Slower | Same |
| Richard Burns Rally | ||||||
| Default | AA AF | Default | AA AF | Default | AA AF | |
| 1024 | Slower | Faster | Slower | Faster | Same | Faster |
| 1280 | Slower | Faster | Slower | Faster | Faster | Faster |
| 1600 | Slower | Faster | Slower | Faster | Faster | Faster |
| FEAR | ||||||
| Default | AA AF | Default | AA AF | Default | AA AF | |
| 1024 | Faster | Faster | Faster | Faster | Faster | Faster |
| 1280 | Same | Faster | Faster | Faster | Faster | Faster |
| 1600 | Same | Same | Faster | Faster | Faster | Faster |
| Quake 4 | ||||||
| Default | AA AF | Default | AA AF | Default | AA AF | |
| 1024 | Slower | Same | Slower | Slower | Slower | Same |
| 1280 | Slower | Same | Slower | Slower | Slower | Same |
| 1600 | Slower | Same | Slower | Slower | Slower | Same |
| SCCT | ||||||
| Default | AA AF | Default | AA AF | Default | AA AF | |
| 1024 | Faster | Faster | Same | Same | Faster | Faster |
| 1280 | Faster | Faster | Same | Same | Faster | Faster |
| 1600 | Faster | Faster | Same | Same | Faster | Faster |
(compared to a single card - for example a 6600 SLI compared to a single 6600).
Things are looking very grim for SLI. Crossfire can offer a performance increase where SLI couldn't. In a game that reacts to both Crossfire and SLI, Crossfire generally offer more of an increase (except for F.E.A.R). Even the updated parts can't offer something close to what Crossfire has. Results from the use of embedded game profiles are not that different, so the existence of game profiles does not guarantee you'll have an increase in performance with SLI - it only promises that SLI will work without problems. One glaring example of this is Quake 4. Despite the presence of an embedded profile, we hardly see any difference in performance (or in this case, lack of) with SLI multi rendering.
There's also games that don't' work well with any multi rendering solutions. SW: KOTOR and Homeworld 2 fall under this category. In these games, SLI is definitely worse than Crossfire. The driver overhead is so limiting, that you'll experience a drastic fall in frame rates if you choose to force SLI multi rendering (Homeworld 2). Using the embedded application profile didn't help. In contrast, Crossfire degrades beautifully here - while there is a minimal performance increase to be gained by forcing AFR with this game, some display corruption did occur with default settings. So ATI felt it was wiser to simply disable AFR for Homeworld 2 and SW: KOTOR.
Forcing Multi Rendering
One of the advantages that SLI has over Crossfire is the ability to force SLI multi rendering from the driver panel. Crossfire users are less fortunate and must resort to renaming executable files to achieve this. For users with an SLI setup, Forceware 84.21 offer three SLI rendering mode options: SLI single rendering, SLI multi rendering and SLI anti aliasing. With the new Forceware 91.31, users can now experiment with different SLI multi rendering scheme. There are now three other options in place: Alternate Frame Rendering, Split Frame Rendering (Scissor Mode) and Alternate Frame Rendering 2 instead of one catch-all SLI multi rendering mode. The purpose is quite clear: when a new game comes out, users can force a multi rendering mode that works best with that game, at least until a new driver (or SLI profile) is available.Naturally, we're intrigued. We want to put this feature to the test and to do so, we took two games that are really friendly to multi rendering solutions: Splinter Cell: Chaos Theory and Richard Burns Rally. First of, Splinter Cell: Chaos Theory. Below are the results from the Lighthouse demo, taken with different SLI multi rendering modes. We then renamed the executable files to trick the driver from applying any application specific tweaks and re run the test - just what happens when the driver encounter a new game (that might be SLI friendly).
Splinter Cell: Chaos Theory
| Forceware 91.31 Default | |||
| Auto Select | |||
| Resolution | 1024 x 768 | 1280 x 1024 | 1600 x 1200 |
| Min | 76.61 | 58.27 | 42.81 |
| Avg | 111.66 | 95.47 | 72.39 |
| Max | 194.04 | 16.85 | 142.13 |
| Alternate Frame Rendering | |||
| Resolution | 1024 x 768 | 1280 x 1024 | 1600 x 1200 |
| Min | 76.41 | 58.1 | 42.78 |
| Avg | 111.44 | 95.5 | 72.38 |
| Max | 193.51 | 169.07 | 142.17 |
| Split Frame Rendering | |||
| Resolution | 1024 x 768 | 1280 x 1024 | 1600 x 1200 |
| Min | 76.31 | 57.98 | 42.8 |
| Avg | 111.39 | 95.47 | 72.38 |
| Max | 193.29 | 169.83 | 142.28 |
| Alternate Frame Rendering 2 | |||
| Resolution | 1024 | 1280 x 1024 | 1600 x 1200 |
| Min | 76.45 | 58.22 | 42.8 |
| Avg | 111.51 | 95.48 | 72.39 |
| Max | 193.44 | 169.68 | 142.17 |
Splinter Cell: Chaos Theory - renamed to squash.exe
| Forceware 91.31 Default | |||
| Auto Select | |||
| Resolution | 1024 x 768 | 1280 x 1024 | 1600 x 1200 |
| Min | 27.55 | 22 | 17.52 |
| Avg | 73.99 | 55.55 | 41.13 |
| Max | 150.26 | 139.69 | 133.25 |
| Alternate Frame Rendering | |||
| Resolution | 1024 x 768 | 1280 x 1024 | 1600 x 1200 |
| Min | 27.55 | 22 | 17.52 |
| Avg | 73.98 | 55.55 | 41.13 |
| Max | 149.47 | 139.94 | 134.2 |
| Split Frame Rendering | |||
| Resolution | 1024 x 768 | 1280 x 1024 | 1600 x 1200 |
| Min | 27.57 | 21.99 | 17.51 |
| Avg | 73.98 | 55.56 | 41.13 |
| Max | 149.85 | 139.77 | 134.13 |
| Alternate Frame Rendering 2 | |||
| Resolution | 1024 x 768 | 1280 x 1024 | 1600 x 1200 |
| Min | 27.53 | 22 | 17.51 |
| Avg | 73.97 | 55.55 | 41.13 |
| Max | 150.7 | 139.88 | 133.85 |
all numbers are in frame rates per second except for 1024 x 768, 1280 x 1024, and 1600 x 1200
Ouch. Forcing different SLI multi rendering doesn't seem to work, despite the fact that the application is SLI friendly. Obviously, to get SLI working properly with this game, NVIDIA had to implement some specific application tweaks This is not good. This means there's no point in giving users control over what SLI multi rendering scheme to use. Apparently, we still have to wait for a beta driver (or a new reference driver) that has a working SLI profile for new games. In other words, the ability of forcing SLI multi rendering is useless (at least in this example).
[Previous Page]
[Go to top]
[Next Page]