Radeon HD 4870 CrossFire: Driver Update From 8.6 to 8.8
If you believe the conjecture, the Catalyst 8.8 driver should introduce a considerable performance improvement in CrossFire mode. Particularly when it comes to communication between two graphics cards, ATI is able to optimize in many games.
| Assassins Creed in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
|---|---|---|---|---|---|---|---|
| HD 4870 CF (512 MB) 8.6 | 56.0 | 56.0 | 56.8 | 42.3 | 43.4 | 43.7 | 298.2 |
| HD 4870 CF (512 MB) 8.8 | 55.4 | 55.9 | 56.7 | 42.6 | 42.1 | 42.9 | 295.6 |
| Difference in performance in percentage | -1.1 | -0.2 | -0.2 | 0.7 | -3.0 | -1.8 | -0.9 |
The new graphics driver doesn’t appear to increase the frame rates.
| Call of Duty 4 in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
|---|---|---|---|---|---|---|---|
| HD 4870 CF (512 MB) 8.6 | 146.9 | 135.4 | 125.7 | 129.6 | 118.3 | 111.7 | 767.6 |
| HD 4870 CF (512 MB) 8.8 | 147.5 | 140.0 | 129.6 | 135.4 | 116.8 | 107.3 | 776.6 |
| Difference in performance in percentage | 0.4 | 3.4 | 3.1 | 4.5 | -1.3 | -3.9 | 1.2 |
The fluctuations are a little more extreme, but they’re in the tolerance region for driver optimizations.
| Crysis High in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
|---|---|---|---|---|---|---|---|
| HD 4870 CF (512 MB) 8.6 | 29.6 | 27.4 | 27.0 | 24.4 | 21.8 | 20.7 | 150.9 |
| HD 4870 CF (512 MB) 8.8 | 29.3 | 25.8 | 26.7 | 24.2 | 21.3 | 20.4 | 147.7 |
| Difference in performance in percentage | -1.0 | -5.8 | -1.1 | -0.8 | -2.3 | -1.4 | -2.1 |
Again, we aren’t seeing any appreciable gain after updating to the Catalyst 8.8 driver set. In fact, these losses are generally with a reasonable margin of error.
| Crysis Very High in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
|---|---|---|---|---|---|---|---|
| HD 4870 CF (512 MB) 8.6 | 27.6 | 25.0 | 22.4 | 22.1 | 19.1 | 14.5 | 130.7 |
| HD 4870 CF (512 MB) 8.8 | 26.7 | 24.8 | 20.9 | 22.0 | 18.0 | 12.9 | 125.3 |
| Difference in performance in percentage | -3.3 | -0.8 | -6.7 | -0.5 | -5.8 | -11.0 | -4.1 |
The new graphics driver sinks the frame rates in Very High setting and in some of the higher resolutions by six percent to 11 percent.
| ET: Quake Wars in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
|---|---|---|---|---|---|---|---|
| HD 4870 CF (512 MB) 8.6 | 59.9 | 57.8 | 54.2 | 58.1 | 55.1 | 52.9 | 338.0 |
| HD 4870 CF (512 MB) 8.7 | 65.0 | 65.6 | 65.2 | 62.6 | 62.7 | 62.5 | 383.6 |
| Difference in performance in percentage | 8.5 | 13.5 | 20.3 | 7.7 | 13.8 | 18.1 | 13.5 |
The use of the Catalyst 8.7 instead of the Catalyst 8.8 unfortunately meant using different graphics drivers in this test versus the others, which you wouldn’t typically do in the real world. As a result, the considerable performance gain of up to 20 percent is interesting to observe, though it’s still a shame we weren’t able to provide comparative numbers for 8.8.
| Half-Life 2 Episode 2 in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
|---|---|---|---|---|---|---|---|
| HD 4870 CF (512 MB) 8.6 | 103.2 | 97.7 | 98.0 | 102.6 | 97.5 | 96.2 | 595.2 |
| HD 4870 CF (512 MB) 8.8 | 104.7 | 99.5 | 98.3 | 102.5 | 99.0 | 97.1 | 601.1 |
| Difference in performance in percentage | 1.5 | 1.8 | 0.3 | -0.1 | 1.5 | 0.9 | 1.0 |
Half-Life 2 Episode 2 displays no performance increase due to the driver update.
| Mass Effect in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | 12808xAA | 16808xAA | 19208xAA | Overall |
|---|---|---|---|---|---|---|---|---|---|---|
| HD 4870 CF (512 MB) 8.6 | 93.2 | 91.8 | 90.9 | 89.6 | 86.8 | 74.0 | 88.9 | 86.0 | 72.2 | 773.4 |
| HD 4870 CF (512 MB) 8.8 | 92.6 | 91.7 | 89.3 | 87.3 | 86.2 | 74.7 | 88.0 | 84.0 | 72.7 | 766.5 |
| Difference in performance in percentage | -0.6 | -0.1 | -1.8 | -2.6 | -0.7 | 0.9 | -1.0 | -2.3 | 0.7 | -0.9 |
The results lie within the measurement tolerance; no improvement can be seen.
| MS Flight Simulator X SP2 in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
|---|---|---|---|---|---|---|---|
| HD 4870 CF (512 MB) 8.6 | 24.3 | 24.5 | 24.9 | 25.3 | 25.2 | 25.3 | 149.5 |
| HD 4870 CF (512 MB) 8.8 | 24.5 | 24.9 | 24.4 | 24.6 | 25.0 | 24.7 | 148.1 |
| Difference in performance in percentage | 0.8 | 1.6 | -2.0 | -2.8 | -0.8 | -2.4 | -0.9 |
Here, again, there are no improvement or losses worth mentioning; the values lie within the margin for error.
| World in Conflict in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
|---|---|---|---|---|---|---|---|
| HD 4870 CF (512 MB) 8.6 | 46.9 | 48.1 | 47.0 | 46.4 | 46.4 | 44.6 | 279.4 |
| HD 4870 CF (512 MB) 8.8 | 46.0 | 46.9 | 45.6 | 43.4 | 46.0 | 43.2 | 271.1 |
| Difference in performance in percentage | -1.9 | -2.5 | -3.0 | -6.5 | -0.9 | -3.1 | -3.0 |
The driver update generates a minor loss in performance, apart from the -6.5 percent in the 1280 resolution (4xAA); other fluctuations can be considered measuring tolerances.
This is observed in the piece and should not affect performance.
Can you please translate this into english?
I have no idea what you mean to say with your comment, other than the obvious (that you disagree on some level with the article)
Proof of my claim in case someone unexpectedly doubts it : http://www.opel.cc/xfx8800gtx/xfx.htm
That was the first failing card. The second I haven't even bothered sending in, and just bought an 4870 instead of waiting months.
That's a good,true and useful conclusion, but it's not a general one, only specific to this particular set of drivers.