Do New Drivers Really Boost Performance?
GeForce GTX 280: The Benefit Of CPU Overclocking
There are two tables for each game and each of the tested graphics card combinations. The second shows the performance differences between the normal CPU clock rate (2.93 GHz) and the overclocked CPU (3.47 GHz). The aim of overclocking is to find out what sort of benefit there is to faster processors in a day when Nvidia readily claims you don’t need a high-end processor to realize great gaming performance.
The driver versions and overclocked CPUs are listed directly after the chip name—the resolutions go from 1280x1024 to 1680x1050 to 1920x1200 pixels, first without anti-aliasing (0xAA) and then in the same order with anti-aliasing (4xAA, and in Mass Effect 8xAA as well). The two upper lines of figures represent the frame rates; the lower line is the difference in performance in percentage.
Assassins Creed in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
---|---|---|---|---|---|---|---|
GTX 280 (1024 MB) 177.92 | 63.3 | 63.3 | 63.3 | 50.2 | 49.8 | 48.2 | 338.1 |
GTX 280 (1024 MB) CPU@3.47 | 74.1 | 71.0 | 65.8 | 56.2 | 53.7 | 48.2 | 369.0 |
Difference in performance in percentage | 17.1 | 12.2 | 3.9 | 12.0 | 7.8 | 0.0 | 9.1 |
As can be expected, more CPU performance can be seen in the lower resolutions without anti-aliasing. The 18.2 percent overclocking makes itself felt at 1280x1024 pixels with a very good result of 17.1 percent. With 1920x1200 pixels and anti-aliasing, even the GTX 280 has reached its maximum capabilities; 2560x1600 pixels or a faster CPU are hardly necessary—the limit has been reached and can be measured.
Call of Duty 4 in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
---|---|---|---|---|---|---|---|
GTX 280 (1024 MB) 177.92 | 169.9 | 150.2 | 132.3 | 149.9 | 130.3 | 112.7 | 845.3 |
GTX 280 (1024 MB) CPU@3.47 | 169.1 | 147.5 | 131.4 | 148.2 | 130.1 | 112.7 | 839.0 |
Difference in performance in percentage | -0.5 | -1.8 | -0.7 | -1.1 | -0.2 | 0.0 | -0.7 |
This result is surprising. In the first resolution you assume a measuring error; the minor performance loss continues constantly throughout and also affects the HD 4870. The game is overloaded with the current graphics cards; more CPU power to the test system has a negative effect.
Crysis High in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
---|---|---|---|---|---|---|---|
GTX 280 (1024 MB) 177.92 | 37.3 | 33.1 | 31.5 | 29.4 | 26.2 | 23.3 | 180.8 |
GTX 280 (1024 MB) CPU@3.47 | 40.6 | 36.5 | 34.7 | 32.7 | 28.0 | 23.3 | 195.8 |
Difference in performance in percentage | 8.8 | 10.3 | 10.2 | 11.2 | 6.9 | 0.0 | 8.3 |
With High quality, a faster CPU can get a few percentage points more out of Crysis and the GTX 280, and the 18.2 percent overclocking yields an extra 10 to 11 percent higher frame rates. As we saw with Assassins Creed, the GTX 280 reaches its limits with Crysis at 1920x1200 pixels with anti-aliasing (4xAA).
Crysis Very High in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
---|---|---|---|---|---|---|---|
GTX 280 (1024 MB) 177.92 | 32.4 | 27.6 | 23.2 | 25.2 | 19.1 | 15.2 | 142.7 |
GTX 280 (1024 MB) CPU@3.47 | 35.0 | 27.8 | 23.2 | 25.4 | 19.0 | 15.2 | 145.6 |
Difference in performance in percentage | 8.0 | 0.7 | 0.0 | 0.8 | -0.5 | 0.0 | 2.0 |
With Very High image quality, the CPU is limited only in 1280 resolution. In all the higher modes or with anti-aliasing activated, the graphics card lacks power and the overclocked CPU offers no additional benefits. This means Very High mode in Crysis will continue to be a good method of testing graphics cards, as it takes even a GTX 280 to the limit.
ET: Quake Wars in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
---|---|---|---|---|---|---|---|
GTX 280 (1024 MB) 177.92 | 85.5 | 82.1 | 82.5 | 84.4 | 82.8 | 81.9 | 499.2 |
GTX 280 (1024 MB) CPU@3.47 | 93.0 | 93.1 | 92.4 | 92.3 | 91.1 | 90.2 | 552.1 |
Difference in performance in percentage | 8.8 | 13.4 | 12.0 | 9.4 | 10.0 | 10.1 | 10.6 |
Quake Wars reacts to the CPU power as could be expected, given that the frame rates are oriented toward the main processor. If you are able to increase the CPU clock rate further, the graphics card will allow itself to be coaxed into providing additional frames. The 18.2 percent CPU overclocking is converted into up to 13.4 percent more performance.
Stay On the Cutting Edge: Get the Tom's Hardware Newsletter
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
Half-Life 2 Episode 2 in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
---|---|---|---|---|---|---|---|
GTX 280 (1024 MB) 177.92 | 103.8 | 98.6 | 98.3 | 102.9 | 98.6 | 98.4 | 600.6 |
GTX 280 (1024 MB) CPU@3.47 | 115.8 | 110.1 | 110.0 | 114.9 | 110.2 | 109.8 | 670.8 |
Difference in performance in percentage | 11.6 | 11.7 | 11.9 | 11.7 | 11.8 | 11.6 | 11.7 |
Half-Life 2 Episode 2 is also considerably limited by the CPU. This can be seen by the constant frame rates and the even increases in all the resolutions. If you can introduce more CPU power to the system, you will get further frame gains. The 18.2 percent additional CPU clocking results in an overall performance increase of almost 12 percent for the game.
Mass Effect in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | 12808xAA | 16808xAA | 19208xAA | Overall |
---|---|---|---|---|---|---|---|---|---|---|
GTX 280 (1024 MB) 177.92 | 96.7 | 94.8 | 92.0 | 96.7 | 84.8 | 62.0 | 93.7 | 82.5 | 60.1 | 763.3 |
GTX 280 (1024 MB) CPU@3.47 | 107.6 | 105.1 | 96.6 | 105.0 | 86.0 | 62.2 | 101.2 | 82.5 | 60.1 | 806.3 |
Difference in performance in percentage | 11.3 | 10.9 | 5.0 | 8.6 | 1.4 | 0.3 | 8.0 | 0.0 | 0.0 | 5.6 |
Experience shows that Mass Effect ( based on the UT3 Engine) runs using 100 percent of the CPU load, which means every little bit extra should be visible. In the lower resolutions, the performance increase is good, but with anti-aliasing the GeForce GTX 280 reaches its limit as of 1680x1050 pixels.
MS Flight Simulator X SP2 in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
---|---|---|---|---|---|---|---|
GTX 280 (1024 MB) 177.92 | 32.7 | 32.7 | 32.2 | 34.6 | 34.4 | 33.7 | 200.3 |
GTX 280 (1024 MB) CPU@3.47 | 37.8 | 37.7 | 37.4 | 39.2 | 38.9 | 38.4 | 229.4 |
Difference in performance in percentage | 15.6 | 15.3 | 16.1 | 13.3 | 13.1 | 13.9 | 14.5 |
Flight Simulator X is hungry for CPU performance and the graphics card can surely provide more speed when it has the help of additional CPU power. The 18.2 percent CPU overclocking translates to as much as 16 percent higher frame rates. The general assumption is that Flight Simulator X is dependent solely on the CPU. But this is not true. If you take the test values of the GeForce 9600 GT for comparison, you can see that at 1920x1200 pixels (4xAA) without a powerful graphics card, the higher CPU performance is no longer realized. For the frame rates to increase, both a powerful GTX 280 and an even more powerful CPU are necessary.
World in Conflict in fps | 12800xAA | 16800xAA | 19200xAA | 12804xAA | 16804xAA | 19204xAA | Overall |
---|---|---|---|---|---|---|---|
GTX 280 (1024 MB) 177.92 | 51.3 | 49.6 | 46.6 | 44.2 | 37.3 | 33.2 | 262.2 |
GTX 280 (1024 MB) CPU@3.47 | 56.7 | 56.5 | 52.8 | 45.5 | 38.4 | 32.7 | 282.6 |
Difference in performance in percentage | 10.5 | 13.9 | 13.3 | 2.9 | 2.9 | -1.5 | 7.8 |
Without anti-aliasing, the increase in frame rates is very good, but both the game and the graphics card could handle more CPU power, which is still limited by the main processor. With anti-aliasing it’s not as good—even the GTX 280 starts to reach its limit. With 1920x1200 pixels and 4xAA it is visibly at the limit; increased CPU performance or test resolutions using 2560x1600 pixels will not make any difference.
Current page: GeForce GTX 280: The Benefit Of CPU Overclocking
Prev Page GeForce GTX 280: Driver Update from 177.39 to 177.92 Next Page GeForce 9600 GT: Driver Update From 175.16 to 177.92-
cangelini LedrosTo bad you used 8.8 drivers instead of the 8.9 ones. Real smart people.Reply
This is observed in the piece and should not affect performance. -
neiroatopelcc SoulLance"Unfortunately for AMD and its customers, the PowerPlay optimization for energy-efficient 2D operation is not included in the Catalyst 8.8 driver for the Radeon HD 4870. Quake Wars refuses to start using the Catalyst 8.8, but it works fine with the Catalyst 8.6 and 8.7."But neither should comments like this appear when everyone on the planet reads a labels that clearly say "Use latest drivers from manufacturer".Reply
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) -
randomizer He's saying you shouldn't criticise newer drivers because most hardware and software packaging has a label, or a section in the manual, telling you to use the latest drivers. So latest = best as far as he is concerned.Reply -
In reference to overclocking graphics cards, I'm pretty sure that XFX card warranties cover overclocking to some degree, though you'd have to double check their warranty fine print.Reply
-
neiroatopelcc SneppyIn reference to overclocking graphics cards, I'm pretty sure that XFX card warranties cover overclocking to some degree, though you'd have to double check their warranty fine print.Having had experience with two failing xfx cards (both 8800gtx's actually), I've got to warn you though, that their warranty is a bit fictional. While they may actually cover broken products, you can easily risk waiting months for a replacement.Reply
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. -
cryogenic We all know there were cases in the past when "Newer drivers provided substantial performance improvements" in various applications or games. Analyzing performance with a few select driver sets can can't lead to the conclusion that drivers are irrelevant, or that from a "professional" stand point the latest drivers shouldn't be chosen.Reply
-
randomizer CryogenicAnalyzing performance with a few select driver sets can can't lead to the conclusion that drivers are irrelevantBut it can lead to the conclusion that current drivers are irrelevant, or rather less relevant.Reply -
cryogenic randomizerReply
That's a good,true and useful conclusion, but it's not a general one, only specific to this particular set of drivers.