Looking Past the
Present
Undoubtedly, gamers are the primary market for 3D
graphics consumer cards. As with any other product, the
faster the performance, the higher the price you must pay.
Look at high end graphics cards such as ATI's Radeon
X1900XTX and NVIDIA's GeForce 7900GTX - these cards are
selling at about USD 450 to 600. If you're not satisfied
with the performance of these cards, the very fortunate of
you can use a pair of them for either a Crossfire or SLI
setup. If so, be prepared to spend around USD 1.000 to
1450.With that kind of serious money, gamers want the best. But that's true for any users, even those buying a motherboard with an integrated graphics card want to get their money's worth. That's why there are so many media (ourselves included) offering product reviews - we want to know if these products are worth the money we're paying for. Although performance and quality is king, there's usually another factor that most reviews can't capture - longevity. After all, if you just recently spent up to USD 1000 on new hardware, you don't want your new rig running to be too slow 3 / 6 months down the road. Now, some of you might be lucky enough to upgrade your graphics cards, processor and memory every 6 months, but not everyone is that lucky. Even with a buyback / trade in program, you typically have to fork the difference in price between your old card and the new one.
From experience, mainstream cards are typically good / fast enough for one year of games. High end cards usually fare better, offering some additional headroom so you can use it for about a year or at most a year and a half. In the graphics card business, that's the time it takes for the next generation to arrive (usually). In this case, choosing the right graphics card is even more important. However, the past six months is somewhat an anomaly in this regard. With ATI, the delay with R520 (Radeon X1800) have pushed the card's time to market very near its successor, the R580 (Radeon X1900). In fact, the Radeon X1800 may well be the only high end card to have a 3 months old life span. With NVIDIA, we basically saw small variations of the same design over the past year: the GeForce 7800GTX, 7800GTX 512 MB and 7900GTX. The most glaring differences between them are clock speeds and amount of memory.
We recently finished our review for both ATI's Radeon X1900XTX and NVIDIA's GeForce 7900GTX (which you can read here and here). Both cards are very fast, offering high enough frame rates for games available today. They're perform equally as well, with the Radeon X1900XTX having an edge for better image quality with anti aliasing and HDR plus AA support. On the other hand, the GeForce 7900GTX is faster when running without AA and AF and provide a much more effective transparency antialiasing. Although it's available with the GeForce 7800GTX, transparency antialiasing is only usable with the GeForce 7900GTX (at 4x). However, as we noted above, there's no way of knowing how these cards handle the new batch of games coming this year (or the year after). Other review websites and media usually rely on synthetic benchmarks for this purpose, however we don't really value the results from such benchmarks. For one, they are, by nature, synthetic benchmarks and not games - they 'simulate' what we may be seeing in games so they should never be used as the only metric in reviewing performance. Two, time and time again, vendors usually optimize their hardware and software to run benchmark well but these optimizations don't necessarily translate to real world performance.
Having finished the reviews of both cards, we find ourselves asking the question: how will these cards cope with newer games. NVIDIA seems more than happy to 'fix' everything with a driver update (Battlefield 2, Oblivion, Tomb Raider: Legend), much to the anguish of users. Users with ATI cards is somewhat better off - new games usually work right of the bat without any major (graphical) issues. New drivers may offer performance increases, but we believe there are no dubious optimizations at work here - ATI's developer team is still exploring the new architecture. So, we're more concerned with the GeForce 7900GTX than the Radeon X1900XTX in this regard.
The Problem
Let's make one thing absolutely clear - there's no way we can be 100% sure how the GeForce 7900GTX will behave in future games. What we can do here is make a calculated guess. Now let's look at the facts:- Both developers and hardware vendors seem to agree that the next generation of games and graphics engines will likely be more shader than texture dependent.
- Higher resolution will also be likely be the trend, both in gaming resolution and art (textures).
- We will also probably see more games making use effects such normal / parallax mapping as well as dynamic lighting and HDR rendering.
More shader use means the graphics card must be fast enough at processing shaders. There are two approach this problem - run the card (core) faster or process more shaders per cycle. NVIDIA chose to use a higher core clock for the GeForce 7900, keeping pretty much the same shader pipeline found in the GeForce 7800 series. Benchmarking shaders (in isolation, without the influence of bandwidth or even depth buffer) is very hard to do - you have to rely on a purely synthetic test since shaders used in current games are still very short and not very complex. So, while we will briefly touch this issue, we won't focus on future shaders performance on the GeForce 7900.
Higher resolution is easy enough - most reviews provide benchmarks from 1024 x 768 and up to 1600 x 1200 (or higher). When using a higher resolution frame buffer, the card must draw more pixels and process depth information more efficiently (in the depth buffer). Higher resolution textures will not only take more space (even when compressed) but also take more bandwidth (or latency due to decompression before processing). This is one area is definitely quite easy to test. The card must have a high fillrate, more efficient depth buffer management and memory bandwidth. There is also antialiasing and anisotropic filtering to consider. Antialiasing is usually related to fillrate, while anisotropic filtering is pretty much related to memory bandwidth.
While effects are related to shaders, we will focus more on their bandwidth related aspects. Normal / parallax mapping puts more strain on bandwidth because these effects make use of normal maps (and an additional height map for parallax mapping). Dynamic lighting can be done in a variety of ways, but usually involve rendering the scene from the light's perspective and translating the information to determine which part is lit or inside the shadows. Naturally, this falls under fillrate, just like HDR.
Now, if we were using synthetic benchmarks, all we have to is run one with normal / parallax mapping, dynamic lighting done through shaders at various resolutions ranging from 1024 x 768 to 1600 x 1200. Then we examine the fillrate and bandwidth test to see whether the graphics card have enough of both. But of course, we won't do that. This time, we will examine efficiency, not just pure scores. By looking at efficiency, we can see just how the GeForce 7900GTX fare under certain conditions like high fillrate and bandwidth demand. Why? Simple really, the GeForce 7900 (in fact, the entire GeForce 6 and 7 series) seems to be very sensitive to memory bandwidth.
Performance
The benchmark results in this article are taken from our Radeon X1900XTX and GeForce 7900GTX reviews. Naturally, the test setup is the same on both articles, the only difference being the drivers used. An important note: we're using Catalyst 6.2 drivers with the Radeon X1900XTX since those were the drivers that were used in that review. FYI, the latest drivers are 6.4 and performance with these drivers are slightly higher. So keep that in mind when you're looking at the numbers. NVIDIA's Forceware 84.21 was used with the GeForce 7900GTX (and 7800GTX).We'd like to thank both Tagan and Kingston for supplying with the additional power supply and 1 GB memory modules for this article.
Our test setup
AMD Athlon 64 3500+ socket 939
2 x 1024 MB Kingston KVR 3-3-3 PC3200 DDR-SDRAM
MSI K8N NForce 4 SLI motherboard
ASUS EN7800GTX TOP GeForce 7800GTX DDR3 256 MB graphics card
(running at standard and reference clocks - core 486 MHz / memory 675 MHz (1350 MHz effective) and core 430 MHz / memory 600 MHz (1200 MHz effective))
Gigabyte Radeon X1900XTX DDR3 512 MB graphics card
Leadtek WinFast PX7900GTX TDH 512 MB graphics card
Maxtor DiamondMaxPlus9 80 GB Serial ATA 8 MB buffer
ASUS E-616 DVD-ROM
Tagan TG530-U15 530 watts ATX/BTX power supply
Windows XP Professional with Service Pack 2 installed
ATI Catalyst 6.2 reference driver
NVIDIA Forceware 84.21 reference driver
NVIDIA NForce 4 6.70 reference driver
Creative SoundBlaster Live! 24 bit 5.12.1.512 driver.
DirectX 9.0c
Performance
Compared to the GeForce 7800GTX
Our first three benchmarks are quite old, but they serve their purpose well. Devoid of SM 2.0 and 3.0 shaders, these games still rely mostly on textures, so they're perfect for testing anti aliasing and anisotropic filtering performance. The table below is pretty self explanatory, but in case you're wondering, "Performance Lost" (in percent) here means the degree of performance lost (compared to default settings) when we enable both AA and AF. Keep in mind that there are fluctuations in the results, although it's very small (typically around 1-2 percent). For each resolution, we calculate an average performance lost for the minimum, average and maximum fps. If you just want a summary of these numbers, just look at the bottom of the table - you can see the weighted average performance lost from 1024 x 768 to 1600 x 1200.Call of Duty
|
|
7800GTX |
|
|
7800GTX TOP |
|
|
7900GTX |
|
|
|
|
Default | AA AF | Performance Lost | Default | AA AF | Performance Lost | Default | AA AF | Performance Lost |
| 1024 |
|
|
|
|
|
|
|
|
|
| Min | 82 | 81 | 1.22% | 81 | 79 | 2.47% | 79 | 78 | 1.27% |
| Avg | 224.65 | 178.91 | 20.36% | 225.92 | 189.56 | 16.09% | 201.11 | 190.75 | 5.15% |
| Max | 503 | 312 | 37.97% | 538 | 337 | 37.36% | 404 | 339 | 16.09% |
| Average |
|
|
19.85% |
|
|
18.64% |
|
|
7.50% |
| 1280 |
|
|
|
|
|
|
|
|
|
| Min | 83 | 84 | -1.20% | 83 | 83 | 0.00% | 80 | 79 | 1.25% |
| Avg | 213.85 | 132.92 | 37.85% | 217.51 | 146.22 | 32.77% | 197.42 | 158.77 | 19.58% |
| Max | 427 | 206 | 51.76% | 459 | 231 | 49.67% | 413 | 271 | 34.38% |
| Average |
|
|
29.47% |
|
|
27.48% |
|
|
18.40% |
| 1600 |
|
|
|
|
|
|
|
|
|
| Min | 82 | 68 | 17.07% | 81 | 77 | 4.94% | 80 | 79 | 1.25% |
| Avg | 193.49 | 96.27 | 50.25% | 201.95 | 107.77 | 46.64% | 192.84 | 123.74 | 35.83% |
| Max | 342 | 138 | 59.65% | 358 | 156 | 56.42% | 383 | 182 | 52.48% |
| Average |
|
|
42.32% |
|
|
36.00% |
|
|
29.85% |
|
|
|
|
|
|
|
|
|
|
|
| Weighted Average |
|
|
30.55% |
|
|
27.37% |
|
|
18.59% |
That's just incredible - look at how much frame rates we're losing with the GeForce 7800GTX - 20 to 40 percent from 1024 x 768 to 1600 x 1200. The GeForce 7900GTX with its higher clocked core allowed us to reach frame rates much higher at higher resolutions. No doubt, you'll notice that the penalty is similar between the GeForce 7800GTX at 1024 x 768 and GeForce 7900GTX at 1280 x1024. So, we'll likely to see the same trend, it's on different resolutions. (1280 x 1024 as opposed to 1024 x 768). Overall, we can see the GeForce 7900GTX has pretty much half the penalty of the GeForce 7800GTX with AA and AF.
Homeworld 2
|
|
7800GTX |
|
|
7800GTX TOP |
|
|
7900GTX |
|
|
|
|
Default | AA AF | Performance Lost | Default | AA AF | Performance Lost | Default | AA AF | Performance Lost |
| 1024 |
|
|
|
|
|
|
|
|
|
| Min | 59 | 62 | -5.08% | 60 | 65 | -8.33% | 60 | 63 | -5.00% |
| Avg | 205.75 | 199.67 | 2.96% | 206.41 | 205.47 | 0.45% | 200.5 | 201.32 | -0.41% |
| Max | 407 | 412 | -1.23% | 409 | 414 | -1.22% | 389 | 394 | -1.29% |
| Average |
|
|
-1.12% |
|
|
-3.03% |
|
|
-2.23% |
| 1280 |
|
|
|
|
|
|
|
|
|
| Min | 64 | 51 | 20.31% | 64 | 58 | 9.38% | 58 | 61 | -5.17% |
| Avg | 203.89 | 153.26 | 24.83% | 204.84 | 171.07 | 16.48% | 198.81 | 189.87 | 4.50% |
| Max | 406 | 268 | 33.9% | 409 | 302 | 26.16% | 389 | 359 | 7.71% |
| Average |
|
|
26.38% |
|
|
17.34% |
|
|
2.35% |
| 1600 |
|
|
|
|
|
|
|
|
|
| Min | 63 | 35 | 44.44% | 62 | 40 | 35.48% | 61 | 47 | 22.95% |
| Avg | 197.63 | 106.63 | 46.05% | 200.29 | 121.35 | 39.41% | 194.58 | 142.37 | 26.83% |
| Max | 404 | 174 | 56.93% | 404 | 199 | 50.74% | 384 | 230 | 40.10% |
| Average |
|
|
49.14% |
|
|
41.88% |
|
|
29.96% |
|
|
|
|
|
|
|
|
|
|
|
| Weighted Average |
|
|
23.09% |
|
|
18.73% |
|
|
10.03% |
Here we can see pretty much the same thing, only this time the GeForce 7900GTX has less than half the penalty of a GeForce 7800GTX. If you look at the details, you can see that there's virtually no penalty whatsoever on the GeForce 7900GTX except at 1600 x 1200. That's actually not surprising since this is an RTS game. Compared to the GeForce 7800GTX, the GeForce 7900GTX retain 20 percent more of its frame rate under default settings. We wonder though, if we were able to reach even higher resolutions such as 2048 x 1536, will we see the same trend as Call of Duty.
Richard Burns Rally
|
|
7800GTX |
|
|
7800GTX TOP |
|
|
7900GTX |
|
|
|
|
Default | AA AF | Performance Lost | Default | AA AF | Performance Lost | Default | AA AF | Performance Lost |
| 1024 |
|
|
|
|
|
|
|
|
|
| Min | 108 | 105 | 2.78% | 107 | 102 | 4.67% | 106 | 103 | 2.83% |
| Avg | 153.5 | 149.69 | 2.48% | 157.33 | 154.69 | 1.68% | 152.13 | 150.9 | 0.81% |
| Max | 229 | 221 | 3.49% | 229 | 226 | 1.31% | 228 | 219 | 3.95% |
| Average |
|
|
2.92% |
|
|
2.55% |
|
|
2.53% |
| 1280 |
|
|
|
|
|
|
|
|
|
| Min | 106 | 84 | 20.75% | 108 | 95 | 12.04% | 105 | 102 | 2.86% |
| Avg | 153.81 | 124.29 | 19.19% | 156.64 | 137.18 | 12.42% | 152.44 | 148.41 | 2.65% |
| Max | 223 | 218 | 2.24% | 234 | 212 | 9.40% | 223 | 216 | 3.14% |
| Average |
|
|
14.06% |
|
|
11.29% |
|
|
2.88% |
| 1600 |
|
|
|
|
|
|
|
|
|
| Min | 101 | 63 | 37.62% | 108 | 71 | 34.26% | 105 | 90 | 14.29% |
| Avg | 153.43 | 92.77 | 39.54% | 156.38 | 104.77 | 33.00% | 151.47 | 129.68 | 14.38% |
| Max | 219 | 155 | 29.22% | 233 | 178 | 23.61% | 219 | 207 | 5.48% |
| Average |
|
|
35.46% |
|
|
30.29% |
|
|
11.38% |
|
|
|
|
|
|
|
|
|
|
|
| Weighted Average |
|
|
17.48% |
|
|
14.71% |
|
|
5.60% |
The results we're seeing in this game is just way out there - the GeForce 7900GTX is more than three times more efficient than the GeForce 7800GTX. Most of the difference can only be seen at 1600 x 1200, where the GeForce 7900GTX can make optimal use of its higher clocked core and memory. If we were to 'superimpose' the trend we're seeing with Call of Duty (and Homeworld 2), the penalty is similar between a GeForce 7800GTX running at 1280 x 960 and a GeForce 7900GTX running at 1600 x 1200. The relatively small difference (around 3-5 percent) is likely normal variations between runs.
All right, we've seen that the GeForce 7900GTX is more efficient than the GeForce 7800GTX, at their own clocks. It's most likely that the higher clocked core and memory that's making the difference here. After all, a 220 / 200 MHz difference is nothing to scoffed about - that's actually close to half the core clock of the GeForce 7800GTX. You might be wondering why we're not seeing even lower penalty on the GeForce 7900GTX. The answer is in the memory clock - it's only about 1/3 faster (1200 : 1600 MHz). That's about 30 - 35 percent, which corresponds to the difference we're seeing. So, even though theoretically we should be able to get higher frame rates, there's not enough memory bandwidth to keep the core fed. This would explain the drastic change of penalty at certain resolutions.
F.E.A.R
|
|
7800GTX |
|
|
7800GTX TOP |
|
|
7900GTX |
|
|
|
|
Default | AA AF | Performance Lost | Default | AA AF | Performance Lost | Default | AA AF | Performance Lost |
| 1024 |
|
|
|
|
|
|
|
|
|
| Min | 47 | 35 | 25.53% | 46 | 40 | 13.04% | 42 | 41 | 2.38% |
| Avg | 103.94 | 67.22 | 35.32% | 111.48 | 73.74 | 33.85% | 113.01 | 92.34 | 18.29% |
| Max | 242 | 157 | 35.12% | 271 | 177 | 34.69% | 295 | 295 | 0.00% |
| Average |
|
|
31.99% |
|
|
27.19% |
|
|
6.89% |
| 1280 |
|
|
|
|
|
|
|
|
|
| Min | 44 | 27 | 38.64% | 43 | 29 | 32.56% | 40 | 36 | 10.00% |
| Avg | 82.92 | 47.97 | 42.15% | 89.33 | 53.19 | 40.45% | 99.22 | 65.11 | 34.37% |
| Max | 173 | 108 | 37.57% | 192 | 124 | 35.42% | 241 | 241 | 0.00% |
| Average |
|
|
39.45% |
|
|
36.14% |
|
|
14.79% |
| 1600 |
|
|
|
|
|
|
|
|
|
| Min | 34 | 19 | 44.12% | 38 | 21 | 44.74% | 41 | 26 | 36.59% |
| Avg | 58.95 | 33.36 | 43.40% | 65.78 | 37.39 | 43.16% | 80.2 | 45.9 | 42.78% |
| Max | 116 | 73 | 37.07% | 130 | 82 | 36.92% | 169 | 103 | 39.05% |
| Average |
|
|
41.53% |
|
|
41.61% |
|
|
39.47% |
|
|
|
|
|
|
|
|
|
|
|
| Weighted Average |
|
|
37.66% |
|
|
34.98% |
|
|
20.38% |
We mentioned in our GeForce 7900GTX review that this card is fast enough if you want to play F.E.A.R at 1280 x 960 with AA and AF. You can look at the benchmark numbers, but by using efficiency as a metric it's easier to see why. We're losing only about half the frame rates compared to the GeForce 7800GTX. However, at 1600 x 1200, the penalty is pretty much the same. so its very likely even the shader processing power from the higher clocked GeForce 7900GTX is still not enough. This is different to what we're seeing at 'traditional' texture dependent games - a more gradual (less drastic) decline going from 1024 x 768 to 1280 x 960 like we saw with Call of Duty.
Quake 4
|
|
7800GTX |
|
|
7800GTX TOP |
|
|
7900GTX |
|
|
|
|
Default | AA AF | Performance Lost | Default | AA AF | Performance Lost | Default | AA AF | Performance Lost |
| 1024 |
|
|
|
|
|
|
|
|
|
| Min | 69 | 57 | 17.39% | 70 | 64 | 8.57% | 69 | 64 | 7.25% |
| Avg | 124.51 | 87.5 | 29.73% | 125.23 | 96.39 | 23.03% | 124.51 | 108.01 | 13.26% |
| Max | 168 | 135 | 19.64% | 168 | 148 | 11.90% | 168 | 156 | 7.14% |
| Average |
|
|
22.25% |
|
|
14.50% |
|
|
9.21% |
| 1280 |
|
|
|
|
|
|
|
|
|
| Min | 68 | 37 | 45.59% | 69 | 42 | 39.13% | 61 | 48 | 21.31% |
| Avg | 116.96 | 59.25 | 49.34% | 119.44 | 66.83 | 44.05% | 117.22 | 79.55 | 32.13% |
| Max | 171 | 102 | 40.35% | 167 | 113 | 32.34% | 164 | 129 | 21.34% |
| Average |
|
|
45.09% |
|
|
38.51% |
|
|
24.93% |
| 1600 |
|
|
|
|
|
|
|
|
|
| Min | 66 | 27 | 59.09% | 68 | 30 | 55.88% | 60 | 38 | 36.67% |
| Avg | 97.05 | 43.5 | 55.18% | 104.9 | 49.07 | 53.23% | 112.97 | 59.29 | 47.52% |
| Max | 154 | 81 | 47.40% | 161 | 87 | 45.96% | 156 | 101 | 35.26% |
| Average |
|
|
53.89% |
|
|
51.69% |
|
|
39.81% |
|
|
|
|
|
|
|
|
|
|
|
| Weighted Average |
|
|
40.41% |
|
|
34.90% |
|
|
24.65% |
Although the overall penalty is still in the GeForce 7900GTX favor, we can all see it beginning to lose footing at 1280 x 1024. We do see that the trend is less drastic than with the GeForce 7800GTX. If we were to use an even higher resolution, such as 2048 x 1536, it's more likely the penalty will be close to 50 percent on the GeForce 7900GTX - just like what we're seeing on the GeForce 7800GTX at 1600 x 1200. Even with an SLI setup, we will likely not have any aster frame rates, so 1600 x 1200 is the end of the line for the GeForce 7900GTX.
Serious Sam II
|
|
7800GTX |
|
|
7800GTX TOP |
|
|
7900GTX |
|
|
|
|
Default | AA AF | Performance Lost | Default | AA AF | Performance Lost | Default | AA AF | Performance Lost |
| 1024 |
|
|
|
|
|
|
|
|
|
| Min | 45 | 45 | 0.00% | 46 | 45 | 2.17% | 42 | 43 | -2.38% |
| Avg | 59.55 | 54.63 | 8.27% | 59.63 | 57.21 | 4.05% | 56.76 | 56.07 | 1.22% |
| Max | 84 | 68 | 19.05% | 84 | 74 | 11.90% | 79 | 79 | 0.00% |
| Average |
|
|
9.11% |
|
|
6.04% |
|
|
-0.39% |
| 1280 |
|
|
|
|
|
|
|
|
|
| Min | 45 | 37 | 17.78% | 45 | 38 | 15.56% | 42 | 40 | 4.76% |
| Avg | 58.85 | 41.73 | 29.09% | 59.48 | 46.62 | 21.61% | 56.62 | 53.33 | 5.81% |
| Max | 82 | 51 | 37.80% | 84 | 57 | 32.14% | 79 | 71 | 10.13% |
| Average |
|
|
28.22% |
|
|
23.10% |
|
|
6.90% |
| 1600 |
|
|
|
|
|
|
|
|
|
| Min | 45 | 26 | 42.22% | 44 | 26 | 40.91% | 43 | 32 | 25.58% |
| Avg | 52.33 | 29.35 | 43.91% | 56.43 | 32.9 | 41.69% | 56.33 | 41.06 | 27.11% |
| Max | 67 | 36 | 46.27% | 74 | 41 | 44.59% | 79 | 52 | 34.18% |
| Average |
|
|
44.13% |
|
|
42.40% |
|
|
28.96% |
|
|
|
|
|
|
|
|
|
|
|
| Weighted Average |
|
|
27.16% |
|
|
23.85% |
|
|
11.82% |
The picture we're seeing here is pretty similar - you only bought one more resolution with the GeForce 7900GTX. But let's not detract from the fact that at 28.96 percent penalty hit is still very good and understandably so coming from a GeForce 7 architecture. What do we mean by this? It's very likely that NVIDIA made only little tweaks here and there on the GeForce 7900GTX to maximize performance and get so high a clock. In a word - the GeForce 7900GTX is pretty much a 'speed bump'. Like all speed bumps, since efficiency is pretty much the same (on the same clock), the GeForce 7900GTX will most likely behave pretty much like the GeForce 7800GTX. At some point with a game that's heavy enough (either in shaders or textures), you'll begin to see a drastic drop in performance.
Splinter Cell: Chaos Theory
|
|
7800GTX |
|
|
7800GTX TOP |
|
|
7900GTX |
|
|
|
|
Default | AA AF |
|
Default | AA AF |
|
Default | AA AF |
|
| 1024 |
|
|
|
|
|
|
|
|
|
| Min | 62.66 | 45.02 | 28.15% | 70.58 | 50.74 | 28.11% | 76.91 | 62.94 | 18.16% |
| Avg | 100.05 | 78.1 | 21.94% | 107.88 | 87.35 | 19.03% | 115.02 | 103.14 | 10.33% |
| Max | 178.93 | 155.45 | 13.12% | 198.55 | 166.23 | 16.28% | 198.8 | 192.49 | 3.17% |
| Average |
|
|
21.07% |
|
|
21.14% |
|
|
10.56% |
| 1280 |
|
|
|
|
|
|
|
|
|
| Min | 43.89 | 31.11 | 29.12% | 49.42 | 35.15 | 28.87% | 63.78 | 44.66 | 29.98% |
| Avg | 73.34 | 54.79 | 25.29% | 82.1 | 61.77 | 24.76% | 102.66 | 77.86 | 24.16% |
| Max | 146.1 | 136.75 | 6.40% | 160.93 | 139.08 | 13.58% | 186.06 | 153.15 | 17.69% |
| Average |
|
|
20.27% |
|
|
22.40% |
|
|
23.94% |
| 1600 |
|
|
|
|
|
|
|
|
|
| Min | 32.51 | 22.97 | 29.34% | 36.6 | 25.92 | 29.18% | 47.94 | 33.13 | 30.89% |
| Avg | 54.79 | 39.99 | 27.01% | 61.2 | 45.23 | 26.09% | 79.68 | 57.61 | 27.70% |
| Max | 136.75 | 105.53 | 22.83% | 144.22 | 119.57 | 17.09% | 150.75 | 135.42 | 10.17% |
| Average |
|
|
26.40% |
|
|
24.12% |
|
|
22.92% |
|
|
|
|
|
|
|
|
|
|
|
| Weighted Average |
|
|
22.58% |
|
|
22.56% |
|
|
19.14% |
Splinter Cell: Chaos Theory is quite an anomaly - we're seeing the penalty hit stays pretty much the same across the results from three graphics cards. There are several things at play here. For one, this game is not pixel fillrate limited, but shader limited. So, we're still within the limits of these cards bandwidth window, even at 1600 x 1200 with AA AF. Two, architecturally the two are the same and now we can see just how similar. Kudos to the developers at Ubisoft. They managed to pretty much 'maintain' the penalty rate even with two cards that have a 220 / 200 MHz difference in clocks.
HDR
Serious Sam II
|
|
7800GTX |
|
|
7800GTX TOP |
|
|
7900GTX |
|
|
|
|
Default | HDR |
|
Default | HDR |
|
Default | HDR |
|
| 1024 |
|
|
|
|
|
|
|
|
|
| Min | 45 | 42 | 6.67% | 46 | 44 | 4.35% | 42 | 42 | 0.00% |
| Avg | 59.55 | 50.39 | 15.39% | 59.63 | 54.5 | 8.60% | 56.76 | 57.53 | -1.36% |
| Max | 84 | 63 | 25.00% | 84 | 71 | 15.48% | 79 | 78 | 1.27% |
| Average |
|
|
15.68% |
|
|
9.48% |
|
|
-0.03% |
| 1280 |
|
|
|
|
|
|
|
|
|
| Min | 45 | 30 | 33.33% | 45 | 33 | 26.67% | 42 | 43 | -2.38% |
| Avg | 58.85 | 36.78 | 37.50% | 59.48 |