by Sergey Lepilov
12/10/2009 | 04:55 PM
An advanced graphics card or a couple of such cards in a CrossFireX or SLI configuration is at the heart of each gaming computer. However, when they upgrade or buy new hardware, users often wonder what CPU they might need to match such a graphics subsystem. It is when new top-end products from AMD or Nvidia come out that such questions arise in abundance and arguments at computer forums may grow really hot. But perhaps the issue is not really worth the argument?
I will try to clear things out today by benchmarking the performance of the fastest single-chip graphics card available today, the ATI Radeon HD 5870, and a CrossFireX configuration built out of two such cards. The task is not difficult but takes a lot of time. I needed to perform tests on three platforms at two display resolutions in three synthetic benchmarks and 15 games, focusing on the performance benefits brought about by overclocking the CPUs and platforms. Anyway, the three week’s work has produced really interesting results I’m ready to share them with you. First, I’ll tell you about the testing methods and hardware I used.
First let us say a few words about the hardware. Considering that the Radeon HD 5870 is a top-end product ($399), it is unlikely to be installed into a computer with a dual- or single-core processor. However, this is theoretically possible. So I took two quad-core processors from AMD and Intel and added a dual-core processor which produced valuable results.
Here are the hardware components that we used to build our test systems:
On each of the three platforms the graphics cards were benchmarked singly and in CrossFireX mode at three CPU frequencies, the other system frequencies being as similar as possible. The quad-core Intel Core i7 was tested at its default frequency of 2.67GHz as well as at increased frequencies of 3.4GHz and 4.1GHz (this is the maximum clock rate I could achieve with my hardware). Overclocking was done by changing the base frequency and the multiplier.
I enabled the Load Line Calibration feature in the BIOS of my ASUS P6T Deluxe mainboard and the core voltage changed from the default 1.2V to 1.3825V depending on the frequency. The system memory (DDR3) worked at a frequency of 1.6GHz in the first two cases and 1.56GHz in the third case, its voltage being set at 1.64V. Thus, the memory frequency remained almost the same all the time. Coupled with identical timings (7-7-7-14_1T), this helped maintain very similar speeds of reading from and writing to system memory as well as its access time. The other overclocking parameters were left at their defaults in the mainboard’s BIOS (set at Auto).
It was harder with the AMD Phenom II X4 965 processor because my mainboard could not increase the base frequency higher than 256MHz. Increasing the HTLink and North Bridge frequencies as high as possible without losing stability, I overclocked the CPU to 4.1GHz at 1.525V. Besides, I tested the Phenom II X4 at its default 3.4GHz (with a voltage of 1.4V) and at a frequency of 2.67GHz (which is the default frequency of the Intel Core i7 and produces something in between the Phenom II X4 910 and X4 920 models).
Despite my using fast DDR2 modules in the system, the integrated memory controller of the AMD processor imposed some limitations resulting in a performance reduction. I could not use CAS Latency 4, Command Rate 1T and a step-up memory divisor. Unfortunately, I did not have two good 2GB modules and it was not right to sacrifice 2GB of RAM (by removing two modules) to benchmark modern games in Windows 7. So, I tried to squeeze what I could out of my components. At a frequency of 908MHz I set the main and secondary timings to minimum stable values.
I want to note that I did not deliberately slow down the AMD platform. I only tried to overclock it as best as I could with the components at hand. Running a little ahead, I must confess that my tests did not reveal any slowing down.
And finally, the third platform is with an Intel Core 2 Duo E8400 whose default frequency is 3.0GHz and maximum frequency (with air cooling) is over 4.3GHz at 1.55V, but I tested it at the same frequencies as the two previous CPUs. This will show us the difference between CPUs of different architectures in games and benchmarks as well as the practical benefit of having more processor cores, the operating frequency being the same.
I used the same DDR2 memory modules on this platform as on the AMD platform, but the frequencies and timings were different due to the peculiarities of overclocking an Intel CPU. At a CPU frequency of 4.1GHz, the memory worked at 1092MHz. At CPU frequencies of 3.4GHz and 2.67GHz, the memory frequencies were 1023MHz and 1066MHz, respectively. The main timings were higher in every case than on the AMD platform: 5-5-5-12_2T. I did not change the additional timings.
The use of different system memory, DDR3 for the Intel Core i7 platform and DDR2 for the AMD Phenom II X4 platform, makes it impossible to compare these two platforms directly. And such a comparison was not in the scope of the test session. Besides, the difference in performance between DDR2-800 (5-5-5-15_2T) and DDR3-1600 (7-7-7-20_1T) on AMD platforms is very small. And in my system I had DDR2 running at 908MHz rather than at 800MHz as in the mentioned article while the main timings were set at 5-4-4-12, so the difference from DDR3-1600 (7-7-7-20_1T) was even smaller.
One more pitfall for the AMD platform in this test session was PCI Express. The Gigabyte GA-MA790GP-DS4H mainboard is based on the AMD 790GX chipset which supports two ATI Radeon cards in CrossFireX only as PCIe x16 + PCIe x8, as opposed to both platforms from Intel which support two PCIe x16. I can’t say definitely if this provokes a performance hit on the AMD platform because I could not carry out a comparative test on a mainboard with the AMD 790FX chipset (supports two PCIe x16). I’m going to keep this fact in mind when analyzing the results.
By the way, I had planned to test with a smaller CPU frequency increment, but this proved to be unnecessary as you’ll see in the Performance section.
I used two graphics cards, which we reviewed before: an AMD Radeon HD 5870 and a HIS Radeon HD 5870. One of them was tested as a single graphics card in the system and they both were also tested in CrossFireX mode:
Now let’s move on to software and benchmarking tools that we used. All tests were performed in Windows 7 Ultimate RTM x64 operating system with the following drivers:
The graphics cards were tested in two resolutions: 1280x1024 and1920x1200. Of course, we all understand that the first resolution has no practical value for powerful graphics cards like the ones tested as well as for CrossFireX configuration on these graphics cards. However, the difference in platforms performance and CPU dependence will be more obvious than in 1920x1200. Our monitor doesn’t support resolutions above 1920x1200, but it is a minor issue, because very few gamers use higher screen resolutions anyway and the tested graphics cards cannot provide sufficient performance to ensure comfortable gaming experience in 2560x1600 resolution.
The tests were performed in two image quality modes: “High Quality” (HQ) without any image quality enhancements and “HQ+ AF16x+AA4/8x” with enabled 16x anisotropic filtering and 4x full screen anti-aliasing (or 8x FSAA if the average framerate was high enough for comfortable gaming experience). We enabled anisotropic filtering and full-screen anti-aliasing from the game settings or configuration files. If the corresponding options were missing, we changed these settings in the Catalyst Control Panel. Vertical sync was always off in driver control panel.
The list of applications and benchmarks includes three popular synthetic benchmarking suites and 15 games of various genres. All games were updated with the latest patches available at the time of tests. Here is the complete list of tests used with the settings (all games listed in their release order):
Here I’d like to add that if the game allowed recording the minimal fps readings, they were also added to the charts, which is of great importance to CPU dependence analysis. Luckily 8 games out of 15 allow it. We ran each game test or benchmark twice and took the best result for the diagrams, but only if the difference between them didn’t exceed 1%. If it did exceed 1%, we ran the tests at least one more time to achieve repeatability of results.
I want to note one thing before proceeding to the tests. When benchmarking my Radeon HD 5870 or Radeon HD 5870 CrossFireX on the platforms with CPU frequencies of 2.67GHz, I saw in some games that the average and bottom frame rate was higher at 1920x1200 than at 1280x1024. This is not an error in the diagram, but a real fact. I can’t give you a thorough explanation but I am sure that when such a thing happens, it means that the graphics subsystem is too fast for the given CPU or platform at large.
In the diagrams below the results of the Intel Core 2 Duo platform are marked orange, those of the AMD Phenom II X4 platform are marked green, and those of the Intel Core i7 platform are marked blue. Semi-synthetic benchmarks go first.
The Radeon HD 5870 is clearly CPU-dependent, both singly and in CrossFireX mode, in the good old 3DMark 2006. The only exception is the high-quality test mode at 1920x1200 where the single card does not speed up as fast as in the other three cases.
Talking about the platforms and CPUs, the Intel Core i7 is expectedly in the lead. Clocked at 3.4GHz, it is successfully competing with the 4.1GHz Phenom II X4. Although the Core architecture and the Intel Core 2 Duo processor are over 3 years old already, the dual-core CPU (when properly overclocked) is quite competitive to newer and more expensive CPUs.
The newer 3DMark Vantage relies less heavily on the CPU, yet the Radeon HD 5870 and its CrossFireX configuration lack CPU resources in the Performance mode. This is clearer with the dual-card configuration, of course. As for the CPUs and platforms, the Intel Core i7 is in the lead again whereas the Intel Core 2 Duo is not as confident as in 3DMark 2006.
No comments about the CPU-related limitations. But I can note that the CrossFireX technology brings about the same performance gain on each of the three tested platforms.
In this game we can see the Radeon HD 5870 depend on the speed of the platform until the high-quality mode at 1920x1200. The CrossFireX configuration responds eagerly to the increased performance of the platform even at the FSAA + AF settings. The bottom speed grows up faster than the average frame rate.
The results on the AMD platform are surprising. The CrossFireX technology brings but a small performance gain, which is downright poor in comparison with what we get on the Intel platforms. Could it be the lack of eight PCIe lanes? If so, why didn’t we see this in the three previous tests where the CrossFireX mode boosted the speed to the same extent on all the platforms? I’ll give you my explanation later on. Right now, let’s move on to the next game.
Easy to see, the Intel Core i7 platform is the optimal choice for one or even two Radeon HD 5870 cards as it easily beats the other two platforms. The performance of the single Radeon HD 5870 does not grow up much when the CPU is overclocked whereas the CrossFireX subsystem speeds up in nearly every test mode and resolution. The AMD Phenom II X4 is disappointing again. When overclocked, it is no faster than the platform with the Intel Core 2 Duo. I must warn you, though, that the AMD Phenom II X4 is inexplicably poor in Crysis and World in Conflict. Let’s see what we have in the other games.
That’s a nice picture. The single Radeon HD 5870 only reacts to the increase of the platform speed at 1280x1024 at the low-quality settings, but the Radeon HD 5870 CrossFireX configuration delivers a linear performance growth until 1920x1200 at the high-quality settings. The quad-core CPUs show their worth at last. Even the weakest of them, the AMD Phenom II X4 at 2.67GHz, beats the Intel Core 2 Duo overclocked to 4.1GHz. The Intel Core i7 beats the AMD Phenom II X4 just as easily, though. It must be noted that CrossFireX technology works smoothly on the AMD platform in this game, delivering as high a performance gain in percent as on the two Intel platforms.
This game’s results are similar to those of Unreal Tournament 3: the single Radeon HD 5870 is quite satisfied with the junior quad-core CPU while the pair of such cards eagerly welcomes a powerful CPU with fast platform, boosting the average frame rate.
As you can see, you don’t have to worry about your CPU if you want to play S.T.A.L.K.E.R. on a single Radeon HD 5870 card. An Intel Core 2 Duo at 2.67GHz is more than enough for this game. It’s different with the CrossFireX configuration for which the overclocking of the Intel Core i7 is only unrewarding. The Core i7 platform is fast even at the default CPU frequency. As opposed to that, overclocking the Intel Core 2 Duo and AMD Phenom II X4 from 2.67 to 3.4GHz helps improve the average frame rate and get a considerable increase in the bottom speed. When the CPU frequency is increased from 3.4 to 4.1GHz, the speed grows up too, but not that much. This is only true for the resolution of 1280x1024 but there is little sense in playing S.T.A.L.K.E.R. at that resolution. When it comes to 1920x1200, overclocking is unrewarding.
Again, the single Radeon HD 5870 does not need an advanced processor. There is a performance growth at 1280x1024 with low-quality settings, but only on the Intel Core 2 Duo and at the first frequency of the AMD Phenom II X4 platform. The Intel Core i7 platform is always fast and the Radeon HD 5870 does not deliver better results in Far Cry 2 if you overclock the CPU.
The CrossFireX tests suggest that two Radeon HD 5870 cards can make a good use of a fast CPU. Save for the resolution of 1920x1200 at high-quality settings, the Intel Core i7 enjoys a hefty advantage over the two other CPUs. If you’ve got two Radeon HD 5870 and want to play Far Cry 2, you may want to consider upgrading your CPU.
Call of Duty 5: World at War responds to the increased speed of the platform, too. The efficiency of CrossFireX depends on the CPU as well as the platform at large. There is almost no performance growth on the weak Intel Core 2 Duo. The results are higher on the faster quad-core AMD Phenom II X4 but it is only with the Core i7 that the Radeon HD 5870 pair shows its best at 1920x1200 with high-quality settings. From a practical point of view, the speed of 80fps delivered by the single Radeon HD 5870 with Intel Core 2 Duo at 2.67GHz is already quite enough. Unfortunately, I could not measure the bottom frame rate, which might be the guiding factor for choosing the lowest allowable CPU for playing Call of Duty 5: World at War comfortably.
Left 4 Dead is not a resource-consuming game, either (and I’m planning to replace it with Left 4 Dead 2 in my future reviews), so it is hard to see that some CPU is insufficient for a Radeon HD 5870 or a CrossFireX tandem of two such cards. We can see that there is little difference between the CPUs with the single Radeon HD 5870 in the 8x MSAA + 16x AF mode at 1920x1200 whereas the CrossFireX tandem speeds up only on the Intel Core i7 platform. The lack of bottom speed data reduced the practical value of these numbers, though.
CrossFireX technology is not supported by Warhammer 40 000: Dawn of War II and the results get only worse when you turn it on. However, the results of the single graphics card are quite indicative of the performance of the platforms and the influence of the CPU because this game proves to be quite demanding in this respect. Overclocking each platform helps increase the average and bottom frame rates. Unfortunately, the quad-core AMD Phenom II X4 loses even to the Intel Core 2 Duo, let alone the Intel Core i7. The latter delivers an impressive increase in bottom speed, which is most necessary for comfortable play.
BattleForge is one more real-time strategy, but it is more indifferent to the speed of the platform than Warhammer 40 000: Dawn of War II. For the single Radeon HD 5870, any of the three tested CPUs suffices. For two graphics cards, there is but a small difference in the low-quality test mode. There is no difference as soon as you turn on full-screen antialiasing and anisotropic filtering.
The last real-time strategy on my list is Stormrise. It produces the same results as BattleForge but we can see some growth of the bottom frame rate.
Once again the single Radeon HD 5870 does not really need an advanced CPU. The Intel Core 2 Duo at the reduced frequency of 2.67GHz is the only CPU to lag behind the others in the low-quality test mode. But at 3.4GHz this CPU does not limit the performance of today’s fastest single-chip graphics card. The Radeon HD 5870 CrossFireX tandem depends on the CPU in the low-quality test mode. Take note that the quad-core CPUs perform well here: the AMD Phenom II X4 at 2.67GHz equals the Intel Core 2 Duo at 4.1GHz. The Intel Core i7 at 2.67GHz is not slower than the Phenom II X4 at 4.1GHz. But when we turn on full-screen antialiasing and anisotropic filtering, the graphics subsystems don’t depend on the CPU and the frame rate is limited by the capacity of the CrossFireX tandem.
The Radeon HD 5870 does not show much dependence on the CPU in this game. The difference between the platforms is only noticeable at 1280x1024, but there is no point in choosing a faster CPU because the frame rate is already high enough. The CrossFireX tandem is clearly limited by the CPU.
This is the only test where the Intel Core i7 at 2.67GHz is the slowest CPU with the single graphics card. However, it goes ahead at the first step of overclocking. The single Radeon HD 5870 depends on the CPU until 1920x1200 at the high-quality settings. The two Radeon HD 5870 cards in CrossFireX mode respond eagerly to the overclocking of the three CPUs in every test mode and resolution.
The third and the most resource-consuming scene of the benchmark integrated into Resident Evil 5 shows that the overclocked dual-core CPU is quite enough for a single Radeon HD 5870. And when you’ve got two such cards in CrossFireX mode, an Intel Core i7 920 at its default frequency will suffice.
I carried out the power consumption measurements using a special PSU. To create maximum load I launched FurMark 1.7.0 in stability check mode at 1920x1200 and Linpack x64. These two programs load heavily the graphics card and CPU, respectively, so we can determine the peak power draw of the whole system. You can see the results in the diagram:
Everything is normal here except for the results of the AMD Phenom II X4 platform overclocked to a CPU frequency of 4.1GHz. This platform needs a power supply with almost as much wattage as you need for the platform with an Intel Core i7 overclocked to the same frequency.
Now, I’ll try to sum everything up. First of all, the single Radeon HD 5870 does not depend as much on the CPU as it is supposed to. According to my tests (the left part of the diagrams), the Radeon HD 5870 is quite satisfied with an overclocked dual-core CPU manufactured two years ago. And you can even leave the CPU at its default frequency at the high-quality settings and 1920x1200. The only exceptions are Left 4 Dead and Warhammer 40000: Dawn of War II. The former game is not a problem for modern top-end Radeons while the latter, on the contrary, calls for a quad-core CPU, preferably from Intel.
The Radeon HD 5870 CrossFireX configuration is more CPU-dependent than the single such card. But it is not possible to say what CPU is sufficient for this graphics subsystem as this depends on the particular game, settings and resolutions. For example, at 1920x1200 with 4(8)x FSAA and 16x AF an Intel Core i7 at 2.67GHz is quite enough for S.T.A.L.K.E.R.: Clear Sky, Far Cry 2, BattleForge, Tom Clancy's H.A.W.X. and Resident Evil 5 as well as for Unigine Heaven Demo. In the other games and semi-synthetic benchmarks the Radeon HD 5870 tandem calls for a faster CPU. The question is whether you really need those 150-200 frames per second.
I also want to say a few words about the platform with the dual-core CPU as compared with the quad-core CPU. Winding up this test session I carried out an interesting experiment. I connected two system cases with a Radeon HD 5870, one with an Intel Core 2 Duo at 4.1GHz and another with an Intel Core i7 at 4.1GHz, simultaneously to the same monitor. Then I selected the most CPU-indifferent resolution of 1920x1200 with FSAA and AF and launched a few games. I switched between the two systems by choosing the input on the monitor (analog/DVI-I) and used two mice and two keyboards.
Using this setup, I could quickly switch from the system with a dual-core CPU to the system with a quad-core CPU and feel the difference between them in the same parts of five games. Excepting S.T.A.L.K.E.R.: Clear Sky, I easily felt the difference such games as Tom Clancy's H.A.W.X., Resident Evil 5, Call of Duty 5: World at War and Crysis. The gameplay was much smoother on the quad-core CPU, without occasional jerks as on the Core 2 Duo. Game levels were loaded faster, too. So, the numbers don’t always provide the full picture, yet I hope that this test session will help you decide what CPU you need for one or two Radeon HD 5870 cards.