The Clarkdale Experiment: Mea Culpa
by Loyd Case on May 4, 2010 1:57 PM ESTThe Game is Afoot
I ran a number of games, plus one synthetic benchmark. All tests were run at two different resolutions and detail settings: 1440x900, low graphics settings, and 1920x1080 at fairly high settings. The first, low resolution set of tests represent performance weighted towards the CPU, and are not graphics bound. The second set of tests represent playable resolutions and detail settings, and can be graphics bound in some cases.
· Unigine Heaven (synthetic). The 1440x900 benchmark was run in DX10 mode, the 1920x1080 benchmark run in DX11 mode, with hardware tessellation enabled.
· S.T.A.L.K.E.R.: Call of Pripyat. The lower resolution test was run with the low detail, dynamic presets. The 1080p benchmark was run in DX11 mode, high preset, with hardware tessellation enabled.
· Far Cry 2: Action benchmark. This is one of the included scenarios with the Far Cry 2 benchmark tool, and is pretty CPU intensive – but also represents the game’s more combat-intensive scenes. The low preset was used for the 1440x900 test, the high preset (DX10 enabled) with the higher resolution test.
· Battle Forge. This RTS is CPU intensive, but also implements DX11 graphics in the highest detail modes.
· Tom Clancy’s H.A.W.X. This title is an arcade styleflight sim, with lots of DX10 eye candy.
· DiRT2 (Demo): This rally racing title supports DX11 and DX9, but not DX10.
Let’s take a look at the results.
When we keep the resolution and detil levels low (tessellation is off in the low resolution test), the CPU differences are noticeable. Once we dial up the graphics pain, though, the difference is negligible.
Heaven is a synthetic test that pushes certain DX11 features. Let’s see how real games fare.
There are some intriguing tidbits in the data. Some of them are a little head-scratching.
Remember, if we run a game at low resolution, with detail levels dialed down, then the game is more likely to be CPU bound than graphics bound. In other words, the CPU impact is exaggerated. If we just look at the low detail results, we see clear wins across the board for Lynnfield. Clearly, if we only ever play games on small displays with crappy visuals, Lynnfield wins out.
The head-scratching part comes in when we dial up the graphics pain. In theory, as you dial up resolution and detail levels, the graphics hardware plays a bigger role, and differences in CPU performance become less noticeable. That’s not true with radical differences in CPU – a slow Celeron compared to a fast Core i7 would run games much more slowly, no matter what GPU is installed. Still, with somewhat similar classes of processors, you’d expect minimal differences.
What we see here, though is a mixed bag. The Far Cry 2 Action scene, for example, still demonstrates a noticeable performance edge for Lynnfield, while we see small, but not insignificant, differences in favor of Clarkdale on Battle Forge and HAWX. Still, the overall edge lies in favor of Lynnfield, with the slower clocked quad core CPU winning in Far Cry 2, DiRT2 and STALKER. These results are pretty much repeatable with multiple runs, too. The numbers vary slightly, but the pattern doesn’t change.
45 Comments
View All Comments
doron1 - Tuesday, May 4, 2010 - link
There is a huge disparity between the advertised results in this article's Far Cry 2 and Toms Hardware's ones, in there there is practically no difference between six (!!) different cpu configurations (including some overclocked results) and the gpu used is the same hd5850..See for yourself
http://www.tomshardware.com/reviews/core-i3-gaming...
Hope you get this cleared up as I'm kinda confused here..
FATCamaro - Tuesday, May 4, 2010 - link
Mistake definitely. He has messed up somewhere, but given his amateurish level it is hard to predict where.nubie - Tuesday, May 4, 2010 - link
I am quite confused, I thought the difference between a large system and a small system was solely the motherboard, so perhaps benching them with an identical CPU would have made more sense.You can get a smaller PC and still use a full size ATX motherboard and graphics card http://www.gtrtechcorp.com/gt3-info/product-info (nearly twice as tall, but also about 1/4 the width)
In fact the GT3 case comes with the PSU and is only $10 dearer than the case you chose, knocking a lot off your budget. The only concession I see is that you must use a laptop-style optical drive. This is made up with the ability to run dual secondary storage drives, so you could have a small (read inexpensive) SSD for the system and a large standard drive for media/games.
That system doesn't really nail any desirable system build metric (It should be able to hit one for >$2k). The common saying is size, performance, price, pick 2. This loses on being small (you can build using a case of half the volume with a full ATX Motherboard). Performance wise it is stuck using a different chip with concessions to performance made in the name of integrated graphics (I guess this is technically Intel's fault, they haven't released a performance version of these processors, and I assume you can put the faster chip in the small PC build?). That leaves the cost of the system. Too much. I would budget about $800 for a mini-gaming system, and it would hit nearly the same performance, enough that you wouldn't notice in-game between the two systems.
This leaves the same taste in my mouth as Tim Allen's ~$60k front-wheel drive Cadillac with 400-hp, lukewarm, but I can appreciate parts of it intellectually. http://www.caranddriver.com/reviews/car/00q4/tim_a...
(I realize it is supposed to be a "quiet" system, sort of. I think there may be a way to design your own case with a single central fan cooling the PSU/GPU/CPU quite nicely. That may be more my personal interest/taste than the builder of this system.)
ClagMaster - Tuesday, May 4, 2010 - link
Nice experiment that held no surprises for me.The i5 750 with four real cores and 8 MB of L3 cache is hard to beat for $190.
For gaming or media multi-threaded applications the i5 750 is going to beat the i5 661. Four real cores is better than two real cores and two make-believe cores. Memory latency is much better on the i5 750 because of the on-die memory controller. I think this helps performance significantly despite the lower clock frequency. The memory controller on the i5 661 is on the graphics unit which has to operate even if it’s not providing graphics processing. So is the PCIe controller for a discrete graphics card. That’s why the power consumption of the i5 661 at 87W for two cores is relatively higher than the 95W for four cores of the i5 750. Hyperthreading is much "hyped" with gains perhaps 10%.
The i5 750 or i7 860 is my first choice for a mainstream gaming rig if I were to build one today – hands down. I would build this on a Gigabyte GA-P55A-UD3P Mobo, 8GB of G.SKILL ECO DDR3 1600 (1.35V), a DIAMOND 5850PE51G Radeon HD 5850 1GB graphics card and a couple of Western Digital Caviar Black WD1002FAEX 1TB harddrives.
For Home theater or Office PC where I need only 2 CPUs, the AMD Athlon II 250 ($60) with a 890G motherboard ($130) is a much better value and adequate performance than a i5 661 ($210) on a H57 motherboard ($120). The 890G provide better graphics. The SB850 provides native SATA 6G with PCIe 2.0 connectivity while the H55/H57 provides SATA 6G with PCIe 1.0 connectivity.
ClagMaster - Tuesday, May 4, 2010 - link
One other thing.If Intel really wanted to make me happy with Clarksdale, they should get rid of that worthless GPU and replace the freed real estate with a four-core 32nm processor operating at 3.2 Ghz as an i5 750 replacement. Then the memory controller and PCIe controller would have much less latency and this processor should be able to operate at 65W.
Jalek99 - Tuesday, May 4, 2010 - link
Too many variable in play. Wouldn't the ASUS board support the other processor, making a true test of the different CPU's and not the bus performance of the boards among other things?vol7ron - Tuesday, May 4, 2010 - link
I like the article and the idea.My biggest problem is that the motherboards and HD can play a huuuuge role in benchmarks. You wouldn't think HDs would, since once the data is loaded into memory, you'd think it'd be insignificant, but it isn't.
Motherboards, on the other hand, are known to have a devastating impact on results. Especially when it comes to accessing RAM. Even at stock timings and settings, motherboard manufacturers have been known to optimize data access paths, to give off a sense of "turbo."
It's tough to say that neither of the above would be influential in the benchmarks, especially when the CPUs are similar.
jonup - Tuesday, May 4, 2010 - link
Most MBs on the same chipset (H55 and P55 perform very close) perform similarly. There are synthetic mamory variances that usually do not materialize into real world performance.Only time MBs on the same chipset perform differently is when the manufacturers decide to OC at default settings. But the reviewer should have made a note if that was the case - it only takes running CPU-Z to notice that the CPU is OCed by 50-60MHz.
Jalek99 - Wednesday, May 5, 2010 - link
But why leave those variables in play? From what you're saying, you could bench an ASUS against a PC-Chips board and expect the same results. Maybe you could, but it seems a minor thing to swap processors to isolate one variable.vol7ron - Wednesday, May 5, 2010 - link
I've seen the same chipset with almost the same cpu-z settings (speeds, timings, proc) result in very different benchmarks.I'm not saying that they are unequal in this particular test, but I am saying that this is a variable that should not be overlooked.
Bottle necks can be created from almost any mechanism of a pc (eg ram, gpu, hd), even the psu could cause a performance impact due to the variation in power efficiency, though this is often minimal. Small things add up, though.
vol7ron