The title of the article may have confused you, but it's easy to explain: indeed, Core 2 Duo E6850 is the latest update in the Core 2 Duo family based on the 65-nm core. That is it's a new processor based on the old core. On the other hand, Core 2 Duo E8200 operates at 2.66 GHz (this frequency has been conquered by this family long ago), even though it's based on the new 45-nm core. That is we have the new core operating at the old frequency. We already wrote about peculiarities of the new core. So we'll not repeat ourselves and proceed straight to our tests. Unfortunately, we haven't obtained test results of the Core 2 Duo E6750 yet, which is the ideal competitor to the E8200. So we compared the latter with the X6800 (a dual-core processor based on the old core, but operating at a higher clock rate) and the QX6700 (the same frequency, the old core, but four cores). By the way, looking at our comparison, some conclusions are evident even without test results of the E6750. What concerns the E6850, it should be compared with the QX6850 (the same architecture, but it has two cores versus 4 cores). That's exactly what we did.
* "2 x ..." means per core;
** For AMD processors this is memory controller bus clock rate;
*** Measured differently for Intel and AMD processors; impossible to compare directly.
Windows XP Professional x64 edition SP1
3ds max 9 x64 edition
Maya 8.5 x64 edition
Lightwave 3D 9 x64 edition
MATLAB R2006a (188.8.131.52) x64 edition
Pro/ENGINEER Wildfire 2.0
Photoshop CS2 (9.0)
Visual Studio 2005 Professional
Apache HTTP Server 2.2.4
CPU RightMark 2005 Lite (1.3) x64 edition
7-Zip 4.42 x64 edition
FineReader 8.0 Professional
Monkey Audio 4.01
OGG Encoder 2.83
Windows Media Encoder 9 x64 edition
Canopus ProCoder 2.01.30
Windows Media Video VCM 9
Half-Life 2 1.0
Quake 4 1.3
Call of Duty 2 1.2
Serious Sam 2 2.07
Supreme Commander 1.0.3220
Essential Foreword to Charts
Our test method has two peculiarities of data representation: (1) all data types are reduced to one—integer relative score (performance of a given processor relative to that of Intel Core 2 Duo E4300, given its performance is 100 points), and (2) detailed results are published in this Microsoft Excel table, while the article contains only summary charts by benchmark classes. We will nevertheless focus your attention on detailed results, when needed.
3D Modeling and Rendering
Quad-core processors were destined to win in this group of tests because of rendering. It's strange to see the E6850 fail to outperform the X6800 despite its higher clock rate and faster bus. Results of the E8200 are quite expectable: even its 1.5 times as large L2 Cache fails to make up for the clock rate lower by 333 MHz.
The number of cores is practically irrelevant in this group of tests, because none of such applications supports multiple cores (this feature is partially supported by MATLAB, but we didn't see it in practice—perhaps, the built-in benchmark uses "wrong" functions). So the QX6700 compares well with the E8200... and we can see that CADs are quite indifferent to the new core. The QX6850 is faster than the E6850 by 2%, but we are inclined to write it off to a measurement error.
Digital Photo Processing
Adobe Photoshop again demonstrates excellent support for multiple cores (including quad-core processors), so the QX6850 outperforms the E6850 by 21%. That's the first time when we see some effect from using the new core: the E8200 even outperforms the old extreme X6800 by one point, although the latter operates at a much higher frequency.
Equal results of the QX6850 and the E6850 indicate that either Visual Studio 2005 cannot use more than two cores very well... or we have to change our compilation project. However, let's not put a cart before a horse—this task is for the next version of our test procedure. Equal results of the X6800/E8200 illustrate advantages of the new core.
Quad-core processors are traditionally "retarded" in this test, so we may ignore them here. The fast bus of the E6850 helps this processor heavily outperform the X6800. However, the E8200 with the same bus and larger cache just catches up with the X6800.
In our article about the new 45-nm core we already noted its significant advantage in CPU RightMark. This chart is another proof: the E8200 outperforms the E6850 by 15%!
Quad-core processors do not demonstrate remarkable results here—we already wrote that to all appearances such software could use two cores at best. The E8200 based on the new core performs very well: it even outperforms the E6850 a little.
We don't see any advantages of the new core here.
It's the old group of tests, which have lost their relevance because of high predictability of their results. No comments.
The advantage of the new core is noticeable, though not that apparent: without it, the E8200 would have lost even more seriously.
Fantastic results of the new core do not need any comments.
See how big a difference exists between the professional and home scores in the comparison of E6850 versus QX6850. In the first case, the E6850 is outperformed by 13%, in the second case—by just 3%. Considering our criteria for grouping tests, the answer to the question about using quad-core processors in home computers is pretty much obvious. The new core represented by the E8200 puts up excellent performance: its total score is similar to that of the X6800, which leaves no doubts about its competition with the E6700/6750.
Estimated Power Consumption
We have predictable results here, we are only a little confused by higher power consumption of the E8200 (versus the E6850) in idle state. But we have no statistics on dual-core processors based on the 45-nm core, so this artifact can be anything, including peculiarities of a given sample.
On the whole, we've said everything in our comments to final charts. Even if we judge by the current prices, dual-core processors on the 45-nm core look much more promising than old 65-nm processors. According to the two-year old trends, this fact is a sign of coming purification of the product range of Core 2 Duo processors.