Benchmarking Results

Intel Core2 Extreme QX9650 Overclocking Testbed
Processor Intel Core2 Extreme QX9650
Quad-core, 3.0GHz, 2x6MB L2 Cache, 9x Multiplier, 1333FSB
Comparison Processor Intel Core2 Extreme QX6850
Quad-core, 3.0GHz, 2x4MB L2 Cache, 9x Multiplier, 1333FSB
Motherboard Asus P5E3 Deluxe / BIOS 0703
Memory OCZ DDR3 PC3-14400 (DDR-1800) Platinum Edition
Memory Settings 7-7-7-15 (DDR-1600 ~ DDR-1700), 2N
Cooling D-tek FuZion CPU water block
EK FC8800-GTX/Ultra full coverage GPU block
ThermoChill PA120.3 radiator
Dual Laing D5 pumps in series
1/2" ID (3/4" OD) Tygon tubing
6x Yate Loon D12SL-12 120x25mm fans @ 12v
Power Supply Corsair TX650W
Video Card MSI 8800 Ultra (660/1050)
Video Driver NVIDIA ForceWare Release 169.12 (beta)
Hard Drives 2x Western Digital 150GB (RAID 0)
10K RPM SATA 3Gbps 16MB Buffer
Operating Systems Windows XP Professional SP2 (2GB)
Vista Ultimate 64-bit (4GB)

No overclocking review is complete without sharing the results of some standard benchmarking suites. The 3DMark series of benchmarks, developed and provided by Futuremark, are among the most widely used tools for benchmark reporting and comparisons. We measure Intel Core 2 Extreme QX9650 general graphics performance using 3DMark06, and because results from Windows XP and Vista are considerably different, we include both.


Core
2 Extreme QX9650 General Graphics Performance

When we compare these results to those taken with a 65nm QX6850, we see clock-for-clock gains of less than 1%. However, the 45nm QX9650 is able to reach a much higher overclock of 4.4GHz, netting us enough improvement to push above 16,000 3DMarks without even altering our graphic card's stock clocks. At these speeds 18K+ with a single overclocked 8800GTX/Ultra is entirely within the realm of possibility (a score which we were in fact able to achieve).


45nm
vs. 65nm General Graphics Performance

PCMark Vantage, Futuremark's latest system performance benchmarking suite, picks up where PCMark05 left off. Years in the making, this program offers the ability to test every conceivable aspect of total system performance either individually or all at once. Since Vantage only targets one operating system (Vista), we take advantage of the 64-bit executable included in the package running it in favor of the 32-bit file. Because Vantage (x64) is more of a complete benchmarking suite than the other tests we run, it makes sense that more than just pure CPU computing power influences results. For that reason, we have also shown the FSB settings used for each run. The additional memory bandwidth generated by the higher memory speeds does have an effect.


Core
2 Extreme QX9650 General System Performance

CINEBENCH R10 is another benchmarking program that performs radically different under both Windows XP and Vista. The developers at MAXON have done a tremendous job optimizing the code for 64-bit computing and it shows - unlike the graphics testing performed above, we find Vista maintains a clear advantage. As we can see, CINEBENCH R10 scales beautifully with additional cores and increased frequencies.


Core
2 Extreme QX9650 Rendering Performance

When placed together with results collected with a QX6850, the performance improvements offered by the QX9650 become immediately apparent. In this case, the Enhanced Core 2 processor is able to hold a commanding lead, thanks mainly to the massively improved frequency scaling.


45nm
vs. 65nm Rendering Performance

TechARP recently published a great x264 benchmark that essentially measures the time it takes a test system to encode a short, DVD-quality 720x480 progressive MPEG-2 video stream to a compressed, high-quality x264 file. Results are provided in the form of frames per second, and together with the known source file frame count it is possible to calculate the total encode time (shown below in seconds). The x264 CODEC makes very efficient use of multiple cores and is highly dependent on available processing power. As such, the benchmark makes a great choice for highlighting CPU performance. Although the CODEC does not yet include SSE4 optimizations, a new feature introduced with Penryn, the QX9650 has no problems making short work of our QX6850.


45nm
vs. 65nm General Encoding Performance

We hope to expand future testing to include real-world gaming results from some of the newest titles like Crysis, Call of Duty 4: Modern Warfare, Unreal Tournament 3, and Gears of War. Stay on the lookout for these results and others in our next look at the QX9650 when we pair this capable processor with the best motherboards companies like ASUS, Gigabyte, MSI, abit, DFI and Foxconn have to offer.

Overclock That CPU Final Thoughts
Comments Locked

56 Comments

View All Comments

  • Griswold - Thursday, December 20, 2007 - link

    I think the other explanation for the delay is much more likely than this horsecrap.
  • tomoyo - Wednesday, December 19, 2007 - link

    This is a great article from both a readability and technical standpoint, I found it very enjoyable to actually read the commentary, which I cannot say is true for 95% of tech reviews. Also the graphs and information presented were extremely useful. I love the demystifying of the vdroop issue to prove that it's a GOOD thing.
  • Amiteriver - Wednesday, December 19, 2007 - link

    Finally a review with some understandable explanations of Bios settings and there realation ships. As the motherboards change you get new Bios names for things and your eyes glaze over wondering what the heck some of the terms mean. Look forward to future articles of popular motherboards Bioses. I will be printing this article out. Thanks much
  • Amiteriver - Wednesday, December 19, 2007 - link

    Finally a review with some understandable explanations of Bios settings and there realation ships. As the motherboards change you get new Bios names for things and your eyes glaze over wondering what the heck some of the terms mean. Look forward to future articles of popular motherboards Bioses. I will be printing this article out. Thanks much
  • Sczee - Wednesday, December 19, 2007 - link

    Quite possibly the best article on computer hardware I have ever read. I learnt a lot about overclocking from it and plan to put it to use tonight. I'll be looking out for your next piece Kris.
  • Bozo Galora - Wednesday, December 19, 2007 - link

    ya, the X48 is being delayed because some top tier mobo manuf needs to get rid of its X38 inventory. My guess is Asus, because they have the muscle to influence Intel. Bleh.

    So I bought all this super xmas bargain stuff, and now I am going to have to let it gather dust - lol

    http://www.digitimes.com/mobos/a20071217PD205.html">http://www.digitimes.com/mobos/a20071217PD205.html
  • Bozo Galora - Wednesday, December 19, 2007 - link

    and the X48 does seem to have a 450-460 cap, at least so far

    http://en.hardspell.com/doc/showcont.asp?news_id=1...">http://en.hardspell.com/doc/showcont.asp?news_id=1...
  • bryanW1995 - Wednesday, December 19, 2007 - link

    Best article I've read here in a long time. Good job.
  • Regs - Wednesday, December 19, 2007 - link

    So lets say I have a 65nm Core Duo running between 0.850V-1.3525V. These are the product specs, which I guess .850 is the low limit and 1.325 is the high limit. Why does the voltage have to decrease depending on load? Is it just as simple as "supply and demand"? How does running the CPU at 1.352v run the risk of instability?
  • Aivas47a - Wednesday, December 19, 2007 - link

    Great article. You guys have really been distinguishing yourselves with in-depth work on overclocking the last few months: exploring obscure bios settings, tinkering with "extreme" cooling -- keep it up!

    My experience with a qx9650 so far is very similar to yours: easy scaling to 4 ghz, difficult scaling after that with 4.2 ghz being the practical max for regular operation (folding, etc.).

    One issue I will be interested to see you address in the future is fsb overclocking on yorkfield. So far I am seeing yorkfield top out at lower fsb (450-460) than was possible for kentsfield on a comparable P35 or X38 platform. That is not so significant for the unlocked Extreme Edition chips, but could make it difficult to achieve the magic 4 ghz with the q9550 and especially the q9450.

Log in

Don't have an account? Sign up now