I wanted to make full use of the unlocked multiplier on the QX6700, so I am including several clock speed variations of Kentsfield. This will give us the indication of what to expect when other clock speeds are released in the near future.

For instance, moving the multiplier down to 9x allows us to run the Kentsfield as a "Core 2 Quad Q6600" which is expected to be released in January. Furthermore, increasing the multiplier to 11x gives us a clock speed of 2.93 GHz, which matches the Core 2 Extreme X6800. This will tell us exactly how much benefit we're getting from running 4 cores instead of 2, completely eliminating clock speed as a factor.

On the AMD side, the X2 4600+ was the fastest Athlon64 available in stores, so that's what I used. Since the 5000+ is finally showing up in stores, I have overclocked our 4600+ to 2.60 GHz, matching the clock speed of the 5000+. However, memory timings had to be lowered, and the HT is running slightly higher than it would at stock speed. The tests we're performing here are much more reliant on CPU speed however, so these did not come up as an issue.

To show how similarly priced Core 2 Duo processors compare to these AMD parts, I lowered the multiplier of the Conroe chip to match that of an E6600 (2.40 GHz) and E6400 (2.13 GHz). Both of these models sit right between the 4600+ and 5000+ in price (the E6400 is even cheaper than the 4600+, and the E6600 beats the 5000+ by about $100).

Intel Motherboard: Gigabyte GA-965P-S3 (Reviewed here)
AMD Motherboard: Foxconn C51XEM2AA-8EKRS2H (nForce 590 SLI) (Reviewed here)

Memory: 2 GB OCZ PC2 6400 SOE Urban Elite 3-4-4-15 (Reviewed here)
Video Card: GeForce 7600 GT 256MB
Audio: Onboard Audio (disabled in all tests but SYSMark)
Hard Drive: WD Raptor 36GB 10K RPM SATA

Windows XP 32 bit with SP2 was used. All the latest chipset, video, and sound drivers were used. SMP was enabled in all tests when possible.

Thanks, as always, to Thermaltake for providing us with the power supply for all our testing. We're using their modular ToughPower 750W now, and it hasn't skipped a beat.

I have designed the graphs to tell multiple stories, depending on what you're looking for. You should pay particular attention to the following:

• Compare the top 2 numbers; this will tell you how the two top Extreme Edition models compare to one another officially.
• Compare the X6800 @ 2.93 GHz to the Kentsfield @ 2.93 GHz - this will tell you exactly how much benefit the two extra cores are giving.
• Compare the Q6600 to the E6600. This will tell you how much benefit four cores gives you at a lower clock speed.
• Compare the two AMD processors to the 2 bottom Intel processors. This will tell you how AMD's fastest models (FX notwithstanding) compare to Intel's middle-range Conroe CPUs.
• Orange bars always represent Intel quad cores. Blue represents Intel dual cores. The green bars are for the AMD dual core chips.

With all that in mind, let's get on with the show!

As always, we kick things off with SiSoft Sandra's CPU arithmetic and multimedia benchmarks. This gives us an idea of the raw computational power of each CPU before moving on to real world performance.

As you can see, SiSoft Sandra's CPU arithmetic tests make full use of all 4 cores on Kentsfield. This allows each quad core to exactly double the respective dual core result. It looks like the shared cache and oversaturated FSB do not affect performance so far.

Same thing goes for the CPU multimedia tests.

Next Page: (PCMark 05; CineBench 9.5)

Page 1: Introduction Page 2: Test Config Page 3: PCMark 05; CineBench 9.5 Page 4: Video Encoding Page 5: Audio Encoding Page 6: Gaming Page 7: SYSMark 2004 SE Page 8: WinRAR Performance; Conclusion