|Introduction and Configurations|
Though we're running the risk of sounding a bit platitudinal, this is definitely the IDF moment we're sure many of you have been waiting for. We've reported to you on Intel's various technology announcements and architectural initiatives over the past few days. And, from what we could garner from our sessions with Intel at this fall's IDF, the company certainly has a lot of bleeding-edge technology on tap that by all rights should dramatically affect the computing industry landscape for years to come.
We've heard from both sides of the fence, nearly all year long, that the age of the multi-core processor is upon us and multi-threaded applications are the wave of the future. The benefits of SMP (symmetric multi-processing) have been know for many years now but until only recently have desktop applications and usage models been able to take advantage of multi-thread capable systems. The advent of the dual-core X86 desktop processor has ushered in a whole new era of computing architectures, performance levels, capabilities in software and of course it doesn't just end there.
Intel's quad-core processor, code named "Kentsfield" is going to be launched in volume this November. It is targeted at raising the bar once again, potentially doubling the performance level of the Core 2 Duo architecture that was marked clearly as the most successful processor product launch in Intel's long history. Kentsfield is in fact an integration of two dual-core Core 2 Duo die on a single package substrate and those cores then share a common front side system bus. That's a pair of dual-core CPUs in a single LGA775 socket and on paper the specs are more than promising.
But can we really exploit the available throughput of a quad-core Intel Core 2 processor? Will this new high-end CPU aimed to woo the hearts of the computing enthusiast, with its subtly-sexy "QX" branding, really make good on its perceived benefits in the benchmarks and real-world usage scenarios? Intel allowed us some behind closed doors private time benchmarking systems based on their new Kentsfield Core 2 Extreme QX6700 processor. The following pages are intended to offer you some insight into what a "controlled" benchmarking effort with Kentsfield was able to yield for us in various performance metrics. Though Intel furnished us with all the test system configured in the following benchmarks, we're very confident in the numbers, as we were at last year's IDF when we first saw Core 2 Duo in action. This time around, it's a quad-core assault.
Intel's own D975XBX2 "BadAxe" motherboard was configured as a test-bed for our Kentsfield QX6700 test system. The machine was well endowed with a dual-GPU GeForce 7950GX2 graphics card in a single PCI Express X16 slot and a 2Gig Corsair PC8500 DDR2 RAM kit. On-board integrated audio was utilized and a single Seagate Barracuda 7200.10 320Gig drive was configured with Windows XP SP2. On with the benchmarks...
|PCMark05 and 3DMark06|
Initial testing was performed on the two Intel-configured machines with Futuremark's PCMark05 benchmarking tools suite. This is a synthetic benchmarking package with basic scripted computational work loads of common functions performed by the average desktop user.
"The CPU test suite is a collection of tests that are run to isolate the performance of the CPU. The CPU Test Suite also includes multithreading: two of the test scenarios are run multithreaded; the other including two simultaneous tests and the other running four tests simultaneously. The remaining six tests are run single threaded. Operations include, File Compression/Decompression, Encryption/Decryption, Image Decompression, and Audio Compression" - Courtesy FutureMark Corp.
The quad-core Core 2 Extreme QX6700 is only showing about a 14% performance advantage over the dual core X6800 chip in the base CPU test module. We should note that an Athlon 64 FX-62 dual core processor scores around 5700 in the PCMark05 CPU test module.
The overall score actually shows the QX6700 slightly slower than the dual core Core 2 chip. Memory bandwidth with the quad-core system, is going to be undoubtedly slightly lower on a per-core basis, since four cores versus two are sharing front side bus bandwidth. Thus the lower overall weighted score. That and coupled with fact that the Core 2 Extreme X6800 has a slightly higher clock speed on each of it's dual cores and it becomes more clear how the numbers have scaled here in these synthetic benchmark runs.
3DMark06's built-in CPU test is a multi-threaded "gaming related" DirectX metric that's useful for comparing relative performance between similarly equipped systems. This test consists of two different 3D scenes that are generated with a software renderer, which is dependent on the host CPU's performance. This means that the calculations normally reserved for your 3D accelerator are instead sent to the central processor. The number of frames generated per second in each test are used to determine the final score. We've got the results from the CPU test and the overall score here...
Futuremark's 3DMark06 CPU test paints a dramatically different picture however. The quad-core Core 2 Extreme QX6700 boasts nearly 2X the performance number that is reported for the Core 2 Duo X6800 system, a dramatic increase to be sure. For another frame of reference, an Athlon 64 FX62 scores around 2090 in the 3DMark06 CPU test module.
The overall 3DMark06 score is weighted in with graphics subsystem performance as well. So the numbers are much tighter here since identical graphics cards were used in the test systems. Regardless, in 3DMark's DX9-based gaming environments, the QX6700 is around 8% faster even though it's held back more so by the graphics cards installed.
|POV Ray and 3ds Max|
POV-Ray, or the Persistence of Vision Raytracer, is a high-end open source tool for creating 3D graphics artwork. We tested with POV-Ray's standard included benchmarking model on both test machines and recorded the scores reported for each. Results are measured in pixels-per-second throughput.
Kentsfield flexes its muscles again and the quad-core QX6700 is some 80+% faster than the dual-core X6800-based system. Those of you into 3D rendering and artwork generation will clearly see dramatic performance enhancements with Intel's new quad-core Core 2 Extreme QX6700.
3ds Max 8 from Autodesk offers 3D modeling and rendering capabilities and it's an industry standard tool for game development. Modeling, rendering and animation are all part of the software's feature-set but we're focusing on the most CPU intensive rendering requirements in our 3ds Max benchmark.
Again, the 2.66GHz quad-core Core 2 Extreme QX6700 shows itself to significantly faster than the 2.93GHz dual core Core 2 chip in 3ds Max rendering workloads; on the order of about 63% faster to be more specific.
|DivX, Sony Vegas and our Conclusion|
DivX video compression and encoding is next and it's an excellent vehicle for testing raw CPU throughput in a very popular usage model for the average end user.
In digital video compression and encoding with DivX, the Core 2 Extreme QX6700 shows a 40+% performance advantage over the dual-core X6800 though it has roughly a 300MHz clock speed deficit.
Sony Vegas is a digital video editing package from Sony Media Software and our testing with it helps expand on the general coverage of digital video manipulation and playback. Vegas mixes audio and video files as well for the final output, versus DivX which is simply digital video conversion and encoding. In this test the quad-core QX6700 is about 50% faster than the dual core Core 2-based system -- another impressive performance gain to say the least.
That wraps up our Kentsfield performance coverage for now. However, we'll be back in the coming weeks with a full HotHardware analysis of Intel's new quad core processor line-up, of that you can be certain. Obviously we have only scratched the surface in terms of exploring the performance profile of the new Core 2 Extreme QX6700 processor. Though Intel only let us run a few synthetic tests, as well as a few 3D rendering and digital video processing tests, there are many more popular application and game titles that we intend on running the QX6700 through its paces with in the weeks ahead. In summation of the data we've presented you with here today, our early observations tell us that Intel's new quad-core Kentsfield chip scales anywhere from 50 - 100% faster than a dual-core Core 2 chip, clock for clock, depending on the type of applications you'll be running.
Kentsfield's performance was as expected to be honest but obviously applications that make better use of multithreading are where the major gains will likely be seen. We don't expect Kentsfield to run standard Office applications that much faster for example, unless there's a lot of multi-tasking thrown in the mix, but that almost goes without saying. Also we know many of you would like to know what the new Intel quad-core architecture can do for various gaming scenarios and we plan to expand on that significantly in our follow-up coverage. On a side note, we're aware that AMD's 4X4 product offering is expected to hit our test bench sometime in Q4. It will be very interesting to compare what four cores sharing a Hypertransport link can do versus four Core 2 cores sharing a single front side bus. We'll be here burning the midnight oil when the time comes and we're hearing that time will be soon. For now Kentsfield looks more than promising for the enthusiast, power user and workstation professional.