|Introduction and Specifications|
To say this year has been a bit of a whirl-wind adventure would be an immense understatement. We've seen Intel move from a back-seat position versus the Athlon 64 X2 with their Pentium D architecture, to commanding virtually all of the lime-light as of late with Core 2 Duo, not to mention running away with nearly all of the performance benchmark wins. If 2006 was the year of the dual-core CPU, then Intel left its mark deep in PC processor history as probably the biggest come-back story the industry has ever seen. Core 2 Duo's steep volume ramp in the market was also indicative of the success of the new processor, which was marked as the strongest in the company's history. Speaking of which, "history" seems to be a recurring theme here. If Intel left their mark on the year of the dual-core processor, they're about to literally usher in the year of the multi-core Desktop system with the new Core 2 Extreme QX6700 processor we have on our test bench today. History in the making? Perhaps. But that's what cutting-edge semiconductor technology is all about -- moving forward at a relentless pace, never pausing to look back.
Today Intel is officially launching a new Quad-Core processor, formerly known by the company's internal code name, "Kentsfield". Kentsfield is the direct result of Intel's immense manufacturing and process technology prowess. Based on the company's 65nm manufacturing process, one of the very few fabs of its kind in high volume production at this time, the Core 2 Extreme QX6700 is an integration of two Core 2 Duo dual-core processors on a single substrate for a total of 4 CPU cores in a single LGA 775 socket. Targeted at serious PC Enthusiasts and Workstation Power Users, this new quad-core CPU is poised to set new land-speed records in benchmark test suites. On this first day of November (PST, for our East Coast readership), Halloween may be over but the gory details of Intel's Quad Core Assault are next.
Over the course of the past year or so, we have posted a wealth of information related to Intel's Core microarchitecture and Core 2 Duo and Extreme processors here at HotHardware.com. For more background on the technologies employed by the Core microarchitecture and Intel's platform as a whole, we suggest taking a look at few of these related articles. They contain detailed explanations of some of the features common to Intel's legacy products, compatible chipsets, and the new Core 2 Duo and Core 2 Extreme processors:
We cover some specifics regarding Intel's 65nm manufacturing process in our 955XE / i975X evaluation and outline Intel's AMT (Active Management Technology) and IVT (Intel Virtualization Technology), among other things, in our July Core 2 Duo Performance and Chipset update. The other articles listed above will also give you some background as to how the Core 2 Duo's performance has changed over the last few months, leading up to today's official launch of the new quad-core Core 2 Extreme QX6700 processor.
Beyond these various data points there really isn't too much more to know about Kentsfield, other than the fact that the chip is simply a pair of dual-core, Core 2 Duo E6700 processors at 2.66GHz, on a single package substrate, sharing a single 1066MHz QDR (Quad Data Rate) System Bus connection to the Northbridge. That's four cores and 8MB of on-die cache all in a single LGA775 socket. In other words, double the cores, double the cache and double the fun, but of course that also takes the QX6700's TDP (Thermal Design Power) up a notch from the Core 2 Duo's 65-75 Watt range to roughly double to 130 Watts. This puts this new Intel quad-core CPU within range of Intel's legacy Pentium Extreme Edition 965 dual-core chip from a power and heat dissipation perspective. Not too shabby at all, when you consider how much faster the architecture has proven itself to be over the past few months and the fact that we're talking four cores in total here, versus only two in the Pentium EE 965.
|Quad Core Roadmap and Kentsfield Vital Signs|
Before we delve into Intel's new Core 2 Extreme QX6700 in detail, let's step you through Intel's current roadmap product offerings and provide the proper perspective from both a model branding and process technology standpoint.
Intel's roadmap clearly shows that the Pentium D will be completely phased out sometime in the first half of 2007, in favor of newer Core 2 Duo product offerings. In addition we see that other Core 2 Quad models will be introduced to backfill behind the Core 2 Extreme QX line-up. Quad-core targeted for the mainstream seems to be in our not so distant future. In fact Intel has confirmed that a Core 2 Quad Q6600 (2.4GHz with locked core multiplier) will be launched in January '07. Finally, in the 2008 time frame, Intel plans to have their 45nm nodes online and in volume. Octal cores by chance? As always, we dare to dream...
The Core 2 Extreme QX6700 looks like any other Intel processor in a Land Grid Array (LGA) 775 package. There is virtually no difference in its design, save for a few different termination resistors on the underside of its body area, in the center power and ground grid.
Quick readings with CPU-Z and our motherboard BIOS report that we have a Core 2 Quad CPU here that supports SSE3 and EM64T instructions, comes with a stock multiplier of 10X with a 1066MHz FSB and 8MB of on-chip cache, 4MB per dual core die, specifically. Additionally, as you can see in the BIOS screen capture, the Core 2 Extreme series of processors has unlocked multipliers. Our chip actually reported a ratio range from a max of 20X to a min of 6X.
|Power and Heat Characteristics|
There's no question, as the number of CPU cores scales in these new multi-core processor architectures, thermal challenges scale right along with them. We're sure many of you are interested in knowing how Kentsfield runs under pressure and what sort of cooling solutions need to be provisioned for optimal performance. We setup our test-bed with a Zalman CPNS9500 all copper CPU cooler. This cooler is easily our favorite for high load stress testing in the labs as it is both hugely efficient, effective, and very quiet.
In short, the new Core 2 Extreme QX6700 ran like a champ with this setup on our open air test bench.
Here we've giving you a view of how the Core 2 Extreme QX6700 performs while idle and also under full load with 100% CPU utilization. As you can see, the QX6700 has very reasonable operational temperatures with the standard off-the-shelf HSF we chose. Granted, the Zalman CPNS9500 is one of the more top-shelf coolers on the market today but we'd dare say that if you're the type to run a processor like the Core 2 Extreme QX6700, you're not going to skimp on something as basic as a CPU cooler. On a side note, the reference cooler that Intel supplied us with for launch was, quite frankly, completely unacceptable. It was ridiculously loud, even though we manually dialed down its fan speed to 60%. And it didn't do nearly as good a job cooling Kentsfield as our Zalman cooler. We hope Intel finds a better cooling solution for their retail CPU bundles but those of you that like to go with bulk "OEM" chips, do yourself a favor and get a solid cooler like the Zalman CPNS9500.
Regardless, within this test setup environment, the Core 2 Extreme QX6700 idles at a mild 31oC. Under full 100% load with a 30 minute burn-in time, the new quad-core CPU only topped out at only 46-47oC. Talk about a huge contrast to Intel's Pentium D thermal characteristics from days gone by. A quad-core Pentium D 965 would have been a toaster of a chip but the Core 2 Extreme QX6700 is a relatively cool operator. Next up - power consumption.
Please keep in mind that we were testing total system power consumption here at the outlet, not just the power being drawn by the processors alone. Also, we should note that we tested all systems with processor clock throttling technologies like AMD Cool-n-Quiet and Intel Speed-Step enabled. Instead this will give you a measure of power consumption of each system at its stock, rated clock speed at idle on the desktop and under load.
The Core 2 Extreme QX6700-based system, built on an Asus P5W DH motherboard with Intel's own 975X chipset actually draws slightly less power at idle versus an Athlon 64 FX-62 on an nForce 590 SLI based board (Asus M2N32 SLI Deluxe). At full load, the QX6700-based system only draws about 8 Watts more than the FX-62/nForce 590 system. On the other hand, ATI's RD580 chipset definitely helps the power consumption profile of the FX-62 obviously, and it puts the dual-core Athlon 64 system more in line with Core 2 Duo. Regardless, again Intel's quad-core Core 2 Extreme QX6700 makes a very respectable showing in terms of overall system power consumption, relatively speaking, versus its dual-core counterparts.
|Overclocking the QX6700|
Though we typically like to tease you with this type of testing, saving it for the tail end of our performance analysis section, we thought we'd dive right into the good stuff here and show you just how many extra clock cycles Intel has in the hopper backing up their new quad-core processor.
For the purposes of this test, we simply cranked up the FSB speed of Kentsfield rather than altering the CPU multiplier. This will be a more realistic scenario, if some of you are interested in a standard Core 2 Quad chip, rather than a likely much more expensive Extreme edition like the QX6700 we tested.
SiSoft SANDRA 2007 SP1 - Kentsfield CPU, Multimedia and Memory Performance
With the help of our standard Zalman air cooler, and a small voltage bump to 1.4V, we were able run the Core 2 Extreme QX6700 at 3.3GHz for hours at 100% load. SANDRA's reported ALU and SSE3 performance for Core 2 Extreme QX6700 overclocked at this speed is, in a word, insane. At over two times the performance of a Core 2 Extreme X6800, it's almost hard to believe. We'll see if these large overclocked gains actually bear fruit in some real world application testing, in the pages ahead as well.
|Our Test Systems and SANDRA 2007|
How we configured our test systems: When configuring our test systems for the following set of benchmarks, we first entered their respective system BIOSes and set each board to its "Optimized" or "High-Performance Defaults." We then saved the settings, re-entered the BIOS and set memory timings for DDR2-800 at 4,4,4,12 1T latency. The hard drives were then formatted, and Windows XP Professional (SP2) was installed. When the Windows installation was complete, we installed the drivers necessary for our components, and removed Windows Messenger from the system. Auto-Updating and System Restore were then disabled, and we set up a 768MB permanent page file on the same partition as the Windows installation. Lastly, we set Windows XP's Visual Effects to "best performance," installed all of our benchmarking software, defragged the hard drives, and ran all of the tests.
We began our testing with SiSoftware's SANDRA, the System ANalyzer, Diagnostic and Reporting Assistant. We ran five of the built-in subsystem tests that partially comprise the SANDRA 2007 suite (CPU, Multimedia, Memory, Cache, and Latency) with the new Core 2 Extreme QX67700. All of the scores reported below were taken with the processor running at its default clock speeds of 2.66GHz.
In terms of the more standard SANDRA scores like the CPU, Multimedia and Memory Bandwidth tests, the Core 2 Extreme QX6700 scales roughly where we expected it to, that is to say that the QX6700 clocks in at slightly more than two times the performance of a Core 2 Duo E6700 at the same clock speed. The Memory Latency numbers, however, look a bit more interesting perhaps, with Kentsfield showing better latency characteristics due to its combined cache area size of 8MB, once the benchmark begin to scale into the larger test range sizes. Peak memory bandwidth shows the QX6700 at roughly two times the E6700 as well. So far the synthetics are painting Kentsfield in a very flattering hue.
|PCMark05: CPU and Memory|
For our next round of synthetic benchmarks, we ran the CPU and memory performance modules built into Futuremark's PCMark05 suite.
"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.
Intel's Core 2 Extreme QX6700 shows itself to be only 25% faster than the Core 2 Duo clock for clock in this test, as we observe comparing E6700 scores to the QX6700. Of course AMD's fastest dual-core processor takes a beating here by all Intel Core 2 CPUs but this isn't anything new, since the Core 2 Duo launch.
"The Memory test suite is a collection of tests that isolate the performance of the memory subsystem. The memory subsystem consists of various devices on the PC. This includes the main memory, the CPU internal cache (known as the L1 cache) and the external cache (known as the L2 cache). As it is difficult to find applications that only stress the memory, we explicitly developed a set of tests geared for this purpose. The tests are written in C++ and assembly. They include: Reading data blocks from memory, Writing data blocks to memory performing copy operations on data blocks, random access to data items and latency testing." - Courtesy FutureMark Corp.
There aren't any mysteries revealed here in Futuremark's PCMark05 Memory performance module. The Core 2 Extreme QX6700 is right on par with an identically clocked Core 2 Duo chip. The variance in the scores is minor and within the test's margin of error.
|Worldbench 5: Office XP and Photoshop 7|
PC World Magazine's WorldBench 5.0 is a Business and Professional application benchmark. The tests consist of a number of performance modules that each utilize one, or a group of popular applications to gauge performance.
Below we have the results from WB 5's Office XP SP2 and Photoshop 7 modules, recorded in seconds. Lower times indicate better performance here, so the shorter the bar the better.
The results we see here with the Core 2 Extreme QX6800 are more indicative of the fact that Microsoft Office XP and Photoshop 7 are older applications that are getting a bit long in the tooth, since they aren't heavily multithreaded. In fact the Office XP test is more of a multi-tasking work load while there are a few filters in Photoshop that are actually multithreaded. Moving forward, we plan to incorporate Adobe's CS2 Photoshop versions in our benchmark test suite, which is a more current version and more capable of taking advantage of multi-processor systems. The long and short of our tests with these two applications, however, is that Kentstfield is only marginally faster than Core 2 Duo for legacy Office-type and Image Editing applications. We'll explore more multithread capable applications next.
|LAME MT and Sony Vegas|
In our custom LAME MT MP3 encoding test, we convert a large WAV file to the MP3 format, which is a very popular scenario that many end users work with on a day-to-day basis to provide portability and storage of their digital audio content.
In this test, we created our own 223MB WAV file (a never-ending Grateful Dead jam) and converted it to the MP3 format using the multi-thread capable LAME MT application in single and multi-thread modes. Processing times are recorded below. Once again, shorter times equate to better performance.
There are many current audio conversion and editing applications that rely on the Lame MP3 encoding engine. Unfortunately, at this time, LAME only supports up to two threads and our Core 2 Extreme QX6700 performs exactly like a E6700 chip. Incidentally, we tried the same exact file conversion with Apple iTunes and came up with similar results. No quad-core optimizations here yet but we're certain it won't take long since this type of application is a natural for multithreading.
Sony's Vegas DV editing software on the other hand is heavily multithreaded as it processes and mixes both audio and video streams. This is a new breed of digital video editing software that takes full advantage of current multi-core architectures.
In the above screen shots, we can see that both the Athlon 64 FX-62, as well as our Core 2 Extreme QX6700 chip, are fully loaded 100% on each of their respective cores. Hence our performance testing shows very clearly the advantages of the additional processing resources available with Intel's new quad-core CPU.
We tossed in our overclocked scores here as well, just to see how things scale but this graph, in general, is very telling. Versus the Core 2 E6700 at the same clock speed, we get a 42% performance advantage with the QX6700 and roughly a 37% advantage over the dual-core Core 2 Extreme X6800 with its 270MHz clock speed advantage. This is not a two-fold increase as you might expect, which is perhaps the result of Kentsfield's shared front side bus architecture. Regardless, no matter how you slice it, the Core 2 Extreme QX6700 is the fastest processor by a wide margin in this test and might we offer, easily the fastest Desktop processor on the planet currently as of this article's launch, at least it is according to the Sony Vegas software.
For this next batch of tests, we ran Kribibench v1.1, a 3D rendering benchmark produced by the folks at Adept Development. Kribibench is an SSE aware software renderer where a 3D model is rendered and animated by the host CPU, and the average frame rate is reported.
We used two of the included models with this benchmark: a "Sponge Explode" model consisting of over 19.2 million polygons and the test suite's "Ultra" model that is comprised of over 16 billion polys.
With Kribibench's software multithreaded, SSE optimized 3D rendering engine, we see a 45 - 75% performance advantage over a similarly clocked Core 2 Duo processor and up to two times the processing power over the AMD Athlon 64 FX-62 dual-core processor. Kentsfield again makes a very strong showing when the software being driven by it is able to take advantage of all of its cores.
POV-Ray, or the Persistence of Vision Raytracer, is a top-notch 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.
This latest version of the POV-Ray is capable of completely taxing all processor cores in a system in a full-up SMP usage model. As you can see in the screen shot above the graph, the Core 2 Extreme QX6700 is at full 100% resource utilization in this test. Furthermore, Intel's Core 2 architecture in general has a significant lead over the Athlon 64 X2/FX architecture here. Even a 2.66GHz Core 2 Duo processor is some 35% faster than a 2.8GHz Athlon FX-62 in this benchmark. Beyond that, it's a complete run-away for the Core 2 Extreme QX6700 which actually more than doubles the performance of the Core 2 Duo E6700, and comes up just short of doubling the performance of the Core 2 Extreme X6800. Pleasantly surprising, to say the least.
|Cinebench 9.5 and 3DMark06: CPU|
The Cinebench 2003 benchmark is an OpenGL 3D rendering performance test, based on the commercially available Cinema 4D application. Cinema 4D from Maxon is a 3D Rendering and Animation tools suite used by many 3D Animation houses and producers like Sony Animation and many others. And of course it's very demanding of system processor resources.
This is a multi-threaded, multi-processor aware benchmark that renders a single 3D scene and tracks the length of the entire process. The time it took each test system to render the entire scene is represented in the graph below (listed in seconds).
In single-threaded mode the Core 2 Extreme QX6700 puts up a time identical to that of a Core 2 Duo E6700 but engage the multi-CPU engine of Cinebench and Intel's new quad-core takes a commanding lead. The QX6700 completes the workload in nearly half the time of a similarly clocked Core 2 Duo or Athlon 64 FX-62 processor. This looks to be something on the order of about a 73% performance gain, clock for clock between Core 2 Quad and Core 2 Duo architectures.
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 that 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.
70% faster than a Core 2 Duo E6700 and 55% faster than a Core 2 Extreme X6800 dual-core CPU, Intel's Core 2 Extreme QX6700 quad-core puts up yet another thrashing in 3DMark06. We're definitely looking at a pattern here now though. Clock tick for clock tick, this new Intel quad-core processor is on the order of about 70-75% faster in many applications that can take full advantage of all four of its cores in multithread mode. It's not quite double the performance (or 100% faster) as having double the available cores might suggest but it is obviously significantly faster in a major-league sort of way. The only exception to this we've seen so far are our POV-Ray benchmark scores.
|Quake 4: Low and High Resolutions|
For our next set of tests, we benchmarked all of the test systems using a custom single-player Quake 4 timedemo. Here, we installed the game's official v1.3 point release which is SMP capable, turned the resolution down to 640x480, and configured the game to run at its "Low-Quality" graphics setting. Although Quake 4 typically taxes today's high-end GPUs, when it's configured at these minimal settings, it is much more CPU and memory bandwidth-bound than anything else.
Since the engine is getting a bit older now and since it clearly wasn't built multithreaded from the ground up, Quake shows very little advantage for the new Intel quad-core CPU. Kentsfield does weigh in a a couple of frames per second faster than the Core 2 Duo E6700, which runs at the same clock speed but the variance is not much to take note of.
When configured at high resolution game settings and thus relying more on the GPU subsystem, QX6700 Kentsfield performance is even more diminished. In fact Kentsfield brings up the rear amongst the Intel processors lineup. Again, legacy code-bases the likes of which run older game engines like Quake 4 are not where you'll find any appreciable advantage with Intel's new quad-core processor. Forward looking engines that are are more targeted toward a multithreaded architecture are where we'll begin to see significantly more tangible benefits.
|F.E.A.R. v1.08: Low and High Resolutions|
For our last set of game tests, we moved on to more low-resolution benchmarking with F.E.A.R. When testing all processors with F.E.A.R, we drop the resolution to 640x480, and drop all of the in-game graphical options to their minimum values to isolate CPU and memory performance as much as possible. However, the in-game "effects" and "advanced computer options" settings, which control the level of detail for F.E.A.R.'s physics engine and particle system, were all set to maximum since these actually do place loads on the CPU rather than GPU.
With GPU limitations almost completely out of the equation, the multi-core processors we tested we left to run as flat out as they possibly could, handling calls from F.E.A.R.'s various other workloads like physics and particle effects and manipulation. Core 2 Extreme QX6700 shows a sizeable almost 60 fps gain over the similarly clocked Core 2 Duo E6700 and even bests Core 2 Extreme X6800's dual-core performance by a solid 34 frames per second.
At high resolution, where our GeForce 7950 GX2 is really the major limiting factor, all of Kentsfield's additional performance is lost. For now, multithreaded game engines that truly exploit new multi-core processor architectures are still only on the horizon, but hopefully that horizon is not too far off. More on this topic in our final analysis and wrap-up, next.
|Our Summary and Conclusion|
General Benchmark And Performance Summary:
Conversely, we were extremely impressed with a few of our more demanding professional rendering applications like POV-Ray, where the Core 2 Extreme QX6700 offered incredibly strong performance that scaled up to 2X the performance of Intel's own Core 2 Duo line and was able to offer these gains in a single-chip footprint. Still, in other situations, like Sony Vegas, Cinebench and 3DMark06, we saw the QX6700 offer huge performance gains again, but also fall short of doubling Core 2 Duo results, with about a 70-75% gain overall.
"Multi-core computing is the new standard for PC games, and we at Epic are thrilled to see Intel leading the industry forward with Core 2 Extreme. Its four high-performance CPU cores enable a new level of realism in games, with realistic physics simulation, character animation, and other computationally-intensive systems."
"The introduction of Quad Core processor based PC's allows Remedy to create real next generation games as demonstrated by our "Alan Wake", a psychological action thriller to be published by Microsoft Game Studios. Dividing complex programming tasks into multiple threads is the way to exploit performance that allows us to create more realistic and dynamically generated environments - and thus enjoy fantastic game worlds like never before."
"Gas Powered Games has had the opportunity with our Supreme Commander title to be on the leading edge of working with Intel's Quad Core. The performance and experience enhancements we've seen implementing our multi-threaded architecture on Intel's multi-core systems has shown us that these technologies will represent the minimum bar for the future of advanced gaming."
"Gamers are asking for more and more realism and immersion, and this of course requires even more power and speed. To meet their expectations, we optimized our more demanding games for multithreading. In this way, Ubisoft can take full advantage of the architecture of multi core processors such as Intel Core 2 Duo and Intel Core 2 Quad Extreme"
"Quad-core will change every aspect of PC gaming. It will change how we create our games, how we provision our service, and how we design our games. The scalability we've seen in graphics over the last few years will now extend to physics, AI, animation, and all the systems which are critical to moving beyond the era of pretty but dumb games."
So with the game industry fully behind quad-core and multi-core architectures (don't forget consoles have all gone multi-core as well), and with the obvious advantages a quad-core system has in virtually all future multithread capable applications, its clear that 2007 will line-up to be the year of the quad-core. Or perhaps we should generalize a bit and say simply the year of the "multi-core" processor. Intel's Kentsfield Core 2 Extreme QX6700 launch today marks a major milestone in computer processor architecture history. The company is the first to market with a single chip integrated quad-core X86 processor and as expected, it more than lit up a few of the benchmarks and applications that could take advantage of it.
AMD's 4X4 dual-socket architecture is waiting in the wings as we're sure many of you are aware. Though the solution isn't going to be as neat or elegant as Intel's single-socket, single-chip product offering, the big question is how will 4X4 scale versus the Core 2 Quad? Will we see the benefits of AMD's HT serial direct connect architecture offer an edge over the aging Intel QDR bus? Or will Intel's obvious higher IPC core CPU advantages be too much for a pair of Athlon 64 X2 processors to overcome? We'll try to provide better clarity for you and answer these questions directly in the weeks ahead.
Keep in mind also however that, Intel has historically gone from an MCM (multi-chip module) design like we've seen here with Kentsfield, to a full fledge monolithic architecture with shared caches. We've seen this with their Presler core Pentium D architecture, which then migrated to the Core 2 Duo's monolithic approach. Intel is likely to take this path again with their 45nm process-based product roadmap, including things like SSE4 and a single large cache pool. So, as they say, there is always another train coming it seems.
Though it's going to carry a hefty price tag at $999 currently, availability is almost immediate (11/6) for those early adopters who dare to step up to Intel's Core 2 Extreme QX6700. If you can contain yourselves long enough (we know many of you, like us frankly might not), mainstream Core 2 Quad Q6600 processors and perhaps other speed bins will ship around January of '07. For now, Intel keeps bringing on the hurt for AMD, turning up the pressure on performance significantly, and introducing to the world the first quad-core X86 processor for the desktop. With this processor launch, Intel's Quad Core assault feels a bit more like a mugging. Not that there's anything wrong with that.