Date: March 10, 2010
Author(s): Rob Williams
It’s official. We’re now entering the six-core realm, thanks to Intel’s Gulftown. The first model, Core i7-980X, is more than capable of delivering the sick scores that our title suggests, and along with it, we can begin to see some major benefits of the 32nm process. To sweeten the deal further, Intel even includes an effective new CPU cooler.
In May of 2005, both AMD and Intel delivered the world’s first desktop dual-core CPU’s, and since then, the processor technology and increase in performance from generation to generation has accelerated to mind-boggling levels. I’m not sure if it’s a sickness, but when I look back to the landscape just two years ago, I can’t help but feel a bit underwhelmed, because the newer models are so much more attractive.
It’s true though. The acceleration of technology is incredible, and it’s actually a bit hard to believe that just five years ago, we were all sporting single-core processors in our desktops. The first dual-core’s came in May of 2005 as mentioned, and a mere year-and-a-half later, Intel launched the first desktop quad-core offering, the QX6700.
So if it only took a year-and-a-half to make the shift from a dual-core to a quad-core, how on earth has it taken another 4.3 years to finally see the industry’s first six-core offering? The reason isn’t due to the fact that it couldn’t be done (I remember Intel talking about Octal-Cores at IDF 2007, and we still don’t see those on the desktop side), but rather because they haven’t been needed.
Just over two years ago, we took Intel’s Skulltrail platform for a spin. For those who may not recall, Skulltrail was Intel’s ultra-high-end solution for those who wanted the best in multi-tasking and the best in overall raw horsepower. As we learned in that article, the company certainly delivered. One thing we also learned, though, was that there was a dire need for better multi-threading support in our common applications, since we found it hard to actually push all of the eight cores we were given.
The multi-threaded situation is much improved as it stands today, and it’s constantly getting better. As we’ll see in our test results, even our real-world benchmarks that we’ve been using for the past year-and-a-half see good benefits with today’s multi-core processors, including the one we’ll be taking a look at today.
Allow me to introduce Intel’s first six-core desktop processor, the Core i7-980X Extreme Edition:
You might be saying to yourself, “A six-core? I can’t even push a quad-core!”, and to be honest, that’s what makes Gulftown so interesting and fun. I admit that didn’t expect much when I first sat down to benchmark the i7-980X, because I already had a good mindset of what I could expect to see. After all, the processor is quite similar to the i7-975, but with +50% the number of cores and threads. But, there’s just something about it… something about having so much power at your finger-tips that cause you to let out an evil laugh. Wait, is that just me?
It’s obvious that the i7-980X has a lot of muscle, but there’s more that makes the chip special than just the core/thread boost. For one, like the recently-released Clarkdale desktop processors, Gulftown (based on Westmere) is built upon a 32nm process. One thing that means less heat and less power consumption than previous architectures, and for proof, you don’t have to look further than our power consumption page.
In addition, there’s also the addition of the AES-NI instruction set, increased L3 Cache (12MB) and also the possibility of improved overclocking.
Before continuing, it’s worth taking a look at Intel’s current Core i line-up:
|Core i7-980X (4)|
|Core i7-975 (1)|
|Core i7-960 (1)|
|Core i7-950 (1)|
|Core i7-870 (2)|
|Core i7-860S (2)|
|Core i7-930 (1)|
|Core i7-860 (2)|
|Core i7-920 (1)|
|Core i5-670 (3)|
|Core i5-750S (2)|
|Core i5-750 (2)|
|Core i5-661 (3)|
|Core i5-660 (3)|
|Core i5-650 (3)|
|Core i3-540 (3)|
|Core i3-530 (3)|
|Pentium G6950 (3)|
Microarchitecture: (1) Bloomfield, (2) Lynnfield, (3) Clarkdale, (4) Gulftown
When I first learned of the Core i7-980X’s release, I was a bit worried of what the price could be. After all, it’s not unlike Intel to charge a sweet premium on its highest-end and special products. Just take a look at Skulltrail, where each CPU was 200MHz faster and $500 more expensive than the high-end QX9650. But, my skepticism was all for naught, and I couldn’t be happier about it.
Like all other top-end launches from Intel in recent memory, the Core i7-980X will give the i7-975 the boot and sell for $999 in sales of 1,000 (the retail price will be a bit more than this). Think about that for a second. Intel just replaced a quad-core with similar architecture with a six-core built on an improved process, is including a robust CPU cooler in the box, and then is selling it for the same price. This might be the first time when I’ve actually considered a $999 processor to be a deal.
Intel hasn’t officially admitted the fate of the i7-975, but it can be assumed that it’s simply going to be removed from the line-up in the near-future. As model names go, the company has pretty much painted itself into a corner, so unless it wants to rename it to the i7-967.5, we’re just going to see it go away. The i7-970 moniker is currently reserved for the next Gulftown variant, which will appear this fall.
That leads me to the next major point about Gulftown. When the i7-980X is officially released to market in the coming weeks, there will be a wide spread before the next follow-up model is released. As mentioned, that will be the i7-970, and we’ll see it this fall. Past that, I’m not quite sure when the next six-core model will hit. One would assume that Intel may release a lesser-expensive model early next year, but it’s really too early to speculate.
For those who are purchasing a Gulftown, the long stretch in the launch schedule is actually a good thing, because for the first time in a while, you can purchase one and have confidence that even later this year, your i7-980X will still rightfully be at the top of Intel’s food chain. It’s all made even more interesting because it’s also the first time in a while where a new processor model can’t simply be matched by an overclocker with a lesser model. Theoretically, for a quad-core to match the multi-threaded performance of this Gulftown at stock speeds, it would have to be overclocked to 5.0GHz. And that’s not even taking the other microarchitecture enhancements into consideration.
Yes, I think Gulftown owners will be pretty pleased about that.
Before we dive in too deep, I wanted to clarify a minor detail. You might have noticed that up to this point, I’ve been calling the i7-980X a “six-core” processor, rather than something like a Sex-Core or Hex-Core. I queried Intel as to the reason for this, and we were told that it’s not even calling Gulftown a “six-core” processor. In fact, it seems to be avoiding the core talk altogether, and it might not be a bad idea given the regular consumer probably isn’t going to care about the number of cores, but would rather just know how kick-ass one model is compared to the next.
As far as AMD is concerned, I’m not quite sure if it shares the same idea, but at least there, it takes no effort to understand the number of cores in its processors, thanks to the simple naming scheme (X2, X3, X4, X6, X12). I might have made the last one up.
So that said, what makes Gulftown what it is, and what’s its positioning? Those questions will be tackled here, beginning with a look into the actual Gulftown die. As you can see in the image below, the layout of the chip isn’t much different from Bloomfield’s. The memory controller still lines the top, and the miscellaneous I/O and QPI accesses are found on each side. Likewise, the L3 cache is placed directly underneath each core (2MB each), while the Queue and Uncore neatly divide up the two sections.
Like Bloomfield, Gulftown is a native six-core processor, not simply two tri-cores fused together. AMD’s upcoming Phenom X6’s will also have a native design, so hopefully we are completely done with the hacked designs we’ve seen in the past, such as with Intel’s own Core 2 Quad line-up.
As a quick recap to Intel’s method of doing things, Gulftown is based on the Westmere architecture, which was first seen with the Clarkdale and Arrandale chips earlier this year. All Westmere chips are based on the 32nm process, and are officially part of Intel’s “Tick” phase. Tock will came later this year with Sandy Bridge, which will also be based on a 32nm process, but will feature a new microarchitecture.
The target audience of Gulftown, or at least the Core i7-980X in particular, is the same audience that all Extreme Edition’s target. It’s for those who want nothing but the best… a chip that delivers unparalleled multi-tasking, incredible performance in multi-threaded applications, and of course, overall flexibility. In previous discussions, I compared moving to a six-core from a quad-core being similar to moving from a large hallway to a living room. The hallway may have allowed a fair amount to get done, but the living room is obviously going to provide you with more breathing-room.
And believe it or not, as we’ll see in some of our tests later, there are many applications that are able to take advantage of a six-core processor. As our current test suite was finalized in late 2008 (we’re in the process of revising the entire thing), not all of our tests are as multi-threaded as others, but even with what we do have, the differences are easily seen.
Another major benefit of Gulftown, as if we needed another one, is that it’s backwards-compatible with the X58 chipset. So if you happened to have purchased an X58 board at launch for the Core i7-965 or any other Bloomfield-based chip, you could go and grab a BIOS update to open up Gulftown support. The support is of course based on a vendor basis, though, so I do recommend double-checking to make sure that your particular board does indeed have a BIOS for Gulftown support.
One of the more interesting features that launched with Bloomfield in late 2008 was the “Turbo” feature. This of course is what allows the processor to boost its clock just a wee bit when it’s being stressed, for improved performance. The method in which it operates can be seen below, but there’s something a bit interesting that caught me off-guard. Can you see it?
With the Lynnfield launch this past fall, and likewise with Clarkdale’s launch this past January, we saw what was essentially a boost to Turbo, where the Turbo on these CPU’s are much more effective than the Turbo on Bloomfield. As we’re dealing with the Westmere microarchitecture here, I half-expected the same kind of robust Turbo performance, but as you can see in the chart above, the height of Turbo will be two extra steps on one core. Compare that to five steps on a Lynnfield chip.
So what’s the reason for the throttling down? My theory is that Intel doesn’t want to exceed the 3.60GHz mark on any of its current CPU’s, and to back me up, just take a look at our table on the previous page. There’s only one CPU that has a Turbo to exceed 3.60GHz, and that’s the Core i5-670, which can go to 3.73GHz. My reasoning there is that it’s a lot safer to hit a higher clock on a dual-core than it is a quad-core, much less a six-core. Intel seems to be playing it safe, and I can’t fault it for that.
Before I wrap up here, there’s just one more thing Gulftown has brought to the table worth mentioning. This if course has been rumored for a while, and it’s absolutely true. Intel has upped the ante and delivered not just a standard CPU cooler we’re used to seeing, but rather a custom-designed enthusiast’s cooler designed for overclocking and overall excellent performance.
It’s not the prettiest cooler out there, but it’s not meant to be. Rather, it’s meant to be effective and if desired, quiet. It’s important to note that this is an Intel-designed cooler, not one the company simply ordered en masse and re-branded. At this time, it’s only going to be included with the Core i7-980X processor and not sold separately. The company could always change its mind based on overall feedback though.
For vendors that sell custom-built PC’s and pair them up with the i7-980X, they have the option to change the branding at the top. I assume the “Intel” logo on the fan, and the general design couldn’t be altered. But it’s definitely a decent option for those companies who want to include the cooler, since it almost comes paired in with the price of the processor alone.
Aside from all that, it’s actually quite an effective cooler, and with it, our Gulftown sample ran cooler than our Core i7-975 using a Thermalright Ultra-120. Intel definitely put some legwork into this one, and if you’re going to be purchasing the i7-980X, you really don’t need to fuss over the CPU cooler, this one is just fine.
Today, most people are completely satisfied when using a dual-core processor, and for those who want a bit more oomph, there are countless quad-core models out there to choose from. But as it stands, many people don’t believe that even a quad-core is worth the money, because they figure that no application or scenario they deal with could even possibly take advantage of all the cores.
I’m willing to bet that that’s not really the case, because even some of our simpler apps we use for testing, performance is better when we increase the cores. The benefits of an increased number of cores comes down to two things… improved multi-tasking, and improved performance where heavily multi-threaded applications are concerned. The bigger and better your CPU, the less you have to worry about processes hanging.
Comparing the Core i7-980X to the quad-core i7-975, Intel touts that performance is increased upwards of 50%. For video effects, we could expect about a 34% increase in overall performance, and then 37% for 3D rendering. For game physics, we’ll see the near 50% increase, as those algorithms tend to be very scalable with the number of available cores.
Bringing up our Skulltrail article once gain, one comment I had at the time was that despite 8 physical cores being available in the PC, it was hard to find any process outside of a synthetic benchmark that could take advantage of all 8. At best, we’d see 6 or 7 cores being used, but not 8. The same can be said here, and Intel admits it. Even though we’re literally seeing a CPU with 50% more cores, we’re not going to see a 50% increase in performance in all areas. For video and 3D, it’s more like a ~35% increase. That’s still rather impressive, and does do well to set itself apart from any quad-core model.
I mentioned before that multi-threaded support in applications and games is constantly improving, and its proven with a look at Intel’s most recent version of a list it develops that gives us a one-stop shop to see which applications and games we should test with. If you are interested in looking at the full list, we have it hosted here.
It its press deck for Gulftown, Intel specifically mentions a couple of applications and games that are either out now, or soon to be out, that take good advantage of multi-core processors. The first mention is of Vegas Pro 9.0c. There, the Vice President of Technology for Sony Creative Software said, “As a result, Vegas Pro 9.0c automatically takes advantage of all 12 CPU threads on the Intel Core i7-980X, resulting in faster rendering of video projects.“.
A video creation tool that can use all 12 threads? Sounds good to me. Unfortunately, we didn’t have time to test out Vegas Pro 9, namely because we didn’t have a project to use with it, but it’s something we’d like to work on given the fact that it’s truly multi-threaded. Aside from that, there’s also Maxon’s Cinebench R11.5, which is the most recent update to one of our favorite benchmarking tools.
The latest version is a vast improvement over the previous ones, as it brings global illumination and ray tracing into things (which almost assuredly will put Intel ahead of AMD in any comparison, given that Intel’s ray tracing performance is currently very tough to beat). The scoring system has also been vastly improved, and we’re back to dealing with simpler numbers. No more of these 1,000’s, but rather we’re down to dealing with single or double digits. The one benefit is that in our quick testing with the software so far, the results are highly repeatable, which makes benchmarking with it all the more enjoyable.
For games, Intel focuses on two real-time strategy titles, R.U.S.E. and Napoleon: Total War. The latter of the two is available right now, and I had hoped to test a bit with it in time for this article, but was unable. The biggest issue is that the style of game is certainly not one to my liking, so I’m more content with waiting for a game to come along that I don’t have to study up on in order to play.
But as you can see the foil below, the extra cores on a CPU can actually increase the animation detail. It’s difficult to tell the difference from two shots that already vary from one another, but the benefits explained here are going to be similar for all games that can take advantage of a CPU. With the extra cores, the physics can be offloaded, which increases performance and results in a better overall look. We do buy big GPU’s for the best graphics, so this is the way the CPU can be used to improve games in other ways.
R.U.S.E. boasts some of the same benefits, but one that caught my eye was the fact that non-important features, such as horses and cows on the landscape, are animated, thanks to the CPU. Also, if you zoom far out of a scene, rather than have a flat texture applied to an object, the extra CPU cores will allow it to remain a real model, albeit a low-res version of it (which is all that’s needed, given the detail wouldn’t be seen anyway).
The overall thought to take away from this page is that the multi-threading landscape is continually improving, and more and more applications and even games are becoming multi-core aware. I still think we have a long way to go before truly multi-threaded games become commonplace, but we’re on the right track.
At Techgage, we strive to make sure our results are as accurate as possible. Our testing is rigorous and time-consuming, but we feel the effort is worth it. In an attempt to leave no question unanswered, this page contains not only our testbed specifications, but also a fully-detailed look at how we conduct our testing.
If there is a bit of information that we’ve omitted, or you wish to offer thoughts or suggest changes, please feel free to shoot us an e-mail or post in our forums.
The table below lists the hardware for our two current machines, which remains unchanged throughout all testing, with the exception of the processor. Each CPU used for the sake of comparison is also listed here, along with the BIOS version of the motherboard used. In addition, each one of the URLs in this table can be clicked to view the respective review of that product, or if a review doesn’t exist, you will be led to the product on the manufacturer’s website.
Please note that for benchmarking the Core i7-980X, we deviated from the Rampage II Extreme that we tested all other Bloomfield’s with in favor of taking Gigabyte’s X58A-UD5 for a spin. This motherboard change won’t effect the performance, but it will change the power consumption just a wee bit. As mentioned before, we’re in the process of upgrading our entire CPU test suite, and the X58A-UD5 will become the base of our new LGA1366 test platform.
AMD AM2+/AM3 Test System
AMD Phenom II X4 965 Black Edition – Quad-Core, 3.40GHz, 1.325v
AMD Phenom II X4 955 Black Edition – Quad-Core, 3.20GHz, 1.325v
AMD Phenom II X3 720 Black Edition – Tri-Core, 2.80GHz, 1.325v
AMD Phenom II X2 555 Black Edition – Dual-Core, 3.20GHz, 1.325v
AMD Athlon II X4 635 – Quad-Core, 2.90GHz, 1.325v
AMD Athlon II X4 620 – Quad-Core, 2.60GHz, 1.375v
AMD Athlon II X3 435 – Tri-Core, 2.90GHz, 1.325v
AMD Athlon II X2 240e – Dual-Core, 2.80GHz, 1.325v
Gigabyte MA790GP-DS4H – 790GX-based, F3 BIOS (01/13/09)
Corsair XMS3 DHX 2x2GB – DDR3-1333 7-7-7-20-2T, 1.65v
Intel LGA1156 Test System
|Processors||Intel Core i7-870 – Quad-Core, 2.93GHz, ~1.25v|
Intel Core i5-750 – Quad-Core, 2.66GHz, ~1.25v
Intel Core i5-661 – Dual-Core, 3.33GHz, ~1.10v
Intel Core i3-530 – Dual-Core, 2.93GHz, ~1.00v
Lynnfield: Gigabyte P55-UD5 – P55-based, F3 BIOS (08/01/09)
Westmere: ASUS P7H55D-M EVO – H55-based, 0503 BIOS (12/02/09)
Corsair XMS3 DHX 2x2GB – DDR3-1333 7-7-7-20-2T, 1.65v
ATI Radeon HD 4870 512MB (Catalyst 8.11)
Intel LGA1366 Test System
ASUS Rampage II Extreme – X58-based, 0705 BIOS (11/21/08)
Gigabyte X58A-UD5 – X58-based, F4 BIOS (02/12/10)
(Gigabyte’s board used only for Core i7-980X)
Intel Core 2 Quad Q9650 – Quad-Core, 3.00GHz, 1.30v (Sim)
Intel Core 2 Quad Q9550 – Quad-Core, 2.83GHz, 1.30v (Sim)
Intel Core 2 Quad Q9400 – Quad-Core, 2.66GHz, 1.30v
Intel Core 2 Quad Q8200 – Quad-Core, 2.33GHz, 1.30v
Intel Core 2 Duo E8600 – Dual-Core, 3.33GHz, 1.30v
Intel Core 2 Duo E8500 – Dual-Core, 3.16GHz, 1.30v (Sim)
Intel Core 2 Duo E8400 – Dual-Core, 3.00GHz, 1.30v
Intel Pentium Dual-Core E5200 – Dual-Core 2.50GHz, 1.30v
ASUS Rampage Extreme – X48-based, 0501 BIOS (08/28/08)
(Sim) represents models that were tested using a faster, but under-clocked processor. For example, for the Q9550, we used the QX9770, since the specs are identical all-around, except for the clock speeds. Those were adjusted appropriately, effectively giving us a Q9550 to test with.
When preparing our testbeds for any type of performance testing, we follow these guidelines:
To aide with the goal of keeping accurate and repeatable results, we alter certain services in Windows Vista from starting up at boot. This is due to the fact that these services have the tendency to start up in the background without notice, potentially causing slightly inaccurate results. Disabling “Windows Search” turns off the OS’ indexing which can at times utilize the hard drive and memory more than we’d like.
To help test out the real performance benefits of a given processor, we run a large collection of both real-world and synthetic benchmarks, including 3ds Max, Adobe Lightroom, TMPGEnc Xpress, Sandra 2009 and many more.
Our ultimate goal is always to find out which processor excels in a given scenario and why. Running all of the applications in our carefully-chosen suite can help better give us answers to those questions. Aside from application data, we also run two common games to see how performance scales there, including Call of Duty 4 and Half-Life 2: Episode Two.
In an attempt to offer “real-world” results, we do not utilize timedemos in any of our reviews. Each game in our test suite is benchmarked manually, with the minimum and average frames-per-second (FPS) captured with the help of FRAPS 2.9.5.
To deliver the best overall results, each title we use is exhaustively explored in order to find the best possible level in terms of intensiveness and replayability. Once a level is chosen, we play through repeatedly to find the best possible route and then in our official benchmarking, we stick to that route as close as possible. Since we are not robots and the game can throw in minor twists with each run, no run can be identical to the pixel.
Each game and setting combination is tested twice, and if there is a discrepancy between the initial results, the testing is repeated until we see results we are confident with.
The two games we currently use for our motherboard reviews are listed below, with direct screenshots of the game’s setting screens and explanations of why we chose what we did.
Autodesk’s 3ds Max is without question an industry standard when it comes to 3D modeling and animation, with DreamWorks, BioWare and Blizzard Entertainment being a few of its notable users. It’s a multi-threaded application that’s designed to be right at home on multi-core and multi-processor workstations or render farms, so it easily tasks even the biggest system we can currently throw at it.
For our testing, we use two project files that are designed to last long enough to find any weakness in our setup and also allows us to find a result that’s easily comparable between both motherboards and processors. The first project is a dog model included on recent 3ds Max DVD’s, which we infused with some Techgage flavor.
Our second project is a Bathroom scene that makes heavy use of ray tracing. Like the dog model, this one is also included with the application’s sample files DVD. The dog is rendered at an 1100×825 resolution, while the Bathroom is rendered as 1080p (1920×1080).
I don’t think the result seen here will come as any surprise. 3D rendering applications tend to have the best multi-threaded goals in mind, and that’s evidenced here with very nice gains over the quad-core i7-975 in both of our render jobs.
Like 3DS Max, Cinema 4D is another popular cross-platform 3D graphics application that’s used by new users and experts alike. Its creators, Maxon, are well aware that their users are interested in huge computers to speed up rendering times, which is one reason why they released Cinebench to the public.
Cinebench R10 is based on the Cinema 4D engine and the test consists of rendering a high-resolution model of a motorcycle and gives a score at the end. Like most other 3D applications on the market, Cinebench will take advantage of as many cores as you can throw at it.
I’m willing to bet that the latest version of Cinebench has improved multi-threading capabilities, but even so, the results seen here on the i7-980X are quite good. The scaling is excellent, and we see just about a 40% increase in our multi-threaded test. Naturally, the single-thread test is on par with the Core i7-975.
Similar to Cinebench, the “Persistence of Vision Ray Tracer” is as you’d expect, a ray tracing application that also happens to be cross-platform. It allows you to take your environment and models and apply a ray tracing algorithm, based on a script you either write yourself or borrow from others. It’s a free application and has become a standard in the ray tracing community and some of the results that can be seen are completely mind-blowing.
The official version of POV-Ray is 3.6, but the 3.7 beta unlocks the ability to take full advantage of a multi-core processor, which is why we use it in our testing. Applying ray tracing algorithms can be extremely system intensive, so this is one area where multi-core processors will be of true benefit.
For our test, we run the built-in benchmark, which delivers a simple score (Pixels-Per-Second) the the end. The higher, the better. If one score is twice another, it does literally mean it rendered twice as fast.
Intel’s ray tracing performance truly shines here, with what amounts to more than a 50% increase between the i7-980X and i7-975. It should be noted, though, that part of the reason for this could be the fact that we’ve updated our version of POV-Ray since the i7-975 was benchmarked, so that might be attributed to the more than 50% increase. Due to how POV-Ray is developed, a certain version will cease to operate after a certain date, so an upgrade is mandatory. That’s not a bad thing for those who want to use the tool for actual ray tracing work, but it makes things difficult for benchmarking.
Photo manipulation benchmarks are more relevant than ever, given the proliferation of high-end digital photography hardware. For this benchmark, we test the system’s handling of RAW photo data using Adobe Lightroom, an excellent RAW photo editor and organizer that’s easy to use and looks fantastic.
For our testing, we take 100 RAW files (in Nikon’s .NEF file format) which have a 10-megapixel resolution, and export them as JPEG files in 1000×669 resolution, similar to most of the photos we use here on the website. Such a result could also be easily distributed online or saved as a low-resolution backup. This test involves not only scaling of the image itself, but encoding in a different image format. The test is timed indirectly using a stopwatch, and times are accurate to within +/- 0.25 seconds.
And here we have a great example of an application that is multi-threaded, but not to the extent that we’re going to see major gains on a six-core or higher processor. For our tests, we’re using Lightroom 2, but 3 is right around the corner, so we can hope to see further improvements made there.
When it comes to video transcoding, one of the best offerings on the market is TMPGEnc Xpress. Although a bit pricey, the software offers an incredible amount of flexibility and customization, not to mention superb format support. From the get go, you can output to DivX, DVD, Video-CD, Super Video-CD, HDV, QuickTime, MPEG, and more. It even goes as far as to include support for Blu-ray video!
There are a few reasons why we choose to use TMPGEnc for our tests. The first relates to the reasons laid out above. The sheer ease of use and flexibility is appreciated. Beyond that, the application does us a huge favor by tracking the encoding time, so that we can actually look away while an encode is taking place and not be afraid that we’ll miss the final encoding time. Believe it or not, not all transcoding applications work like this.
For our test, we take a 0.99GB high-quality DivX H.264 AVI video of Half-Life 2: Episode Two gameplay with stereo audio and transcode it to the same resolution of 720p (1280×720), but lower the bit rate in order to attain a modest file size. This test also utilizes the SSE instruction sets, either SSE2 or SSE4, depending on what the chip supports.
I have to admit that I’m quite impressed here, because the six-core is actually put to good work. We didn’t quite reach the 35% faster mark that Intel stated in the press deck, but did come rather close, at 31%. Depending on the video encoder used, I’m confident 35% could be hit. For our future video encoding tests, we’ll be dropping DivX in lieu of today’s more popular HD formats.
While TMPGEnc XPress’ purpose is to convert video formats, ProShow from Photodex helps turn your collection of photos into a fantastic-looking slide show. I can’t call myself a slide show buff, but this tool is unquestionably definitive. It offers many editing abilities and the ability to export in a variety of formats, including a standard video file, DVD video and even HD video.
Like TMPGEnc and many other video encoders, ProShow can take full advantage of a multi-core processor. It doesn’t support SSE4 however, but hopefully will in the future as it would improve encoding times considerably. Still, when a slide show application handles a multi-core processor effectively, it has to make you wonder why there is such a delay in seeing a wider-range of such applications on the marketplace.
With ProShow, the performance gain is even less than what we saw with TMPGEnc, with an average gain of 27%. That’s still much better than a quad-core, but we’re still not diving deep into the sixth core.
This test here stresses the CPU’s ability to handle multi-media instructions and data, using both MMX and SSE2/3/4 as the instruction sets of choice. The results are divided by integer, floating point and double precision, three specific numbering formats used commonly in multi-media work.
Leave it to a synthetic benchmark to show us just what we wanted to see… a 50% increase. We call these synthetic benchmarks for a reason, because in the real-world, as we’ve seen, it’s very hard to exploit all six cores at once, or even come close to an actual 50% increase.
With each new processor launch, one thing that’s bound to prove faster are mathematical equations, which when all said and done, plays a massive role in a lot of our computing today. The faster an equation can be completed, the faster a math-heavy process can finish.
Sandra includes applications designed to specifically test the mathematical performance of processors, with the main one being the arithmetic test.
As expected, like Sandra’s multimedia test, the arithmetic test does well to use every bit of processor you give it, so we once again see a near-perfect 50% performance gain.
Crypto is a major part of computing, whether you know it or not, and certain processes can prove slower than others, depending on their algorithms. User passwords on your home PC are encrypted, as are user passwords on web servers (like in our forums). Past that, crypto is used in other areas as well, such as with creating of unbreakable locks on files or assigning a hash to a particular file (like MD5).
In Sandra’s Cryptography test, the results are outputted as MB/s, higher being better. Although this is somewhat of an odd metric to go by, generally speaking, the higher the number, the faster the CPU tears through the respective algorithm, which comes down to how fast a password is either encrypted, decrypted, signed, et cetera.
As if we needed this graph to be skewed even more! When we published our Clarkdale launch article, it was humorous to see the AES performance of the i5-661 make all of the competition look like wimps, but the i7-980X makes the situation even worse. Even the SHA test saw a dramatic increase, which is surprising given I didn’t even realize it was truly multi-threaded until this point.
It goes without saying… all your crypto are belong to the i7-980X.
Most, if not all, businesses in existence have to crack open a spreadsheet at some point. Though simple in concept, spreadsheets are an ideal way to either track information or compute large calculations all in real-time. This is important when you run a business that deals with a large amount of expenses.
Although the importance of how fast a calculation takes in an Excel file is, we include results here since they heavily test the mathematical capabilities of each processor. Because Excel 2007 is completely multi-threaded (it can even take advantage of an 8-Core Skulltrail), it makes for a great benchmark to show the scaling between all of our CPUs.
I’ll let Intel explain the two files we use:
Monte Carlo – This workload calculates the European Put and Call option valuation for Black-Scholes option pricing using Monte Carlo simulation. It simulates the calculations performed when a spreadsheet with input parameters is updated and must recalculate the option valuation. In this scenario we execute approximately 300,000 iterations of Monte Carlo simulation. In addition, the workload uses Excel lookup functions to compare the put price from the model with the historical market price for 50,000 rows to understand the convergence. The input file is a 70.1 MB spreadsheet.
Calculations – This workload executes approximately 28,000 sets of calculations using the most common calculations and functions found in Excel*. These include common arithmetic operations like addition, subtraction, division, rounding and square root. It also includes common statistical analysis functions such as Max, Min, Median and Average. The calculations are performed after a spreadsheet with a large dataset is updated with new values and must re-calculate many data points. The input file is a 6.2 MB spreadsheet.
Excel might seem innocent enough, but it’s easily one of the most multi-threaded applications on the market. It’s going to be awful difficult for a regular user to ever require more than one core, but for those with extremely robust spreadsheets replete with calculations, multi-core processors can help. When running our macros here, Excel actually said “Using 12 processors” in the corner, so it can almost be assumed that there really is no limit to the number of cores you can use.
Generally speaking, the faster the processor, the higher the system-wide bandwidth and the lower the latency. As is always the case, faster is better when it comes to processors, as we’ll see below. But with Core i7, the game changes up a bit.
Whereas previous memory controllers utilized a dual-channel operation, Intel threw that out the window to introduce triple-channel, which we talked a lot about at August’s IDF. Further, since Intel integrates the IMC onto the die of the new CPUs, benefits are going to be seen all-around.
Before jumping into the results, we already had an idea of what to expect, and just as we did, the results seen are nothing short of staggering.
Here’s something I wasn’t expecting. For some reason, the bandwidth we saw on the i7-980X was much less than the i7-975, despite using the same memory speeds. Given what Intel told me, I’m willing to bet this could be an issue with our motherboard, but I hope to find out for sure soon. Up until just yesterday, our board was unable to handle DDR3-1600 speeds, but a beta BIOS fixed that. So it can be assumed that just maybe, an issue related to memory still exists. So take the results above with a grain of salt for the time-being.
While the memory bandwidth was a little strange, the cache bandwidth is just about where we’d expect it to be. It’s also interesting to note that the latency is a bit lower on the i7-980X, but it’s again tough to know at this point whether that’s due to the motherboard or the architecture.
How fast can one core swap data with another? It might not seem that important, but it definitely is if you are dealing with a true multi-threaded application. The faster data can be swapped around, the faster it’s going to be finished, so overall, inter-core speeds are important in every regard.
Even without looking at the data, we know that Core i7 is going to excel here, for a few different reasons. The main is the fact that this is Intel’s first native Quad-Core. Rather than have two Dual-Core dies placed beside each other, i7 was built to place four cores together, so that in itself improves things. Past that, the ultra-fast QPI bus likely also has something to do with speed increases.
Our core latency hasn’t changed much from the i7-980X, and it even decreased by 1ns. This is understandable given the different layout of the cores compared to the Bloomfield’s. Core bandwidth has seen an obvious boost, though.
While some popular game franchises are struggling to keep themselves healthy, Call of Duty doesn’t have much to worry about. This is Treyarch’s third go at a game in the series, and a first for one that’s featured on the PC. All worries leading up to this title were all for naught, though, as Treyarch delivered on all promises.
To help keep things fresh, CoD: World at War focuses on battles not exhaustively explored in previous WWII-inspired games. These include battles which take place in the Pacific region, Russia and Berlin, and variety is definitely something this game pulls off well, so it’s unlikely you’ll be off your toes until the end of the game.
For our testing, we use a level called “Relentless”, as it’s easily one of the most intensive levels in the game. It features tanks, a large forest environment and even a few explosions. This level depicts the Battle of Peleliu, where American soldiers advance to capture an airstrip from the Japanese. It’s a level that’s both exciting to play and one that can bring even high-end systems to their knees.
Luckily for hardcore CoD players, the game’s performance doesn’t change with a faster CPU, which is rather impressive. Here, the game ran just as well on our lowly E5200 as it did on our i7-980X.
The original Half-Life 2 might have first seen the light of day close to four years ago, but it’s still arguably one of the greatest-looking games ever seen on the PC. Follow-up versions, including Episode One and Episode Two, do well to put the Source Engine upgrades to full use. While playing, it’s hard to believe that the game is based on a four+ year old engine, but it still looks great and runs well on almost any GPU purchased over the past few years.
Like Call of Duty 4, Half-Life 2: Episode Two runs well on modest hardware, but a recent mid-range graphics card is recommended if you wish to play at higher than 1680×1050 or would like to top out the available options, including anti-aliasing and very high texture settings.
This game benefits from both the CPU and GPU, and the skies the limit. In order to fully top out the available settings and run the highest resolution possible, you need a very fast GPU or GPUs along with a fast processor. Though the in-game options go much higher, we run our tests with 4xAA and 8xAF to allow the game to remain playable on the smaller mid-range cards.
Unlike CoD, HL2: Episode Two does love extra CPU power, and that’s evidenced above, but only at the highly-sporadic 1680×1050 resolution. That resolution has proven to be a chore, because the average FPS can fluctuate a great deal. What’s important to note here is that at our top setting of 2560×1600, the differences are almost zero.
As PC enthusiasts, we tend to be drawn to games that offer spectacular graphics… titles that help reaffirm your belief that shelling out lots of cash for that high-end monitor and PC was well worth it. But it’s rare when a game comes along that is so visually-demanding, it’s unable to run fully maxed out on even the highest-end systems on the market. In the case of the original Crysis, it’s easy to see that’s what Crytek was going for.
Funny enough, even though Crysis was released close to a year ago, the game today still has difficulty running at 2560×1600 with full detail settings – and that’s even with overlooking the use of anti-aliasing! Luckily, Warhead is better optimized and will run smoother on almost any GPU, despite looking just as gorgeous as its predecessor, as you can see in the screenshot below.
The game includes four basic profiles to help you adjust the settings based on how good your system is. These include Entry, Mainstream, Gamer and Enthusiast – the latter of which is for the biggest of systems out there, unless you have a sweet graphics card and are only running 1680×1050. We run our tests at the Gamer setting as it’s very demanding on any current GPU and is a proper baseline of the level of detail that hardcore gamers would demand from the game.
Our previous games didn’t show real differences between CPUs, and Crysis Warhead is no different. You can be rest-assured that no matter your PC, this game is going to run like molasses!
Although we generally shun automated gaming benchmarks, we do like to run at least one to see how our GPUs scale when used in a ‘timedemo’-type scenario. Futuremark’s 3DMark Vantage is without question the best such test on the market, and it’s a joy to use, and watch. The folks at Futuremark are experts in what they do, and they really know how to push that hardware of yours to its limit.
The company first started out as MadOnion and released a GPU-benchmarking tool called XLR8R, which was soon replaced with 3DMark 99. Since that time, we’ve seen seven different versions of the software, including two major updates (3DMark 99 Max, 3DMark 2001 SE). With each new release, the graphics get better, the capabilities get better and the sudden hit of ambition to get down and dirty with overclocking comes at you fast.
Similar to a real game, 3DMark Vantage offers many configuration options, although many (including us) prefer to stick to the profiles which include Performance, High and Extreme. Depending on which one you choose, the graphic options are tweaked accordingly, as well as the resolution. As you’d expect, the better the profile, the more intensive the test.
Performance is the stock mode that most use when benchmarking, but it only uses a resolution of 1280×1024, which isn’t representative of today’s gamers. Extreme is more appropriate, as it runs at 1920×1200 and does well to push any single or multi-GPU configuration currently on the market – and will do so for some time to come.
The results here are just as we expected, with a clean 50% increase on the CPU score.
Before discussing results, let’s take a minute to briefly discuss what I consider to be a worthwhile overclock. As I’ve mentioned in past content, I’m not as interested in finding the highest overclock possible as much as I am interested in finding the highest stable overclock. To me, if an overclock crashes the computer after a few minutes of running a stress-test, it has little value except for competition.
How we declare an overclock stable is simple… we stress it as hard as possible for a certain period of time, both with CPU-related tests and also GPU-related, to conclude on what we’ll be confident is 100% stability throughout all possible computing scenarios.
For the sake of CPU stress-testing, we use LinX. Compared to other popular CPU stress-testers, LinX’s tests are far more gruelling, and proof of that is seen by the fact that it manages to heat the CPU up to 20Â°C hotter than competing applications, like SP2004. Also, LinX is just as effective on AMD processors. Generally, if the CPU survives the first half-hour of this stress, there’s a good chance that it’s mostly stable, but I strive for a 12 hour stress as long as time permits.
If the CPU stress passes without error, then GPU stress-testing begins, in order to assure a system-wide stable overclock. To test for this, 3DMark Vantage’s Extreme test is used, with the increased resolution of 2560×1600, looped nine times. If this passes, some time is dedicated to real-world game testing, to make sure that gaming is just as stable as it would be if the CPU were at stock. If both these CPU and GPU tests pass without issue, we can confidently declare a stable overclock.
When it comes to overclocking, nothing has been simpler to deal with over the past couple of years than Intel’s processors, especially since Conroe. You can always pick up a dual or quad-core and expect a good experience, but for Gulftown, we’re dealing with two extra cores and a different die design, so it’s hard to just assume that we’ll see the same kind of overclocking potential as we’ve become accustomed to.
There’s also the concern that if just one core on the die is less-than-ideal, then it’s going to hold back your top-end overclock. To find a “perfect” chip might be rare, but I think generally speaking, it shouldn’t be too hard for anyone to achieve at least 4.0GHz on the i7-980X, and that’s basically what we accomplished as well, albeit with 50MHz tacked on top of it.
(Click to view LinX stress report)
I’ve always been rather pleased with a 4.0GHz overclock, and for the i7-980X, things were no different. But the more I thought about it, the less impressive such an overclock is on a chip like this is. The reason simply has to do with the fact that with Turbo, the i7-980X hits about 3.50GHz, so in essence, a 4.0GHz overclock is merely 500MHz above stock. Compare that to the 4.55GHz overclock we saw on our Core i3-530 last week. Now that’s an overclock.
In talking to Gigabyte, I was told that 4.2GHz should be possible, but for me, no matter what I tried, I couldn’t achieve stability with LinX. I might have been able to attain 4.1GHz and still retain that same stability, but I left it at 4.05GHz as all I had to do was increase the multiplier and voltage (to 1.350v). Because the Core i7-980X has an unlocked multiplier, you don’t have to touch the BCLK at all.
I’m still pleased with 4.0GHz, because though it might still be only 500MHz above stock, it’s a nice number to see. But do we actually have a reason to go through with it?
Intel Core i7-980X 3.33GHz (Overclock: 4.05GHz)
|Autodesk 3ds Max 2009|
|Cinebench R11.5 |
|Adobe Lightroom 2.0|
Convert 100 RAW to JPEG
HD Video Encode
Mobile Video Encode
HD Video Encode
DVD Video Encode
Big Number Crunch
It’s hard to say, unless you consider a 13% increase in performance to be worth all of the extra stress on the CPU. Personally, I don’t think so, because as it is, the i7-980X is extremely fast to begin with, and really, 10 or so percent really isn’t going to make much of a difference in the real-world. That changes if we’re talking about 25%+, but here, it’s just too difficult to tell.
If you’re running rendering projects or anything heavily math-based that takes hours or even days to complete, then the overclock might be a bit more worthy. After all, 90 hours is a lot better than 100 hours, when it comes to projects of that magnitude.
It goes without saying that power efficiency is at the forefront of many consumers’ minds today, and for good reason. Whether you are trying to save money or the environment – or both – it’s good to know just how much effort certain vendors are putting into their products to help them excel in this area. Both AMD and Intel have worked hard to develop efficient chips, and that’s evident with each new launch. The CPUs are getting faster, and use less power, and hopefully things will stay that way.
To help see what kind of wattage a given processor draws on average, we use a Kill-A-Watt that’s plugged into a power bar that’s in turn plugged into one of the wall sockets, with the test system plugged directly into that. The monitor and other components are plugged into the other socket and is not connected to the Kill-A-Watt. For our system specifications, please refer to our methodology page.
To test, the computer is first boot up and left to sit at idle for five minutes, at which point the current wattage is recorded if stable. To test for full CPU load, LinX is run with 2560MB memory usage for a total of five minutes. During that run, the highest point the wattage reaches on the meter is captured and becomes our “Max Load”.
On the first page of this article, I mentioned that the i7-980X had better power consumption than the i7-975, and here’s proof of that. Unlike most tests up to this point, where all of the Bloomfield processors were tested in a different motherboard, to be more fair here, I popped our i7-975 into the same board we used for Gulftown testing, so the two figures can reliably be compared.
The drop to 32nm has made an obvious improvement here, because both of those CPU’s are based on similar architectures, yet the six-core shaved 45W off at load, and about evens out at idle. When a six-core processor drains less power than a quad-core… that’s what I call impressive.
Where do I begin? When I first received the Core i7-980X, I didn’t think too much of it. Given the facts, I had an idea of what I could expect to see performance-wise. To be simple about it, the i7-980X is an i7-975 with +50% the number of cores and threads and is based on the Westmere architecture. From that alone, it’s fairly easy to surmise what the performance and experience will be like.
But as I sat down and started goofing around with the chip, I couldn’t help but get really excited about what I was dealing with. The experience was similar to the first time I touched a quad-core… it just felt like I had so much breathing room, and couldn’t find enough stuff to throw at it. The same can be said about the i7-980X, but maybe even more so. The Core i7-980X is an unbelievably powerful processor, and a total blast to use. It’s that simple.
I’m still not willing to buy into the fact that everyone needs a six-core processor, but for those who are willing to take the plunge, there’s a lot of fun to be had, and enough power to get your big projects done fast. Although most of our benchmarks that we used for this review aren’t entirely designed to the fullest for multi-core processors (especially 4+ cores), I was pleasantly surprised to see that despite that, the Core i7-980X still gave us a nice boost in many areas.
If you deal with either 3D rendering or media encoding on a regular basis, the Core i7-980X is a blessing. You can see the proof in our results tackling those scenarios… the increases can be huge. Plus, I am not even entirely pleased with some of the benchmarks we did run, such as the video encoding, since DivX is seemingly being used a lot less today. I’d like to follow-up in the future, once we get our new CPU suite sorted out, to see if even larger increases can be seen with today’s more popular HD codecs, such as WMV, MP4 or even MKV.
I should mention again about the superb power consumption we also saw from the i7-980X. As you can see on the previous page, it actually manages to use 45W less at full load compared to the i7-975, despite having two extra cores to deal with, and being built with a very similar microarchitecture. The shift to 32nm sure does deliver some nice rewards.
Wafer with Intel’s Gulftown Dies
One of the more impressive features with the Core i7-980X is the fact that it’s set to sell for $999. It’s not atypical for Intel to price its high-end models at that same price point, but we’ve just made a rather sizable leap here, and that’s not even considering that the company is even including a great CPU cooler in with the price. For those looking to build that new high-end rig, this is going to be hugely appreciated.
By now, you can probably conclude on whether a six-core processor can make a big difference in your life, and I can honestly say, if you are a power user, it certainly can. But if there is one disappointing thing about Gulftown as it stands today, it’s that it’s not going to be accessible to the vast majority of people for a while. But, given the beast we’re dealing with here, that’s not all too surprising.
As mentioned earlier, the fact that Gulftown is off-limits for the cash-conscious consumer works out to the favor for all those who are going to be taking the plunge. It means that they can purchase the Core i7-980X and actually have confidence that it’s going to remain Intel’s highest-end chip for a while. And also as mentioned, because we’re dealing with six-cores here, it means it’s pretty well impossible for someone to take a Bloomfield and overclock it high enough to match the performance, at least at 100% stable.
That’s where AMD’s Phenom X6 chips are going to be very interesting. Chances are good that X6 chips will carry a bit of a premium as well, but you can bet that they’re going to be much more affordable than $999. Given AMD’s current architecture, its X6 chips aren’t going to compete that closely with Intel’s Gulftown, but it’s certainly going to be an option for those looking for a six-core at an easier-to-stomach price.
Intel’s Core i7-980X Extreme Edition
Have a comment you wish to make on this article? Recommendations? Criticism? Feel free to head over to our related thread and put your words to our virtual paper! There is no requirement to register in order to respond to these threads, but it sure doesn’t hurt!
Copyright © 2005-2017 Techgage Networks Inc. - All Rights Reserved.