The catch however is that what we don’t have is a level of clear domination when it comes to single-card solutions. AMD was shooting to beat the GTX 295 with the 5870, but in our benchmarks that’s not happening. The 295 and the 5870 are close, perhaps close enough that NVIDIA will need to reconsider their position, but it’s not enough to outright dethrone the GTX 295. NVIDIA still has the faster single-card solution, although the $100 price premium is well in excess of the <10%>-From Our Radeon 5870 Review, On The GTX 295 vs. The 5870
Let’s get straight to the point, shall we? Today AMD is launching the 5970, their dual-GPU card that finishes building out AMD’s technical domination of the high-end market. With it AMD delivers the absolute victory over NVIDIA’s GTX 295 that the Radeon 5870 couldn’t quite achieve and at the same time sets the new high water mark for single-card performance.
This also marks the last AMD product introduction of the year. The rest of the Evergreen series, composing the sub-$100 low-end parts, will be launching next year.
AMD Radeon HD 5970 AMD Radeon HD 5870 AMD Radeon HD 5850 Stream Processors 2x1600 1600 1440 Texture Units 2x80 80 72 ROPs 2x32 32 32 Core Clock 725MHz 850MHz 725MHz Memory Clock 1GHz (4GHz data rate) GDDR5 1.2GHz (4.8GHz data rate) GDDR5 1GHz (4GHz data rate) GDDR5 Memory Bus Width 2x256-bit 256-bit 256-bit Frame Buffer 2x1GB 1GB 1GB Transistor Count 2x2.15B 2.15B 2.15B TDP 294W 188W 151W Manufacturing Process TSMC 40nm TSMC 40nm TSMC 40nm Price Point $599 $400 $300
The 5970 serves as the nowadays obligatory dual-GPU part. It is 2 Cypress dice mounted on a single, dual-slot video card. AMD clocks it at 725MHz core and 1GHz (4GHz effective) for the GDDR5 memory. The card comes equipped with 2GB of GDDR5, which is split between the two GPUs, giving it an effective memory capacity of 1GB. The card will be selling for $600, at least so long as vendors and retailers hold the line on MSRP.
In practice this makes the card something between a 5850 in Crossfire mode and a 5870 in Crossfire mode. The clocks are the same as the 5850, but here all 20 SIMD units are enabled. This is a 15% clockspeed difference between the 5970 and 5870CF, so officially the 5870CF will continue to be the faster setup. However as we’ll see in a bit, looking at the stock 5970 can be a bit deceiving.
This also brings up the matter of the name of the card. We asked AMD what happened to the X2 tag, and the answer is that they didn’t want to use it since the card was configured neither like a 5850 nor a 5870 – it was closer to a mythical 5860. So rather than call it an odd (or worse yet, wrong) name, AMD just gave it a new model number entirely. We suspect AMD wanted to be rid of the X2 name – their processors go up to X4 after all – but there you go as far as an official reason is concerned. It looks like special multi-GPU tags are now gone in both the NVIDIA and AMD camps.
Moving on, for power, the 5970 uses an 8pin and a 6pin power connector (although the 6pin sits on top of a spot silk-screened for anther 8pin). The TDP is 294W, bringing it in just under the 300W ATX limit. Idle power is 42W, thanks to AMD’s aggressive power optimizations present in the entire 5000 series.
As some of you may have noticed, in spite of the fact that this card is at least a pair of 5850s, it consumes less than the 320W (2x160W) such a setup would. In order to meet the 300W limit, AMD went and binned Cypress chips specifically for the 5970, in order to find chips that could operate at 725MHz at only 1.05v (the 5850 runs at 1.088v). Given the power creep coming from the 4800 series, binning for the best chips is the only way AMD could get a 300W card out.
AMD’s official guidance for this card is that the minimum requirements are a 650W power supply, and they recommend a 750W power supply. The recommended power supply will become more important later on when we talk about overclocking.
Finally, AMD is also launching Crossfire Eyefinity support with the 5970, and thus far only the 5970. Currently Eyefinity doesn’t work with Crossfire mode on any of AMDs cards due to driver limitations. The drivers that the 5970 will be shipping with enable Crossfire Eyefinity support on the 5970 for 22 games – currently AMD is using whitelisting and is enabling games on a case-by-case basis. Crossfire Eyefinity will make its way in to the mainstream Catalyst drivers and be enabled for other cards early next year.
Monday, January 4, 2010
The Radeon HD 5970: Completing AMD's Takeover of the High End GPU Market
GPU Accelerated Flash 10.1 Prerelease
I suppose I could start this article off with a tirade on how frustrating Adobe Flash is. But, I believe the phrase “preaching to the choir” would apply.
I’ve got a two socket, 16-thread, 3GHz, Nehalem Mac Pro as my main workstation. I have an EVGA GeForce GTX 285 in there. It’s fast.
It’s connected to a 30” monitor, running at its native resolution of 2560 x 1600.
The machine is fast enough to do things I’m not smart or talented enough to know how to do. But the one thing it can’t do is play anything off of Hulu in full screen without dropping frames.
This isn’t just a Mac issue, it’s a problem across all OSes and systems, regardless of hardware configuration. Chalk it up to poor development on Adobe’s part or...some other fault of Adobe’s, but Flash playback is extremely CPU intensive.
Today, that’s about to change. Adobe has just released a preview of Flash 10.1 (the final version is due out next year) for Windows, OS X and Linux. While all three platforms feature performance enhancements, the Windows version gets H.264 decode acceleration for flash video using DXVA (OS X and Linux are out of luck there for now).
The same GPU-based decode engines that are used to offload CPU decoding of Blu-rays can now be used to decode H.264 encoded Flash video. NVIDIA also let us know that GPU acceleration for Flash animation is coming in a future version of Flash.
To get the 10.1 pre-release just go here. NVIDIA recommends that you uninstall any existing versions of flash before installing 10.1 but I’ve found that upgrading works just as well.
What Hardware is Supported?
As I just mentioned, Adobe is using DXVA to accelerate Flash video playback, which means you need a GPU that properly supports DXVA2. From NVIDIA that means anything after G80 (sorry, GeForce 8800 GTX, GTS 640/320MB and Ultra owners are out of luck). In other words anything from the GeForce 8 series, 9 series or GeForce GT/GTX series, as well as their mobile equivalents. The only exceptions being those G80 based parts I just mentioned.
Anything based on NVIDIA’s ION chipset is also supported, which will be the foundation of some of our tests today.
AMD supports the following:
- ATI Radeon™ HD 4000, HD 5700 and HD 5800 series graphics
- ATI Mobility Radeon™ HD 4000 series graphics (and higher)
- ATI Radeon™ HD 3000 integrated graphics (and higher)
- ATI FirePro™ V3750, V5700, V7750, V8700 and V8750 graphics accelerators (and later)
It’s a healthy list of supported GPUs from both camps, including integrated graphics. The only other requirement is that you have the latest drivers installed. I used 195.50 from NVIDIA and Catalyst 9.10 from AMD. (Update: The Release Notes now indicate Catalyst 9.11 drivers are required, which would explain our difficulties in testing. ATI just released Catalyst 9.11 but we're having issues getting GPU acceleration to work, waiting on a response from AMD now)
Intel’s G45 should, in theory, work. We tested it on a laptop for this article and since the acceleration is DXVA based, anything that can offload H.264 decode from the CPU using DXVA (like G45) should work just fine. As you’ll see however, our experiences weren’t exactly rosy.
Intel Canceling Larrabee Prime
Larrabee is Dead, Long Live Larrabee
Intel just announced that the first incarnation of Larrabee won't be a consumer graphics card. In other words, next year you're not going to be able to purchase a Larrabee GPU and run games on it.
You're also not going to be able to buy a Larrabee card and run your HPC workloads on it either.
Instead, the first version of Larrabee will exclusively be for developers interested in playing around with the chip. And honestly, though disappointing, it doesn't really matter.
The Larrabee Update at Fall IDF 2009
Intel hasn't said much about why it was canceled other than it was behind schedule. Intel recently announced that an overclocked Larrabee was able to deliver peak performance of 1 teraflop. Something AMD was able to do in 2008 with the Radeon HD 4870. (Update: so it's not exactly comparable, the point being that Larrabee is outgunned given today's GPU offerings).
With the Radeon HD 5870 already at 2.7 TFLOPS peak, chances are that Larrabee wasn't going to be remotely competitive, even if it came out today. We all knew this, no one was expecting Intel to compete at the high end. Its agents have been quietly talking about the uselessness of > $200 GPUs for much of the past two years, indicating exactly where Intel views the market for Larrabee's first incarnation.
Thanks to AMD's aggressive rollout of the Radeon HD 5000 series, even at lower price points Larrabee wouldn't have been competitive - delayed or not.
I've got a general rule of thumb for Intel products. Around 4 - 6 months before an Intel CPU officially ships, Intel's partners will have it in hand and running at near-final speeds. Larrabee hasn't been let out of Intel hands, chances are that it's more than 6 months away at this point.
By then Intel wouldn't have been able to release Larrabee at any price point other than free. It'd be slower at games than sub $100 GPUs from AMD and NVIDIA, and there's no way that the first drivers wouldn't have some incompatibly issues. To make matters worse, Intel's 45nm process would stop looking so advanced by mid 2010. Thus the only option is to forgo making a profit on the first chips altogether rather than pull an NV30 or R600.
So where do we go from here? AMD and NVIDIA will continue to compete in the GPU space as they always have. If anything this announcement supports NVIDIA's claim that making these things is, ahem, difficult; even if you're the world's leading x86 CPU maker.
Do I believe the 48-core research announcement had anything to do with Larrabee's cancelation? Not really. The project came out of a different team within Intel. Intel Labs have worked on bits and pieces of technologies that will ultimately be used inside Larrabee, but the GPU team is quite different. Either way, the canceled Larrabee was a 32-core part.
A publicly available Larrabee graphics card at 32nm isn't guaranteed, either. Intel says they'll talk about the first Larrabee GPU sometime in 2010, which means we're looking at 2011 at the earliest. Given the timeframe I'd say that a 32nm Larrabee is likely but again, there are no guarantees.
It's not a huge financial loss to Intel. Intel still made tons of money all the while Larrabee's development was underway. Its 45nm fabs are old news and paid off. Intel wasn't going to make a lot of money off of Larrabee had it sold them on the market, definitely not enough to recoup the R&D investment, and as I just mentioned using Larrabee sales to pay off the fabs isn't necessary either. Financially it's not a problem, yet. If Larrabee never makes it to market, or fails to eventually be competitive, then it's a bigger problem. If heterogenous multicore is the future of desktop and mobile CPUs, Larrabee needs to succeed otherwise Intel's future will be in jeopardy. It's far too early to tell if that's worth worrying about.
One reader asked how this will impact Haswell. I don't believe it will, from what I can tell Haswell doesn't use Larrabee.
Intel has a different vision of the road to the CPU/GPU union. AMD's Fusion strategy combines CPU and GPU compute starting in 2011. Intel will have a single die with a CPU and GPU on it, but the GPU isn't expected to be used for much compute at that point. Intel's roadmap has the CPU and AVX units being used for the majority of vectorized floating point throughout 2011 and beyond.
Intel's vision for the future of x86 CPUs announced in 2005, surprisingly accurate
It's not until you get in the 2013 - 2015 range that Larrabee even comes into play. The Larrabee that makes it into those designs will look nothing like the retail chip that just got canceled.
Intel's announcement wasn't too surprising or devastating, it just makes things a bit less interesting.
Epic Demonstrates Unreal Engine 3 for the iPod Touch/iPhone 3GS
I got together with Mark Rein last week and he showed me an Unreal Engine 3 tech demo running on a 3rd generation iPod Touch. The same Unreal Engine 3 that powers Gears of War 2, running on an iPod Touch. The engine also works on the iPhone 3GS, and Mark tells me that we’ll see it on another mobile platform at CES (hmm...).
The demo is both playable and has a flythrough. It’s using a modified Unreal Tournament level previously shown off at GDC. A virtual thumbstick on the left side of the screen controls your movement, while tracking your thumb in the lower right corner of the screen controls the camera. Just tap the screen to shoot. Mark said this is a tech test bed and they’re experimenting with several different control schemes including ones with tilt.
In practice, the controls work well. This is just a demo so there was no score or point to the game, I just got to run around and kill a single respawning enemy. And it was fun.
As you can see from the video the frame rate was smooth. There are more visuals to be added, as well as some polishing, but the demo looked very good for an iPhone game.
It requires OpenGL ES 2.0, so the iPhone 2G and 3G won’t work, nor will the older iPod Touch models. It doesn’t really matter though, this is just a starting point.
Epic isn’t announcing any sort of iPhone engine licenses nor are they entering the iPhone game market. Porting UE3 to the iPhone is simply one of many projects being worked on inside a newer, more svelte and innovative Epic Games (wait till you see what’s next...).
Today the iPhone, Tomorrow the World
Mark said they planned to make this available to licensees at some point in the near future. That’s great for end users because it means that any Unreal Engine licensee can now start playing around with making iPhone games based on the same technology. Unfortunately the recently announced, free to the public, Unreal Development Kit (UDK) is Windows only - the iPhone version isn’t included. I’d guess that at some point Epic will change that, it just makes too much sense. Doing so would enable a whole new class of iPhone game development using an extremely polished engine.
It’s all about taking the portable market seriously. While I wouldn’t expect to see any Epic branded iPhone games anytime soon, eventually it wouldn’t be too far fetched to see a full port of Gears of War to something as small as an iPhone. NAND Flash capacities to support multiple 9GB games will be there in another few years, as will GPU horsepower.
Remember that the SoC in the iPhone 3GS is only built on 65nm technology, Intel is about to release its first 32nm chips. You could cram four times as many transistors into the same space at 32nm, roughly 9 times as many at 22nm. Remember that graphics performance scales very well with additional transistors. At what point does the smartphone become more powerful than an Xbox 360? Sometime in the next 3 - 5 years for sure.
And it’s not just about iPhone support. Mark told me that as soon as CES we’ll see Unreal Engine 3 on another mobile platform entirely. More announcements will happen throughout 2010. This isn’t a platform specific thing, it’s about bringing Unreal Engine 3 to the entire portable market.
Final Words
For the end user, Epic just improved the chances of getting better looking games on the iPhone and potentially other portable devices.
For Epic, expanding UE3 into the portable market makes a lot of sense - it will eventually increase the base of paying UE licensees as well as help move the entire portable gaming industry forward. It’s not all altruistic though, by releasing iPhone and other portable versions of UE3 it helps secure Epic’s position as a supplier of game engines, regardless of platform.
For an existing Unreal Engine licensee you now have the ability to compete in the growing iPhone market thanks to Epic, how nice. Once the phone gets powerful enough I expect it’ll be used for more than just playing simple games.
As personal computing moves to more platforms and takes new forms, what we need are technologies that unify development across all devices. There’s no reason that a game you’re writing for an iPhone shouldn’t be built on the same foundation as something you’re writing for a high end console. The difference should be in the game, not in the engine.
This latest move by Epic does validate whatever Apple has been quietly doing all along. From the investments in Imagination Technologies to hiring two previous AMD Graphics CTOs, Apple is clearly interested in gaming (which is funny given the poor state of gaming on the Mac).
Historically Apple likes to enter markets when it believes that it can do something better or at least different. We saw that with the iPhone. The question is, how does Apple plan on providing a different take on gaming?
The remaining hurdles are significant, but not unsurmountable. Playing anything other than a point and touch game on the iPhone can be frustrating. Epic doesn’t address that, but someone else surely will when the time is right.
Engine houses like Epic enable game developers to focus on building the game they want to build, not the underlying technology. With UE3 on the iPhone, we will eventually see more and hopefully better games on future versions of the platform. Not to mention whatever other mobile platforms Epic plans on porting UE3 to as well.
XBMC & Broadcom Bring 1080p Decode Upgrade to ill-equipped netbooks, nettops, Apple TVs
We finally have tons of machines that come in these tiny little boxes, sometimes with pretty interfaces, that hook straight up to our TVs. We’ve been asking for this for years and we finally have it. They’re even downright affordable. In order to make them affordable, they use cheap hardware. In particular, slow CPUs. they’re not that bad. Definitely fast enough to browse a bunch of movies or TV shows, but generally too slow to play back high definition video.
Pine Trail, a great little platform, uses very little power, it just can't play 1080p video
NVIDIA built an entire platform out of addressing this deficiency. It’s called ION, and it mates a fast-enough-for-most-things CPU with a GPU capable of decoding 1080p video. Hooray for NVIDIA. Here’s the problem - ION is not retrofittable.
ION, it's a new system, not an upgrade
If you have a non-ION netbook, nettop, AppleTV, Mac Mini, or other impressively tiny device that you want to use to drive high definition video you can’t. And you can’t upgrade them to enable such support.
It’s not like buying a hard drive or more memory. Well, it wasn’t at least.
In one of my last articles on ION I talked about XBMC being one of the best applications for the platform. It delivers a better UI for watching downloaded content than Windows Media Center does, and the Linux version has full hardware acceleration support for ION. Oh, and it’s free.
Last week it got even better. The next major version of XBMC, codenamed Camelot, came out with a whole bunch of new features. I haven’t had a chance to play with it yet but it definitely looks cool.
And today it just got incredible. Thanks to the efforts of Scott Davilla, the XBMC developers and Broadcom there’s now full support for the Broadcom Crystal HD decoder (BCM970012) in all versions of XBMC. The code has been merged into XBMC as of this morning and will be available in the next release of the project.
The mini PCIe BCM970012 from Logic Supply
The Crystal HD chip is currently available on mini PCIe cards, and is expected to make its way to ExpressCard and PCIe x1 cards in the future. This is huge because many netbooks, nettops and existing devices like the AppleTV or Mac Mini have at least one mini PCIe slot. A $69 mini PCIe card (or $25 on eBay) with the Crystal HD chip on it plus the next version of XBMC can enable full 1080p playback on any of these machines that would otherwise not be able to play high def video. Not to mention that you can get some of these devices second hand or refurbished for much less than the cost of a new ION system.
It’s extra sweet because the driver is open source, so we can expect to see it more than just XBMC. The next official release of XBMC is likely some time away, but support has already been added to the SVN release.
The full press release from XBMC is below.
Broadcom Crystal HD, It's Magic.
1080p HD content playback has always been the Holy Grail for any Media Center application but this has traditionally been difficult; playback using software decode alone requires a very hefty CPU and hardware decoding has only been made available recently using the nVidia's VPDAU technology, available only on Linux. Windows has its own platform specific solutions and poor old OSX has no public APIs available at all. There really is nothing around with a common API that enables hardware accelerated 1080p HD content playback that can also be used under all three major platforms (OSX, Linux, and Windows). Well, that situation is about to change.
Through hard work and the joint efforts of several TeamXBMC/Redhat developers and the Broadcom Media PC Group, cross-platform hardware decoding of mpeg2, h.264 and VC1 video content up to 1080p will be coming to XBMC on OSX, Linux, and Windows via the Broadcom Crystal HD Hardware Decoder (BCM970012). The Broadcom Crystal HD is available now in a mini-PCIE card with ExpressCard and 1X PCIE form factors to follow. This means that the AppleTV and all those lovely new netbooks, Eee Boxes and older Intel Mac Minis have exciting new potential.
This solution has a common programming API, so many 3rd party developers and applications will be able to leverage hardware accelerated video content playback across OSX, Linux, and Windows platforms with minimal source code changes. Best of all, this is an open source solution with full source code for driver and library available for OSX and Linux under a GPL/LGPL license. Wow, this indeed is the Holy Grail and a major score for the open source community as this means no more tainted Linux kernels! Support has already been added to XBMC under the svn trunk. Other media projects such as FFmpeg, MythTV and Xine will soon follow as their developers add
support.
The Linux source code for the driver and library can be found at the Broadcom web site. For OSX, the binaries and source code for the driver and library will be hosted at http://crystalhd-for-osx.googlecode.com. Users in the USA can purchase the Crystal HD mini PCIE card from Logic Supply for $69 and of course, there's always ebay for those outside the USA.
Linux Gaming: Are We There Yet?
Introduction by Jarred
A few months back, I wrote an article looking at battery life on a couple of laptops using several different OSes. Windows XP, Vista, and 7 were the main focus, but I decided to test battery life on Linux running Ubuntu as well. Naturally, the Linux community wasn't happy to see their OS place last in the battery life results. One of the readers actually took the time to offer his help in getting a Linux OS configured "properly", and we started work.
Eventually, we abandoned the follow-up article as I had other pressing matters to get to and troubleshooting Linux from half a continent away is difficult (plus we started with the NV52, and ATI support under Linux is still lacking), but I offered Christopher the chance to write a couple pilot Linux articles for our site. He had shown a much better knowledge of Linux and I figured getting someone passionate about the OS was our best bet. (I'll let Christopher tell you what he does for his "real job" if he so desires.)
As a side note, with Anand's call for more writers, this sort of passion for any technology is what we like to see. Show us you know what you're talking about and you care about the technology, and there's a good chance we can use your skills.
With that out of the way, here's Christopher's first Linux article where he provides a concrete look at something few have ever attempted: Windows Gaming on Linux.
Over the past few years, there has been a common question on the Linux vs. Windows desktop front: does the Linux desktop have the ability to play various major release games, and if so what is the performance difference between the two? Linux is commonly overlooked as a viable gaming platform in most communities. Our intention today is to shed some light on what does and does not work inside Linux, as well as give solid performance data for those looking for another option in the gaming world. Each OS has areas where it shows superiority over the other, but for the sake of staying true to the purpose of this article we will only be focusing on the game performance/functionality differences.
Unfortunately there are very few game releases that support running inside Linux natively. To combat this issue there are a few Linux projects that will allow Linux users to run Windows applications - note that we did not say "emulate Windows". We have selected three Linux projects in order to complete our initial round of testing. Our open source project selection "Wine" is a free, easily downloadable project that is created to support both Windows games and applications. The second selection, "Cedega", is a closed source implementation of Wine focused on gaming. The final selection is Crossover Games which like Cedega is a closed source implementation of Wine allowing for enhanced usability and gameplay over Wine.
Some are probably asking at this point, what is Wine? The Wine project was started in 1993 to allow Windows applications to be run under Linux. Wine takes the Windows API (Application Programming Interface) and implements it in the Linux user space. Since Wine is running in user space and is not a part of the Linux Kernel, it relies on the wineserver daemon to provide your basic Windows kernel functionality as well as other various tasks of X integration.
As a quick recap, there are two basic goals we want to complete in this article. First we want to compare performance and functionality of games between Linux and Windows. Second, we will look at the performance and functionality differences of Wine/Cedega/Crossover Games.
AMD Athlon 2650e & X2 3250e - Better than Atom for SFF Desktops
Dell has really come into its own in the past few years. Shedding the image (and fur?) of yet another cheap PC OEM, Dell has tried to make design a more important part of its manufacturing as of late. Not every design is a win, but the zino HD appears to be the first nettop styled to have mass market appeal:
Dell's Inspiron zino HD, available in more than just pink
As soon as Dell announced the zino HD I was bombarded with emails. Not for a review of the system, but for a review of the processors inside the machine. A quick look at Dell’s tech specs for the zino HD list the following processor options:
They sound fast. But in practice, many of these chips are actually Atom competitors. I asked AMD for two of the CPUs on this list - the Athlon 2650e and the Athlon X2 3250e.
The 2650e is a single-core K8 based Socket-AM2/AM2+ processor that runs at 1.6GHz. It’s got 512KB of L2 cache and the rest of the feature set of the older K8 AM2 processors. As an AM2 chip it’s DDR2 only and physically won’t fit or work in an AM3 motherboard. Keep that in mind before you go sticking one of these where it doesn’t belong.
Based on its specs, the Athlon 2650e is a near perfect competitor for Intel’s Atom. The newly announced D510 runs at 1.66GHz (close enough), but is an in-order architecture with Hyper Threading. Bottom line? The Athlon 2650e should be faster in single threaded apps, but slower in multitasking/multithreaded applications. Great. Atom boasts a lower total TDP as well.
The Athlon X2 3250e fixes the 2650e’s biggest shortcoming by adding a second core. The clock speed drops a bit to 1.5GHz, but the rest of the specs stay the same. TDP goes up from 15W on the 2650e to 22W for the 3250e.
Dell charges an extra $65 for the 3250e upgrade. Just to put things in perspective - $10 more and you could have yourself a complete Pine Trail platform.
AMD Athlon 2650e & X2 3250e - Better than Atom for SFF Desktops
Dell has really come into its own in the past few years. Shedding the image (and fur?) of yet another cheap PC OEM, Dell has tried to make design a more important part of its manufacturing as of late. Not every design is a win, but the zino HD appears to be the first nettop styled to have mass market appeal:
Dell's Inspiron zino HD, available in more than just pink
As soon as Dell announced the zino HD I was bombarded with emails. Not for a review of the system, but for a review of the processors inside the machine. A quick look at Dell’s tech specs for the zino HD list the following processor options:
They sound fast. But in practice, many of these chips are actually Atom competitors. I asked AMD for two of the CPUs on this list - the Athlon 2650e and the Athlon X2 3250e.
The 2650e is a single-core K8 based Socket-AM2/AM2+ processor that runs at 1.6GHz. It’s got 512KB of L2 cache and the rest of the feature set of the older K8 AM2 processors. As an AM2 chip it’s DDR2 only and physically won’t fit or work in an AM3 motherboard. Keep that in mind before you go sticking one of these where it doesn’t belong.
Based on its specs, the Athlon 2650e is a near perfect competitor for Intel’s Atom. The newly announced D510 runs at 1.66GHz (close enough), but is an in-order architecture with Hyper Threading. Bottom line? The Athlon 2650e should be faster in single threaded apps, but slower in multitasking/multithreaded applications. Great. Atom boasts a lower total TDP as well.
The Athlon X2 3250e fixes the 2650e’s biggest shortcoming by adding a second core. The clock speed drops a bit to 1.5GHz, but the rest of the specs stay the same. TDP goes up from 15W on the 2650e to 22W for the 3250e.
Dell charges an extra $65 for the 3250e upgrade. Just to put things in perspective - $10 more and you could have yourself a complete Pine Trail platform.
Intel Arrandale: 32nm for Notebooks, Core i5 540M Reviewed
Clarkdale is the desktop processor, but Arrandale is strictly for my notebooks. The architecture is the same as Clarkdale. You've got a 32nm Westmere core and a 45nm chipset on the same package:
The two-chip solution does matter more for notebooks as it means that motherboards can shrink. Previously this feature was only available to OEMs who went with NVIDIA's ION platform (or GeForce 9400M as it was once known). This is the first incarnation of Intel's 32nm process so it's not quite as power optimized as we'd like. The first mainstream Arrandale CPUs are 35W TDP, compared to the 25W TDP of most thin and light notebooks based on mobile Core 2. Granted the 35W includes the graphics, but it's not always going to be lower total power consumption (more on this later).
The Arrandale lineup launching today is huge. Intel launched 7 Clarkdale CPUs, but we've got a 11 mobile Arrandale CPUs coming out today:
The architecture is similar to Clarkdale. You get private 256KB L2s (one per core) and a unified L3 cache for the CPU. The L3 is only 3MB (like the Pentium G9650) on the Core i5 and Core i3 processors, but it's 4MB (like the desktop Core i5/i3) on the mobile Core i7. Confused yet? I'll have to admit, Intel somehow took a potentially simple naming scheme and made it unnecessarily complex. We also get some low-voltage parts that have 18W TDPs. They run at low default clock speeds but can turbo up pretty high.
Turbo is hugely important here. While Clarkdale's Turbo isn't exactly useful, the TDPs are low enough in mobile that you can really ramp up clock speed if you aren't limited by cooling. Presumably this will allow you to have ultra high performance plugged-in modes where your CPU (and fans) can ramp up as high as possible to get great performance out of your notebook. Add an SSD and the difference between a desktop and a notebook just got even smaller.
Arrandale does have one trick that Clarkdale does not: graphics turbo.
GPU bound applications (e.g. games) can force the CPU part of Arrandale into a low power state, and the GPU can use the added thermal headroom to increase its clock speed. This is a mobile only feature but it's the start of what will ultimately be the answer to achieving a balanced system. Just gotta get those Larrabee cores on-die...
Chipsets are even more complicated on the mobile side:
The Clarkdale Review: Intel's Core i5 661, i3 540 & i3 530
I swear this is the longest it’s taken for an Intel architecture to penetrate the market. We first met Nehalem on November 3rd, 2008. It came to us as a high end quad-core processor and took a full year to make it to more affordable motherboards in the form of Lynnfield. Even with P55 motherboard prices down at the magical $99 marker, Intel relinquished control of the $100 - $200 CPU market without a Nehalem to compete down there. Instead we were left with a choice between Penryn, the update to Intel’s 2006 Conroe architecture, or Phenom II, AMD’s low-cost Nehalem competitor. The choice was simple.
From $100 to $200, your best bet has been AMD. Either through aggressive pricing on quad-core CPUs or the L3-cache-less Athlon II line, AMD controls the $100 - $200 market. Today we meet Intel's first 32nm CPUs, codename Clarkdale, designed to specifically target that $100 - $200 market.
Two cores, Nehale..err Westmere-style
Technically Clarkdale isn’t Nehalem, it’s Westmere. Take Nehalem, use 32nm transistors, add in some new instructions for accelerating encryption/decryption, and you’ve got the makings of Westmere.
Clarkdale uses a dual-core Westmere and sticks it next to a 45nm Intel GMA die. That’s right, meet the first (er, second) Intel CPU with on-chip graphics. Next year we’ll see Sandy Bridge bring the graphics on-die, but until then we have Intel’s tried and true multi-chip-package to tide us over.
We don’t get on-die graphics yet because Intel still hasn’t switched over to its make-everything-at-the-best-process-ever strategy. The 32nm fabs are ramping up with CPU production and the 45nm fabs need something to do. Nearly every desktop and laptop sold in 2010 will need one of these 45nm GMA die, so the fabs indeed have something to do.
It’s not all rosy with Clarkdale unfortunately. Remember the memory controller that Nehalem so graciously integrated? Clarkdale kicks it off die again. The IMC is housed in the 45nm GMA die. It’s still on-package, but not on-die. The benefit is super fast memory access for the graphics core, but slower memory access for the CPU core. In fact, this is a derivative of the memory controller used in older Intel chipsets (e.g. P45/X48).
The CPU connects to the GMA die using QPI, so bandwidth shouldn’t be as big of a problem. Latency is unfortunately hurt as a result. Access times can be longer than older LGA-775 processors thanks to this memory controller design being optimized for FSB architectures. Again, the problem goes away with Sandy Bridge, but today we’re faced with it.
Like Lynnfield, Clarkdale uses Intel’s LGA-1156 socket. Clarkdale should work in all P55 motherboards, but you’ll need a BIOS update. No existing P55 motherboards route video from the socket to a VGA/DVI/HDMI connector, so you’ll need a new motherboard if you want to take advantage of the on-package graphics. Enter the H55, H57 and Q57 chipsets.
A typical H57 I/O layout
The H and Q-series of chipsets feature what Intel calls its Flexible Display Interface (FDI). This is basically a link between the CPU and the chipset that passes along video output. H5x/Q57 motherboards will have a video out on board so you can use Clarkdale’s integrated graphics.
The chipsets differ in price and features. The table below sums it up:
Click to Enlarge
Support for Intel’s Rapid Storage Technology is probably the main reason you’ll want H57 over H55. The difference between H57 and Q57 boils down to security and management features. The H-series is for consumers, the Q-series is for corporate customers. Make sense?
The chips themselves are pretty straightforward. Intel is launching seven desktop Clarkdale processors (and a whole lot more notebook Arrandale chips):
| Processor | Core Clock | Cores / Threads | L3 Cache | Max Turbo | TDP | Price |
| Intel Core i5-670 | 3.46GHz | 2 / 4 | 4MB | 3.76GHz | 73W | $284 |
| Intel Core i5-661 | 3.33GHz | 2 / 4 | 4MB | 3.60GHz | 87W | $196 |
| Intel Core i5-660 | 3.33GHz | 2 / 4 | 4MB | 3.60GHz | 73W | $196 |
| Intel Core i5-650 | 3.20GHz | 2 / 4 | 4MB | 3.46GHz | 73W | $176 |
| Intel Core i3-540 | 3.06GHz | 2 / 4 | 4MB | N/A | 73W | $133 |
| Intel Core i3-530 | 2.93GHz | 2 / 4 | 4MB | N/A | 73W | $113 |
| Intel Pentium G9650 | 2.80GHz | 2 / 2 | 3MB | N/A | 73W | $87 |
The six processors labeled Core i5s and Core i3s all have the same basic architecture. You get per-core 256KB private L2s and you get a 4MB L3 cache shared among both cores (down from 8MB from the quad-core Lynnfield/Bloomfield chips). The i5s get turbo mode while the i3s do not. The i5-661 uses a higher graphics clock and has a higher TDP than the 660. Remember that these are CPU+GPU combos on a single package, so GPU clocks do vary based on model.
The Clarkdale lineup is honestly made up of CPUs that are too expensive. The Core i5 670, 661/660 and 650 are all priced above $170 and aren’t worth the money. The problem is Lynnfield’s turbo mode gives you high enough clock speeds with two threads that there’s no need to consider a dual-core processor. You can buy a Core i5 750, have more cores than any of these Clarkdales and run at close enough to the same frequencies for $196. Or you can buy a Phenom II X4 965 for about the same price and have much better multi-threaded performance. The chips that are most interesting are the Core i3s.
| Processor | Clock Speed | Max Turbo | |
| 2 Cores Active | 1 Core Active | ||
| Intel Core i5-670 | 3.46GHz | 3.60GHz +3.9% | 3.73GHz +7.7% |
| Intel Core i5-661 | 3.33GHz | 3.46GHz +4.0% | 3.60GHz +8.0% |
| Intel Core i5-660 | 3.33GHz | 3.46GHz +4.0% | 3.60GHz +8.0% |
| Intel Core i5-650 | 3.20GHz | 3.33GHz +4.2% | 3.46GHz +8.3% |
| Intel Core i3-540 | 3.06GHz | N/A | |
| Intel Core i3-530 | 2.93GHz | N/A | |
Turbo just isn't as interesting with only two cores. With four cores you used to have to make a tradeoff between good 4 and 2 threaded performance, but Lynnfield fixed that. No one really debates single vs. dual core anymore. The single core turbo modes are great, but aren't worth the money. Pay attention to the i3s.
The Clarkdale Review: Intel's Core i5 661, i3 540 & i3 530
I swear this is the longest it’s taken for an Intel architecture to penetrate the market. We first met Nehalem on November 3rd, 2008. It came to us as a high end quad-core processor and took a full year to make it to more affordable motherboards in the form of Lynnfield. Even with P55 motherboard prices down at the magical $99 marker, Intel relinquished control of the $100 - $200 CPU market without a Nehalem to compete down there. Instead we were left with a choice between Penryn, the update to Intel’s 2006 Conroe architecture, or Phenom II, AMD’s low-cost Nehalem competitor. The choice was simple.
From $100 to $200, your best bet has been AMD. Either through aggressive pricing on quad-core CPUs or the L3-cache-less Athlon II line, AMD controls the $100 - $200 market. Today we meet Intel's first 32nm CPUs, codename Clarkdale, designed to specifically target that $100 - $200 market.
Two cores, Nehale..err Westmere-style
Technically Clarkdale isn’t Nehalem, it’s Westmere. Take Nehalem, use 32nm transistors, add in some new instructions for accelerating encryption/decryption, and you’ve got the makings of Westmere.
Clarkdale uses a dual-core Westmere and sticks it next to a 45nm Intel GMA die. That’s right, meet the first (er, second) Intel CPU with on-chip graphics. Next year we’ll see Sandy Bridge bring the graphics on-die, but until then we have Intel’s tried and true multi-chip-package to tide us over.
We don’t get on-die graphics yet because Intel still hasn’t switched over to its make-everything-at-the-best-process-ever strategy. The 32nm fabs are ramping up with CPU production and the 45nm fabs need something to do. Nearly every desktop and laptop sold in 2010 will need one of these 45nm GMA die, so the fabs indeed have something to do.
It’s not all rosy with Clarkdale unfortunately. Remember the memory controller that Nehalem so graciously integrated? Clarkdale kicks it off die again. The IMC is housed in the 45nm GMA die. It’s still on-package, but not on-die. The benefit is super fast memory access for the graphics core, but slower memory access for the CPU core. In fact, this is a derivative of the memory controller used in older Intel chipsets (e.g. P45/X48).
The CPU connects to the GMA die using QPI, so bandwidth shouldn’t be as big of a problem. Latency is unfortunately hurt as a result. Access times can be longer than older LGA-775 processors thanks to this memory controller design being optimized for FSB architectures. Again, the problem goes away with Sandy Bridge, but today we’re faced with it.
Like Lynnfield, Clarkdale uses Intel’s LGA-1156 socket. Clarkdale should work in all P55 motherboards, but you’ll need a BIOS update. No existing P55 motherboards route video from the socket to a VGA/DVI/HDMI connector, so you’ll need a new motherboard if you want to take advantage of the on-package graphics. Enter the H55, H57 and Q57 chipsets.
A typical H57 I/O layout
The H and Q-series of chipsets feature what Intel calls its Flexible Display Interface (FDI). This is basically a link between the CPU and the chipset that passes along video output. H5x/Q57 motherboards will have a video out on board so you can use Clarkdale’s integrated graphics.
The chipsets differ in price and features. The table below sums it up:
Click to Enlarge
Support for Intel’s Rapid Storage Technology is probably the main reason you’ll want H57 over H55. The difference between H57 and Q57 boils down to security and management features. The H-series is for consumers, the Q-series is for corporate customers. Make sense?
The chips themselves are pretty straightforward. Intel is launching seven desktop Clarkdale processors (and a whole lot more notebook Arrandale chips):
| Processor | Core Clock | Cores / Threads | L3 Cache | Max Turbo | TDP | Price |
| Intel Core i5-670 | 3.46GHz | 2 / 4 | 4MB | 3.76GHz | 73W | $284 |
| Intel Core i5-661 | 3.33GHz | 2 / 4 | 4MB | 3.60GHz | 87W | $196 |
| Intel Core i5-660 | 3.33GHz | 2 / 4 | 4MB | 3.60GHz | 73W | $196 |
| Intel Core i5-650 | 3.20GHz | 2 / 4 | 4MB | 3.46GHz | 73W | $176 |
| Intel Core i3-540 | 3.06GHz | 2 / 4 | 4MB | N/A | 73W | $133 |
| Intel Core i3-530 | 2.93GHz | 2 / 4 | 4MB | N/A | 73W | $113 |
| Intel Pentium G9650 | 2.80GHz | 2 / 2 | 3MB | N/A | 73W | $87 |
The six processors labeled Core i5s and Core i3s all have the same basic architecture. You get per-core 256KB private L2s and you get a 4MB L3 cache shared among both cores (down from 8MB from the quad-core Lynnfield/Bloomfield chips). The i5s get turbo mode while the i3s do not. The i5-661 uses a higher graphics clock and has a higher TDP than the 660. Remember that these are CPU+GPU combos on a single package, so GPU clocks do vary based on model.
The Clarkdale lineup is honestly made up of CPUs that are too expensive. The Core i5 670, 661/660 and 650 are all priced above $170 and aren’t worth the money. The problem is Lynnfield’s turbo mode gives you high enough clock speeds with two threads that there’s no need to consider a dual-core processor. You can buy a Core i5 750, have more cores than any of these Clarkdales and run at close enough to the same frequencies for $196. Or you can buy a Phenom II X4 965 for about the same price and have much better multi-threaded performance. The chips that are most interesting are the Core i3s.
| Processor | Clock Speed | Max Turbo | |
| 2 Cores Active | 1 Core Active | ||
| Intel Core i5-670 | 3.46GHz | 3.60GHz +3.9% | 3.73GHz +7.7% |
| Intel Core i5-661 | 3.33GHz | 3.46GHz +4.0% | 3.60GHz +8.0% |
| Intel Core i5-660 | 3.33GHz | 3.46GHz +4.0% | 3.60GHz +8.0% |
| Intel Core i5-650 | 3.20GHz | 3.33GHz +4.2% | 3.46GHz +8.3% |
| Intel Core i3-540 | 3.06GHz | N/A | |
| Intel Core i3-530 | 2.93GHz | N/A | |
Turbo just isn't as interesting with only two cores. With four cores you used to have to make a tradeoff between good 4 and 2 threaded performance, but Lynnfield fixed that. No one really debates single vs. dual core anymore. The single core turbo modes are great, but aren't worth the money. Pay attention to the i3s.