Slashdot Mirror
Intel Kills Consumer Larrabee Plans
An anonymous reader tips news that Intel has canceled plans for a consumer version of their long-awaited and oft-delayed Larrabee chip, opting instead to use it as a development platform product. From VentureBeat:
"'Larrabee silicon and software development are behind where we had hoped to be at this point in the project,' said Nick Knuppfler, a spokesman for Intel in Santa Clara, Calif. 'Larrabee will not be a consumer product.' In other words, it’s not entirely dead. It’s mostly dead. Instead of launching the chip in the consumer market, it will make it available as a software development platform for both internal and external developers. Those developers can use it to develop software that can run in high-performance computers. But Knuppfler said that Intel will continue to work on stand-alone graphics chip designs. He said the company would have more to say about that in 2010."
So they intend to take a product, who's chief advantage was that it could run old x86 code, and only sell it people who are designing new software? Am I the only one that sees a problem with this?
A nicer way of saying:
Uhm, guys, remember how we were supposed to ship a year ago and said recently we will ship a year from now? Well, add 5 to that now...but we will provide and totally kick ass, promise.
One that hath name thou can not otter
In case you've forgotten what a Larrabee was (like I had), it was Intel's planned graphics / vector processing chip, competing with nVidia and AMD / ATI graphics systems. Here's the Wikipedia article.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
Hmm... I think Intel's plan is for Larrabee GPU's to launch at the same time as Duke Nukem Forever! :)
I would say ATI AMD are about to become the leader. Intel is making it more difficult to ship mobile systems without the craptastic intel graphics cards. Larrabee was supposed to be a decent performance GPU, that would almost be like a co-processor.
AMD has slightly slower CPU's, but their intgerated graphics blow the snot out of the Intel ones, and are getting even better.. What good is a super fast CPU, if you can't play any games, or even do basic stuff without using the power hungry CPU?
What are we going to do tonight Brain?
http://www.brightsideofnews.com/news/2009/12/2/intel-larrabee-finally-hits-1tflops---27x-faster-than-nvidia-gt200!.aspx
way faster than amd's or nvid's hottest....
I spent most of internship in intel arguing with people hyping larabee as the 2nd coming of jesus that it would never happen... And now i can finally say HAH!
NVidia hasn't let ATI do anything. Actually, NVidia is dealing with a series of problems - from serious packaging problems last year to TSMC yield issues now. ATI/AMD has been really effective lately; NVidia historically had a dominant position, but definitely not a monopoly, and I'll say that they have slipped a lot recently. Things change fast in the GPU race, so NVidia may recover quickly. But ATI/AMD have a solid amount of momentum, and the only real execution problem I've seen them make in the last few months in GPUs has been to rely on TSMC.
Take a look at the Dell Zino HD - it combines AMD's 'just enough CPU' with top end GPU to make a very compelling system. Intel has cut NVidia out of the chipsets, so they don't get the synergy that AMD has with ATI.
AMD is definitely better situated for the long haul than NVidia, and actually may be better off than Intel for complete systems.
No, I don't trust in god. He'll have to pay up front, like everybody else.
last i checked their flash disks were pretty kickass
turn up the jukebox and tell me a lie
This is being mis-reported or mis-communicated by Intel, I believe.
The first version of Larrabee silicon isn't going to consumers, that's all.
From the consumer's perspective, it's a delay. Yet to be seen if it's fatal.
Otherwise, who'd want to use it to develop software?
I think the announcement of the 48-core Intel 'Bangalore' chip just recently is not a coincidence.
When I first read about the Larrabee chip, I thought the decision to make it a cache coherent SMP chip to be simply insane - architectures like that are very difficult to scale, as the inter-core chatter scales roughly as the factorial of the number of cores. Remember how Larrabee was designed around a really wide 1024-bit ring bus? I bet that's required because otherwise the cores would spend all of their time trying to synchronize between each other.
So, Larrabee is effectively cancelled, but only a day or two before Intel announced an almost identical sounding part without cache-coherence! It sounds to me like they've given up on the 100% x86 compatibility, and realised that a chip with some extra instructions around explicit software controlled memory synchronization and message passing would scale way better. Without cache coherence, a "many core" chip is basically just an independent unit repeated over and over, so scalability should be almost infinite, and wouldn't require design changes for different sizes. That sounds like a much better match for a graphics processor.
While Intel kept their cards relatively close to their chest, from all of the presentations I've seen, no first-gen Larrabee chip could scale beyond 24 cores even with a 1024 bit bus, while the new Bangalore chip starts at 48 cores. There's no public info on how many lanes Bangalore has in its on-chip bus but based on the bandwidth of its 80 core experimental predecessor, I'm guessing it's either 32-bit or 64-bit (per core).
I don't know about that. Intel's offerings that are slated to come out 1Q - 1H of 2010 could give AMD some problems. Right now AMD has the performance advantage in the server space, but Gulftown will likely trump their offerings. Arrandale also looks quite impressive, especially the quad core i7 with an 18 watt TDP. The cores only run at 1.2 GHz, but with their Turbo boost the chip can clock up to 2.2 GHz. That will offer some amazing battery life for laptops and still provide good performance. I do believe some of the Arrandale processors will have a GPU on die as well. Granted it's an Intel GPU, but it offers some great power and cost savings over having to include a discrete card.
AMD doesn't look to have anything great coming out until late 2010 or early 2011 based on their roadmap. It helps that ATI is kicking ass in the graphics space. Right now they're winning on price and power. If they can get more of their 5800 series out in the market and release the mobile versions of those cards sooner rather than later, they'll be able to push a lot of hardware that way. However, they're not a real threat to Intel until they can get their SOC products out the door and offer a really compelling reason to go with their products.
Settling their legal issues with Intel will also help them a lot in the long run, but they're not out of the woods yet. They're still having financial problems, but if they can get through the next 18 months they'll be in great shape. The fact that they've been ahead of schedule on a lot of their new chips in the last year has probably helped substantially as well. AMD is in good position for the long term, but they need to decent sales in the coming quarters, which may be difficult to do with Intel releasing a lot of great new chips, especially in the mobile market where AMD hasn't been particularly strong recently.
Wow... thanks for your insight! Should have known Intel would be logical even about their failures, and roll them over to something that has a chance of applicability. The only thing I wish they would do is skip the 64-bit crap and make 128-bit architectures that are compatible with both 32- and 64-bit predecessors. It would ease the development of new applications since the life time of 128-bit archs would be decades as opposed to developing all 64-bit apps to only have 128-bit archs appear in 5-10 years.
I'm not sure if you're trolling or not, but 64-bit memory capacity is not "twice" as big as 32-bit, it's 4.3 billion times as big. That's more than just 5 to 10 years of Moore's law, that's more like 50 years. Physical bus widths have nothing to do with architecture bitness either, there are memory buses for 64-bit architectures that only have a few pins.
I have a 4 year CS degree and I can tell you with certainty that that blogger is full of shit. The problems that are already parallelizable, are easily multithreaded with current technology. The problems with serial dependency, are not, and never will be, easily multithreaded.
Rendering graphics is already done, because its easy to split the task of rendering a bunch of pixels into pixel-sized chunks. Each small thread can read from the same shared memory (the scene graph and textures, etc.) and write to a distinct location (its one pixel in the frame buffer).
Encoding video using motion-compensation techniques (basically all modern video codecs) will never be satisfactorily parallelizable because the best bang/bitrate can only be achieved when frames are processed serially. Frames need to be processed as a whole to optimize for panning and other full-scene motion, and the results of the previous frame's motion analysis is typically needed to compute the next delta. You can break the processing up into multiple threads easily enough, but you miss out on opportunities to make the output more efficient or better looking.
When Mr. PseudoScience blogger can parallelize the video encoding problem without so many dependencies that its essentially a serial process, then he should get some credit, not before then.
I might agree with you if ATI/AMD would finally get serious about producing drivers that aren't complete crap. Their hardware is fine, but Linux drivers, as well as OpenGL drivers on Windows just plain suck.
It's not just the video drivers. ATI also has a horrible software stack (SDK, runtime, compiler and documentation) for their Stream GPGPU computing architecture, which is why everybody uses NVIDIA and its excellent CUDA. Generally speaking, ATI has excellent hardware, but such hardware is useless if you don't have a matching software to exploit it.
"I'm never quite so stupid as when I'm being smart" (Linus van Pelt)
but at least they are dedicated graphics solutions
Actually, the 9400m is not. It uses system memory but does a much better job then Intel. It also acts as the memory controller and does system IO. The reason for the parent's comments is that all future Intel CPUs will have integrated memory controllers (like the i7 and i5) and an integrated GPU. Performance will suck but it will make for cheap systems. This will make it difficult for system builders to make a low end system with good graphic performance as the market for such systems will be small. The smaller market will reduce the quality/performance of available parts for those system builders - one of which is Apple.
The problem is, a many-core non cache-coherent x86-like system isn't particularly interesting. The big advantage of Larrabee was that you could treat it like a normal SMP system, including (presumably) running standard multithreaded C code on it. Once you have to deal with memory synchronization explicitly, Larrabee starts to look a lot more (from a programming standpoint) like Fermi, Cypress or whatever other Nvidia/ATI GPUs are out at the time.
There's nothing magic about x86/AMD64 in the HPC world. It's attractive because it is cheap and has good performance. Clusters can, have been, and still are built using POWER and other architectures.
But for "embarrassingly parallel" problems, which are the target application for these chips, cache coherence is often not necessary, and simply imposes a design burden. There are lots of problems where it's better to have 1000x the performance than 1/2 the developer time.
It may not even involve less development time: Others have pointed out that the Unix "fork" mechanism combined with "copy-on-write" at the memory page level would also work, and wouldn't require cache coherency. Similarly, any existing code designed for message-passing supercomputers would work out of the box, with only a recompile using a new library. Developers just have to start thinking in terms of "many processes" instead of "many threads".
I suspect that in the long term (decades), cache coherency will simply not scale, and most computers will use explicit message-passing internally, even at the single processor level. The transition has already started: most new servers are NUMA systems, where there's a concept of "near" and "far" memory visible to the software, and most of the real heavy lifting in PCs are done by the GPUs, some of which do not have complete cache coherency across all cores.
Don't forget about the NVidia Ion platforms. They also use a "just-enough" CPU in Intel's Atom, with higher end NVidia GPUs to run nicely integrated HD set-top boxes. Nice little platforms for MythTV frontends.