Intel Reveals More Larrabee Architecture Details 123
Ninjakicks writes "Intel is presenting a paper at the SIGGRAPH 2008 industry conference in
Los Angeles on Aug. 12 that describes features and capabilities of its
first-ever forthcoming many-core architecture, codenamed Larrabee.
Details unveiled in the SIGGRAPH paper include a new approach to the
software rendering 3-D pipeline, a many-core programming model and
performance analysis for several applications. Initial product
implementations of the Larrabee architecture will target discrete graphics
applications, support DirectX and OpenGL, and run existing games and programs.
Additionally, a broad potential range of highly parallel applications including
scientific and engineering software will benefit from the Larrabee native C/C++
programming model."
Good old SIGGRAPH (Score:5, Insightful)
With the supposed death of Usenet, the closing of PARC, and the general Facebookification of the Internet, its nice to see a bunch of nerds get together and geek out simply for the sake of it.
Re:Good old SIGGRAPH (Score:5, Informative)
Re: (Score:3, Interesting)
Other areas of CS have multiple conferences throughout the year. Graphics has only one, and that's SIGGRAPH. If your paper is not accepted at SIGGRAPH, you are considered to have done nothing worthwhile that year. You could win every special effects award in Hollywood, but no SIGGRAPH paper = no cred.
That's just how it works.
Re: (Score:3, Informative)
Re: (Score:2)
Eh? [parc.com]
Re: (Score:1)
Trying to fight the trend toward specialization? (Score:4, Insightful)
Re:Trying to fight the trend toward specialization (Score:4, Informative)
It almost certainly won't work. In the past, there has been a swing between general and special purpose hardware. General purpose is cheaper, special purpose is faster. When general purpose catches up with 'fast enough' then the special purpose dies. The difference now is that 'cheap' doesn't just mean 'low cost' it also means 'low power consumption,' and special-purpose hardware is always lower power than general-purpose hardware used for the same purpose (and can be turned off completely when not in use).
If you look at something like TI's ARM cores, they have a fairly simple CPU and a whole load of specialist DSPs and DSP-like parts that can be turned on and off independently.
Re:Trying to fight the trend toward specialization (Score:5, Interesting)
It almost certainly won't work. In the past, there has been a swing between general and special purpose hardware.
Except with unified shaders and earlier variations the GPU isn't that "special purpose" anymore. It's basicly an array of very small processors that individually are fairly general. Sure, they won't be CPUs, but I wouldn't be surprised if Intel could specialize their CPUs and make them into a competitive GPU. At the very least, good enough to eat a serious chunk upwards in the graphics market, as they're already big on integrated graphics.
Intel's integrated graphics sells ATI and Nvidia. (Score:3, Insightful)
Basically, once you discover what Intel graphics has not been able to do, you buy an ATI or Nvidia graphics card.
Not everybody plays 3D games (Score:2)
More people in the world need Intel level graphics than need ATI/NVIDIA. This is borne out in sales numbers - Intel is the #1 graphics chip maker and has been so for many years.
Intel graphics has been TERRIBLE. (Score:2)
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
We felt so abused by the previous chipsets... (Score:2)
The last video driver we tested was version 14311 for the 945 chipset. It had a LOT of problems. There was a LOT of denial by Intel [intel.com] that there were problems.
So, I would be very interested to know: Is the video in the 965 chipset better? Is the software trouble-free? How about
Re: (Score:1, Informative)
Nobody is arguing that Intel makes good graphics hardware. They make adequate graphics hardware that the majority use without problems.
Go to any non-gaming office building and tell me how many Intel graphics vs Nvidia and ATI you find. I am willing to bet that most, if not all, of them will be Intel.
Re:Intel's integrated graphics sells ATI and Nvidi (Score:3)
And once you discover what kind of driver support they offer, you go right back to Intel.
The new Intel G45 chipset recently made me order a new motherboard just to replace my video card. It's "fast enough", one might say...
Personally, I can't wait to get all that proprietary crap out of my kernel. Shouldn't have fallen for the temptation in the first place.
Re: (Score:2, Interesting)
Except the part where GPUs have 256-512 bit wide, 2GHz + dedicated memory interfaces and Intel processors are...way, way less. Add that to the ability to write tight code on a GPU that efficiently uses caching and doesn't waste a cycle, compared to the near impossibility of writing such code on the host processor which you share with an OS and other apps... meh.
There might be some good stuff that can be done with this architecture, but I am not convinced it's a competitor to GPUs pound for pound. You have t
Re: (Score:2)
Even with all the advances in shaders GPU's are not quite generalized due to several reasons. Hardcoded data fetch logic (yes there is some support for more arbitrary memory reados but those are limited and take a fairly big performance hit). GPUs also have poor performance for dynamic branching -- sure they support it, but
Re: (Score:2)
General purpose is cheaper, special purpose is faster.
Only sort of. Special purpose is often cheaper, hence the profusion of ASICS. General purpose is more flexible, and so more desirable as a result. Also, special purpose is only cheaper if "general purpose" isn't quite up to the task. Speical purpose is also only cheaper if you're doing it all the time.
For instance, on the low end, MP3 players often have (had?) MP3 decoder ASICS, because it was too expensive to perform on the very small CPU. On a PC, the
Re: (Score:2)
See: x87 FPUs, cryptographic accelerators, video decoding, GPUs, etc.
Re:Trying to fight the trend toward specialization (Score:5, Interesting)
I don't think so. I think the fact is that with the right architecture (which Intel is trying to get into place) which exact core on which processor handles a specific task should become less and less relevant.
What this technology will hopefully provide will be the ability to have a more flexible machine which can task cores for graphics, then re-task them for other needs as they come up. Your serious gamers and rendering heads will still have high end graphics cards, but this would allow more flexibility for the "generic" business build PC's.
Re: (Score:2)
What'll be more interesting is if it fragments the PC market.
If you want a super-fast ray-tr, erm, protein folding application you need one with the Larrabee chipset. If you want to play the latest game you'll need a traditional PC + graphics card. Would it be possible that business PCs turn to Larrabee and home PCs stick with current architectures?
Ray Tracing -v- Rasterization (Score:1)
Neither the summary nor TFA itself mentions the words "Ray Tracing" or "Rasterization" [slashdot.org].
Am I missing something here?
Re:Ray Tracing -v- Rasterization (Score:4, Informative)
No, because the article is about Intel explaining that the purpose of Larrabee is NOT to be specialised like that. It's meant to be a completely programmable architecture that you can use for rasterization, ray tracing, folding, superPi or whatever else you want to program onto it.
Basically, they're trying to say "it's not REALLY a GPU as such, it's actually a really fat, very parallel processor. But you can use it as a GPU if you really want to".
I beg to differ. (Score:2)
The biggest debate in all of graphics-dom [graphixery?] for the last six months or a year has been Ray Tracing -vs- Rasterization.
So what happened?
I just don't understand how you can have an article about next-generation GPU tech and not ask whether the logic gates & data busses are going to be optimized for Ray Tracing or for Rasterization or for both [which would require at least twice the silicon, if not twice the wattage and twice the heat dispensation].
Has Intel completely abandoned the idea
Re:I beg to differ. (Score:4, Insightful)
Oh. (Score:2)
My bad - when something is this irrational, I guess the first suspicion should be politics - instead, I had simply assumed incompetence [or insouciance or absence of inquisitiveness] on the part of the author.
I will work to up my cynicism.
Re: (Score:1)
Another likely reason for acceptance is the last author listed, Pat Hanrahan, a giant in computer graphics (and a truly nice guy). Of course, the review process is supposedly scrupulously blind (the referees don't see the authors' names), but once th
Re: (Score:1)
That debate has for the most part been confined to speculation in various media outlets. It is obvious to everyone in the graphics community that Larrabee will be optimized for rasterization. It would not be possible to sell a graphics card that did not get comparable performance to ATI/NVIDIA in today's games, which means DirectX/OpenGL, which means rasterization. Intel wants to sell Larrabee, so it will have to rasterize extremely well. If Intel has to make a decision between optimizing a piece for ra
OpenGL (Score:1)
I get a warm-fuzzy feeling seeing that OpenGL isn't dead. I was first and best impressed with it when I played NeverWinter Nights, why hasn't it caught on more? Why don't more Open Source Games use it (as opposed to reusing the Quake engine)?
Re:OpenGL (Score:5, Informative)
Re:OpenGL (Score:5, Interesting)
Isn't the point of Larabee to change that? With umpteen Pentium-compatible cores, each one beefed up with vector processing instructions, software rendering might become fashionable again.
Re: (Score:2, Informative)
You still need an API - which OpenGL provides. On the hardware side of things, few chips actually implement the (idealized) state machine that OpenGL specifies, it's always a driver in between that translates the OpenGL model to the chip model.
Re:OpenGL (Score:4, Informative)
OpenGL is just an abstraction layer. Mesa implements OpenGL entirely in software. Implementing it 'in hardware' doesn't really mean 'in hardware' either, it means implementing it in software for a coprocessor that has an instruction set better suited to graphical operations than the host machine.
Sure, you could write your own rasteriser for Larrabee, but it wouldn't make sense to do so. If you use an off-the-shelf one then a lot more people are likely to be working on optimising it. And if you're implementing an off-the-shelf rasteriser, then implementing an open specification like OpenGL for the API makes more sense than making everyone learn a new one, and means that there's already a load of code out there that can make use of it.
Re: (Score:1)
Re: (Score:2)
If you are going to replace the GPU for rendering, then you are going to have to replace the Z-buffer (for depth-testing), triangle rasterisation with hardware texture mapping with fragment shaders, hardware shadow-mapping, and vertex shaders (for character animation). All of this should also work with multi-sampling at HDTV resolutions (2048 x 1500+ pixels). Someone made the observation that the average pixel would be rendered at least seven times, so that would have to taken into account. The only alterna
Re: (Score:2)
I was just promoted by what the other poster mentioned about Quake. The old Quake rendering engine was optimized like crazy for Pentium processors. Larabee is going to be a phalanx of them (well, not quite the original Pentium, something a bit better). The Quake code might run on it rather well, with each CPU rendering a small tile of the display. Of course, Quake's visual effects are hardly state-of-the art nowadays, but it would be an interesting hack.
Re: (Score:2)
I certainly agree. The assembly language used for texture mapping actually took advantage of the separate add/multiply processor units so that the divide-by-z for perspective projection took no extra clock cycles. The geometry for the environment was designed to fit into cache pages. Combined with light-mapping and a complex environment, it was an amazing experience to see.
Re: (Score:2)
There's a difference between the quake engine and OpenGL. OpenGL is just a graphics library, it pretty much just outputs primitives.
The Quake engine manages meshes, does collision detection, handles all the mess of drawing the right textures for the right models, managing lighting etc.
If there were an OSI model for graphics, OpenGL would be layer 4, and the Quake Engine would be layer 5/6.
Larrabee is a GPU (Score:1)
Re: (Score:1)
Well it's really a little of both. The only piece of hardware on Larrabee which is *definitely* GPU-specific is the fixed-function texture sampling hardware (which implements trilinear sampling, anisotropic filtering, and texture decompression). The rest of it is basically just a multicore CPU, where each CPU has something like an expanded MMX/SSE unit (4x the size, with some new and interesting features), and all the 3D stuff (shaders, z-buffering, clipping, etc) is implemented in software on top of that
Believe It When I See It (Score:2, Insightful)
Bearing in mind all the other promises Intel has made about their previous graphics offerings, I'm rather inclined to think that once again this will underwhelm. Especially considering all the crap that's been coming out of Intel about real-time raytracing. (It's always been just around the corner because rasterisation always gets faster.)
That's not to say that it's an interesting bit of tech, but from what I've seen so far it looks like the x86 version of Cell. Of course though it's a PC part and won't be
Re:Believe It When I See It (Score:5, Insightful)
There's a mass high end market (Score:2)
It isn't nearly as big as a more economical (both in terms of power usage and cost) market, but it is still large. ATi and nVidia are not having problems moving their new high end accelerators that draw hundreds of watts. On the contrary, there were some low stocks of the new GeForces after the prices were dropped.
Re: (Score:2)
Uh? Most people don't even know what is the power consumption of their desktop!
And most of those who care, care only because power --> heat --> fan noise and big case.
So I don't think that 300W is an issue (with an heatpipe to avoid fan noise), the software for this kind of new architecture will be the biggest issue: GPUs have evolved 'progressively' providing new feature AND good performance for the existing games, this will be much more difficult for Larrabee.
Not at all like Cell BE (Score:2, Informative)
> so far it looks like the x86 version of Cell
Then you missed the fact that the article says it uses a coherent 2-level cache for inter-core communications; the Cell BE is quite exotic in that it uses DMA transfers and has no memory coherency between the SPEs.
The article doesn't explicitly state that the Larrabee cores are homogeneous, but I would be surprised if they weren't; the Cell cores are somewhat heterogeneous if you want to use the PowerPC core to squeeze the last drop of processing power out of
Re: (Score:2)
Thanks, you just made me have a look at Almasi and Gottlieb (1994) again, where else would I need it than on slashdot. Chapter 10 Section 3 and following is good to have a look at.
Your last statement leaves me puzzled. It seems that you must have something to connect caches and memory with so we could be content with a simple bus which A&G describes as existing in Sequents cost effective machines.
When you read further you come across the KSR1 which uses a hierarchical ring architecture, sporting large c
software mode the return? (Score:1)
Why not just do a software mode driver for em?
that probably would make the 3D gaming market a bit bigger without forcing the people to buy a 3D acelerator card (thing that is kinda impossible to do on most laptops)
Re: (Score:2)
Who forced you to buy a more 3D oriented graphics-card?
These days you can pick your system oriented to your usage, if you want to play alot of games, do alot of encoding or work alot with media, you'll get a more advanced graphicscard and are willing to make a bigger investment in that. If you don't, you're perfectly fine with an integrated graphicscard. The choice is there, and it's to be made by yo
Re: (Score:1)
There are software drivers, but I suspect they'd tend to be far slower than even cheap graphics hardware.
In fact one of the reasons why Intel integrated graphics are so slow is because they do some things (like vertex shaders) in software.
Re: (Score:1)
Just imagine... (Score:2)
Re: (Score:1)
Is this a new slashdot meme? Never finishing the jokes?
Ok well, there was an AI bot running on a larrabee, an Irish Priest and a Soviet Russian who walked into a bar. The Irish priest orders a scotch. Then suddenly...
Re: (Score:2)
There, fixed that for you.
Yes, but what is the extent of Larrabee? (Score:5, Interesting)
Intel's introduction of "Larrabee" is an example. Where will it be used? Only in high-end gaming computers and graphics workstations? Will Larrabee provide video adapters for mid-range business desktop computers?
I'm not the only one who thinks Intel has done a terrible job communicating about Larrabee. See the ArsTechnica article, Clearing up the confusion over Intel's Larrabee [arstechnica.com]. Quote: "When Intel's Pat Gelsinger finally acknowledged the existence of Larrabee at last week's IDF, he didn't exactly clear up very much about the project. In fact, some of his comments left close Larrabee-watchers more confused about the scope and nature of the project than ever before."
The Wikipedia entry about Larrabee [wikipedia.org] is somewhat helpful. But I don't see anything which would help me understand the cost of the low-end Larrabee projects.
Re: (Score:1, Insightful)
As soon as it actually exists somewhere other than Intel's laboratories, they're usually pretty forthcoming on details (to the point we even have specs on how to use their graphics hardware, which is more than we can say for e.g. nVidia.)
OTOH, Larrabee is still Labware, and should be thought of as such. Unless you're willing to sign away your life in NDAs, don't expect to know too much yet.
Re: (Score:2)
According to the wikipedia article, the launch is currently planned for late 2009 or 2010. So it's a good bet that Intel won't have a good idea about what products they will be releasing for another year or so.
There have been many mistakes. (Score:2)
Intel Marketing: ArsTechnica art. about confusion? (Score:2)
One, Two, Four, Many? (Score:2)
This only goes to show that the people at Intel really can't count..
(Firmly tongue in cheeck, of course :)
Re: (Score:1)
Changing the playground... (Score:5, Interesting)
What most people don't seem to realize is that Larabee is not about winning the 3d performance crown. Rather, it is an attempt to change the playground: you aren't buying a 3d card for games. You are buying a "PC accelerator" that can do physics, video, 3d sound, dolby decoding/encoding etc. Instead of just having SSE/MMX on chip, you now get a complete separate chip. AMD and NVIDIA already try to do this with their respective efforts (CUDA etc), but Larabee will be much more programmable and will really pwn for massively parallel tasks. Furthermore, you can plug in as many Larabees as you want, no need for SLI/crossfire. You just add cores/chip like we now add memory.
P.
Re: (Score:2)
The need for SLI/crossfire is because the bandwith needed for multiple cards to work on a frame buffer is too high for even the newest PC memory bus.
Intel's cards are not going to be able to get around this, so we will most likely add a third method of card interconnect to the mess.
Re: (Score:1)
I beg to differ. Nobody will buy Larrabee unless there's software for it, but nobody will write the software until there's a large market of Larrabees out there to run it on. Luckily the entire library of PC games can be made to run on Larrabee via DirectX/OpenGL, and that's Intel's way around this problem. However, in order for this to work the price/performance has to be competitive. Intel's best bet is to use their superior fabrication facilities t
What does this do (for graphics) better than GPU? (Score:2)
Re: (Score:1)
It's way too early to say whether it'll even be equivalent performance-wise to AMD and NVIDIA's GPU designs in Larrabee's release time frame, and it'll be very dependant on its compiler and drivers, but as a concept right now it's hugely interesting i
Re: (Score:1)
Lets say we want to use the numerous deformation techniques that do not work by transforming normals via a matrix (i.e. they must re-compute the normals because the deformation is non linear - FFD's for exam
Potentially, but... (Score:1)
Seriously, manymanycores architectures are nice for public servers that are coded very well. Potentially able to serve N clients at once, the machines running Larrabees will usually bottleneck somewhere else.
For the desktop user, manymanycores mean that the main window will move smoothly in the foreground while anti-blackware, indexes and updates consume the background.
For the power gamer, even manymanycores won't be enough. There's no such thing as
But can intel make good drivers as there on board (Score:2)
But can intel make good drivers as there on board ones suck?
There on board video cards look good on paper but then come in dead last next to nvidia and ati on board video and that is with out use side port ram. ATI new board video can use side port ram.
Gettin' a kick . . . (Score:2)
Re: (Score:2)
This is remarkable, Larrabee is a fairly low resolution area, what are you guys trying to hide?
Also notice the rectangular street patterns, uncanny how it resembles a chip layout.
So are you renaming one of those streets into memory lane, the parking lot in front of the bank into Cache Plaza maybe ...
Anandtech has an excellent article (Score:4, Informative)
That is much more detailed than the one linked in the article summary. It can be found here. [anandtech.com]
Re: (Score:2)
Wow, I always thought Anand was a bit of an Intel fanboy, but does he ever gush of Intel in that one.
Some fun out of context quotes:
Modern 486 (Score:2)
Re: (Score:2)
Re: (Score:2)
What I would like to know is how high you might clock a 486 core built in a modern fab.
Re: (Score:2)
Re: (Score:2)
Yeah, that was sort of what I was expecting. My understanding is (IANAChipDesigner, etc) that to reach high clock rates, you'd need to have deeper pipelines than on the 486. The reason being that each step can then be considered an independent unit with regards to clock propagation.
Re:Good news (Score:5, Insightful)
This is good news for Mac mini and MacBook users.
How so? Has Apple announced that it will adopt Larrabee for the Mac Mini or the MacBook? No. All you have are rumors and speculation by MacRumors and Ars Technica. When Apple says they will adopt the Larrabee GPU, then you can say that it is good news for Mac users of any stripe. Until then, it's just Intel news, not Apple news.
Re: (Score:2)
I think it depends on how much Larrabee will cost, however with what we know so far Apple seems to be heading into multi-CPU architectures, so using Larrabee would make sense.
Re:Good news (Score:5, Informative)
I think it depends on how much Larrabee will cost, however with what we know so far Apple seems to be heading into multi-CPU architectures, so using Larrabee would make sense.
Larrabee costs somewhere between 150 and 300 Watt, so MacBooks and Mac Minis are not likely to use them. Mac Pro, on the other hand, possibly.
Re: (Score:1, Insightful)
Re: (Score:1, Insightful)
Re: (Score:3, Interesting)
The power brick for my Core 2 Duo Mac mini is somewhere around 80 Watts I think. And I'd assume the actual usage is lower than that. Let's say 50~60 Watts for the whole computer (CPU, GPU, hard drive, optical drive, RAM, FireWire, USB, etc).
If Larrabee takes 150~300 Watts, then it's just insane, no matter how many cores it has.
Re: (Score:2)
Mac minis integrate 2.5 inch hard disk drives (ATA in the G4 models and SATA in the Intel models), CPUs and other components originally intended for mobile devices, such as laptops, contrary to regular desktop computers which use lower cost, but less compact and power-saving components. These mobile components help lower power consumption: According to data on the Apple web site, first-generation PowerPC Mac minis consume 32 to 85 Watts, while later Intel Core machines consume 23 to 110 Watts. By comparison, a contemporary Mac Pro with quad-core 2.66 GHz processors consumes 171 to 250 Watts.
Re: (Score:3, Informative)
They've stated that it will be a 150W+ chip on a PCI Express 2 card, as I recall, and is intended as a GPU, though it will be fully programmable and have CPU capability (so when not doing GPU stuff, it could serve as extra CPUs). It is intended to compete in the high end graphics market.
Essentially, it's a clutch of high performance software vector units in parallel with a bunch of CPUs. Graphics scale with each added processor because it is a software driven architecture, whereas traditional GPUs don't s
error (Score:2)
oops - I meant limited ray tracing (not casting). That or painter's algorithm (which optimally uses a sorted list). I never did fully understand ray casting - I jumped from painter's to hardware.
Re:Good news (Score:4, Funny)
Re: (Score:3, Insightful)
Re: (Score:2)
Re: (Score:2)
Clearly Intel means C/C++ with pthreads or Win32 threads. Also, C++ is getting thread support in C++0x.
Re: (Score:1)
This is an important point. Programming for GPUs has gotten easier now that nVidia's CUDA and AMD/ATI's CTM are around. But "easier" is not the same as "easy".
Before CUDA and CTM, if you wanted to write a GPU program, you had to pretend to do graphics. This meant setting up a whole pile of OpenGL or Direct3D nonsense, and then drawing a rectangle to trigger the pixel processor on each data item. At least that nonsense is out the
Re: (Score:2, Informative)
In addition, Fortran to C tools have been around for some years. To say that Fortran is the only scientific language is BS. R [r-project.org], S Plus [insightful.com], Octave [wikipedia.org], matlab [mathworks.com], perl and CUDA [nvidia.com] to name a few. Taking R as an example - it provides an code interface that allows you to write optimised C/C++ routines and utilise those in the language itself.
Grandpa, you're so 1977... (Score:2)
It's true that by the time C became popular, there already existed many libraries in Fortran that had been tested and debugged, so no new development on those lines was needed. But most new scientific software is being written in C. There are a few diehards who will never admit it, but for all practical purposes Fortran is dead as a development language.
Re: (Score:1)
Re: (Score:2)
More good analysis here:
http://www.pcper.com/article.php?aid=602 [pcper.com]