Slashdot Mirror

← Back to Stories (view on slashdot.org)

Nvidia Discloses Details On Next-Gen Fermi GPU

Posted by samzenpus on from the getting-the-skinny dept.

EconolineCrush writes "The Tech Report has published the first details describing the architecture behind Nvidia's upcoming Fermi GPU. More than just a graphics processor, Fermi incorporates many enhancements targeted specifically at general-purpose computing, such as better support for double-precision math, improved internal scheduling and switching, and more robust tools for developers. Plus, you know, more cores. Some questions about the chip remain unanswered, but it's not expected to arrive until later this year or early next."

137 of 175 comments (clear)

  1. But does it... by popo · · Score: 4, Interesting

    ... run Linux?

    --
    ------ The best brain training is now totally free : )
    1. Re:But does it... by Anonymous Coward · · Score: 1, Insightful

      More importantly, does it run physx in a machine that also has a non-nvidia gpu?

      Oh, wait. No, it doesn't.

    2. Re:But does it... by PopeRatzo · · Score: 1, Insightful

      More importantly, does it run physx in a machine that also has a non-nvidia gpu?

      You understand that these gpu's are made by nvidia, right? So how could they run something on a machine with a non-nvidia gpu if the gpu's the article refers to are made by nvidia/I.?

      What exactly were you trying to say? I'm not quite sure.

      --
      You are welcome on my lawn.
    3. Re:But does it... by skarhand · · Score: 4, Insightful

      You could have read the link... Theoretically, you could use an ATI card for graphics and a second Nvidia card just for the physx. Well, not anymore. Nvidia disabled that possibility in the driver. So people with older Nvidia cards who choose to upgrade to the newest radeon 5800 series will lose physx. That kind of business practices remind me of a certain company from Redmond...

    4. Re:But does it... by Anonymous Coward · · Score: 5, Interesting

      Some motherboards have more than one PCI Express slot. Some even come with GPUs built onto the motherboard. In either case, it is entirely conceivable that there may be a GPU present other than the one attached to the display. Then there's the Hydra 200 (look on Anandtech, I'm too lazy to find the link) - a chipset which evenly distributes processing power among multiple GPUs from any vendor to render a scene or do GPGPU computing.

      Nvidia just released new drivers which explicitly disable PhysX acceleration in the presence of a GPU from another manufacturer. For the above stated reasons, this is evil.

    5. Re:But does it... by PopeRatzo · · Score: 1

      Then there's the Hydra 200 (look on Anandtech, I'm too lazy to find the link) - a chipset which evenly distributes processing power among multiple GPUs from any vendor to render a scene

      Really? That's cool.

      Nvidia just released new drivers which explicitly disable PhysX acceleration in the presence of a GPU from another manufacturer.

      That sucks. Do you know if there are other hardware physics solutions on the market besides PhysX? I know of a couple of software solutions, but don't know of any other hardware ones.

      Thanks for explaining the OP for me, AC.

      --
      You are welcome on my lawn.
    6. Re:But does it... by PopeRatzo · · Score: 2, Insightful

      That kind of business practices remind me of a certain company from Redmond...

      Actually, I can think of at least one other major computer manufacturer who makes products that nerf other manufacturers' products. I think they're located in Cupertino.

      --
      You are welcome on my lawn.
    7. Re:But does it... by Dishevel · · Score: 5, Funny

      Now you have done it. Don't you know enough not to wake the fanboys. Now I am gonna have to hear how Jobs is a genius for not allowing people to do things that he believes unimportant with their hardware. That nobody actually needs any more control than Steve gives them. It's not that I agree with them. I just do not want to hear it. Please just let them have their fanboy dreams in peace. Please!

      --
      Why is it so hard to only have politicians for a few years, then have them go away?
    8. Re:But does it... by mwvdlee · · Score: 1

      IBM isn't located in Cupertino, right? Neither is Sun. Which are you talking about?

      (before some pedantic know-it-all kid starts acting up; yes, I do know which company he was talking about, as do all of us).

      --
      Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
    9. Re:But does it... by ioshhdflwuegfh · · Score: 2, Informative

      He is talking about Apple.

    10. Re:But does it... by noidentity · · Score: 1

      But does it run Linux?

      Yes, but only if your main display is connected to a genuine Nvidia graphics card.

    11. Re:But does it... by smoker2 · · Score: 2, Informative

      Wrong. They disabled PhysX when a non-nVidia graphics card is powering the display. The presence of another graphics card is not the driving issue.

    12. Re:But does it... by ciderVisor · · Score: 1

      Thanks for that. I didn't know they were based in Cupertino.

      --
      Squirrel!
    13. Re:But does it... by TJamieson · · Score: 1

      PhysX is the only one that had a true hardware solution (the P1 PhysX cards, previous to nV purchase). That said, AMD has been porting Havok iirc to OpenCL for use on their hardware for fancy things like cloth mesh.

      --
      For the last time, PIN Number and ATM Machine are redundancies!
    14. Re:But does it... by Narishma · · Score: 1

      Yeah it's too bad for the two people who have both an AMD and nVidia card on the same machine AND care about PhysX...

      --
      Mada mada dane.
    15. Re:But does it... by jgtg32a · · Score: 1

      If I knew a while ago that I could have bought an nvidia card to function as a dedicated Physx card I would have picked one up. There's no change to game play but it does add some rather nice effects.

    16. Re:But does it... by hattig · · Score: 1

      AMD is also pushing the Bullet physics API running via OpenCL as a non-proprietary way forward.

    17. Re:But does it... by TJamieson · · Score: 1

      Really? That is some good news as well. More solutions is better, esp if it helps smack nV around a bit and gets devs off CUDA+PhysX.

      --
      For the last time, PIN Number and ATM Machine are redundancies!
    18. Re:But does it... by parlancex · · Score: 1

      Are you retarded?

      I'm a GPU programmer who is intimately familiar with CUDA, and there are a plethora of extremely important TECHNICAL reasons for NVidia's decision. PhysX is advantageous to CPU based physics processing not only because it is COMPUTING the collision detection and kinematics on the GPU, but also because the physics data describing the matrix transforms can be shared between the physics and graphics context for a large speed boost.

      This requires PhysX to work intimately with the graphics card that is responsible for rendering using those transforms. ATI consistently releases products with poor GPGPU features and probably isn't much interested in documenting their architecture enough for NVidia to maintain and debug a version of PhysX which is capable of interacting with ATI cards in this regard.

      If you have an ATI card you can still use CPU processed PhysX with your ATI graphics card, which speaks directly to the fact that this is a TECHNICAL limitation. Jesus Christ, don't let that stop your holy war though.

  2. But it won't have an RPU... by argent · · Score: 1

    They could fit one of Philipp Slusallek's ray-trace processing units in the corner of the chip and never notice the cost in silicon.

    1. Re:But it won't have an RPU... by hattig · · Score: 1

      Fermi has around 750 GFLOPS (clock speeds weren't released as the product is still around three months away from retail) of *double precision* power. It'll do ray tracing just fine. Indeed real time ray tracing was demonstrated yesterday with Fermi. I definitely see some ray tracing engines getting some CUDA/OpenCL love in the very near future now that DP is so fast.

      It's also likely that the 5870 and 5850 could be faster than Fermi for traditional gaming graphics, as these don't need double precision, and the 5000 series has vast single precision power (2.7 TFLOPS for the 5870, 3 TFLOPS for the eventual 5890 that'll emerge when Fermi appears).

    2. Re:But it won't have an RPU... by argent · · Score: 1

      Power efficiency matters. The RPU had about the transistor budget of a Rage Pro.

    3. Re:But it won't have an RPU... by Bitmanhome · · Score: 2, Interesting

      FLOPS aren't directly comparable, as the ATI chips are arranged more like the Itanium, while nVidia looks more like a Core Duo. ATI has more raw power, but uses a smaller percentage of it.

      --
      Not that this wasn't entirely predictable.
    4. Re:But it won't have an RPU... by hattig · · Score: 1

      That's why I said for "traditional gaming graphics" as the ATI arrangement is optimised for that application, i.e., that's probably where an ATI card gets closest to peak performance.

  3. AWESOME by Icegryphon · · Score: 1, Funny

    such as better support for double-precision math

    BEST NEWS EVAR!

    1. Re:AWESOME by ArchMageZeratuL · · Score: 5, Insightful

      To the best of my knowledge, double-precision floating point operations are actually pretty important for some scientific applications of GPUs, and as such this is significant for those using GPUs as supercomputers.

    2. Re:AWESOME by Korin43 · · Score: 2, Insightful

      It could also be useful in raytracing. The official reason POV-Ray hasn't been able to use video cards is that they don't have the required precision. That's probably pre-CUDA though, but "better support" sounds helpful.

    3. Re:AWESOME by jpmorgan · · Score: 1

      NVIDIA has had double precision support since the GT200 line. Performance was poor compared to single precision, but it was there.

    4. Re:AWESOME by AHuxley · · Score: 1

      Then you can dream of a big home made rendering device
      http://helmer2.sfe.se/

      --
      Domestic spying is now "Benign Information Gathering"
    5. Re:AWESOME by PC+and+Sony+Fanboy · · Score: 1

      better support

      Not only is it "better support", but it now has "Awesome Street Cred" and "Totally amazing user experience", thanks to a magical technology licensed from ________ (Pick your favorite scape-goat) famous for lock-in and proprietary design!

    6. Re:AWESOME by Penguinoflight · · Score: 1

      Do you mean the gtx280 line, or the re-badged (again) 8000 series?

      --
      "And we have seen and do testify that the Father sent the Son to be the Savior of the World"
      1 John 4:14
    7. Re:AWESOME by GreatBunzinni · · Score: 2, Insightful
      Actually, that's a fundamental aspect of GPGPU's migration from an interesting oddity to a serious option (if not the obvious choice) in the number crunching world. Just to give you an example, I'm a structural engineering major and, for my graduate thesis, I'm on the process of developing a pair of structural analysis programs (finite elements method applications), a type of problems which basically consist of solving considerably large linear equation systems. That sort of problem is right up GPGPU's alley. Yet, although it's a very affordable piece of technology and, as it was already demonstrated, would bring massive performance improvements to this sort of problem, after analysing the options it was found that, at least at this moment, it would be better to focus on relying on multiple-CPUs through multi-threading instead of jumping into the GPGPU's bandwagon. One of the main reasons that forced GPGPUs not to be seen as a serious option was, in fact, their underwhelming support for double-precision math.

      There were a hand-full of issues behind that decision. One of them was that some GPGPU platforms fail silently, which, in practice, means that you start crunching numbers with less than the expected mantissa and therefore you get considerably larger rounding errors,. This is something that may bring disastrous results. Another issue is that even in some cases the announced double-precision support of some products was a bit flawed, as it failed to comply with IEEE 754, the standard for floating-point arithmetic. Although it didn't complied due to only a hand-full of issues, to rely on GPGPUs to crunch numbers when they don't conform to that standard would mean that someone would be forced to spend a considerable time formally checking what effects that non-compliance would have on the project being developed. That means that that would take precious man-hours from projects which may already be poorly manned, not to mention that that task would be rendered to waste as the next GPGPU generation would either fully support with IEEE 754 or, in the worst case scenario, fail to support it in some other aspect, which would mean that the poor chap assigned to verify the effects of the product's non-compliance would be forced to do everything from scratch, once again.

      So, to sum things up, GPGPU's support for double-precision math is, in fact, great news. It means that everyone may have it's own personal vector-processing super-computer on his desktop. Heck, even on laptops. That may not mean much for the proverbial joe-sixpack (at least not beyond the "oohh... shiny graphics" side of things) but being able to crunch a lot more numbers on the same time frame means the world to anyone writing/using number-crunching software, which is a lot of people.

      --
      Slashdot, fix your code or at least hire someone who is competent at it to do it for you.
    8. Re:AWESOME by SlashWombat · · Score: 1

      It sounds like it might be very nice, but with Intel working towards a GPU integrated on the same chip as a multicore CPU, I wonder how much longer NVIDIA will remain mainstream. Remember, AMD has absorbed ATI, so both major CPU chip vendors are now obviously targeting the same (integrated) end product.Sure, it will initially be targeted at laptops, but I suspect more laptops are sold per annum now than desktop/server boxes combined!

    9. Re:AWESOME by jpmorgan · · Score: 4, Informative

      GTX 280 is a graphics card. The GT200 is the GPU core the GTX 280 card is based on. Likewise the 8800 series graphics cards were based on the G80 chip (and later G92, I think). There were also the G84, G86, G94 that power a number of nvidia's economy or mobile platforms. The Quadro 5600 and 4600 are also G80 based. There were other, cheaper Quadros based on the G84. The Quadro 5800 is based on the GT200 chip. The Tesla 870s were based on G80s, the 1070s are based on GT200. The cards also tend to have different memory interfaces (and amounts), clock rates and even firmware, which is why there are many different cards all based on the same handful of chips.

      So no, I do mean the GT200. The GT200 processor supports double-precision, the G8x and G9x processors do not.

    10. Re:AWESOME by smallfries · · Score: 1

      What kind of performance were you expecting from multi-cores? I'm not too familiar with your application so I was wondering if the code could hit 1 FLOP/cycle for 3-4 GFLOP/core or even 2 FLOP/cycle for 6-8 GFLOP/core? If the programming model for DP on these boards supports what you need then you might be looking at 100x the performance on GPGPU. But then again, if lack of support for saturating error conditions was holding you back before when there was a potential 10x increase performance, is 100x enough to overcome that hurdle? I'm not actually sure how you could provide saturating errors on a vector array anyway, as it implies a different code-path for each core in the vector.

      --
      Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
    11. Re:AWESOME by cheesybagel · · Score: 1

      Nehalem can achieve 51 DP GFLOPs. The NVIDIA GT200 can get 77 DP GFLOPs. This Fermi GPU, although a great advance, is available at a time Intel can manufacture CPUs with twice the cores and will double vector word length in their next processor family. So while GPUs have more GFLOPs, it is still not that much of a difference for people who use dual precision.

    12. Re:AWESOME by smallfries · · Score: 1

      I mixed up per-core and per-chip performance in original post, but assuming that the code can be scheduled to despatch 2 DP ops per cycle per core (reasonable if the code has a simple enough structure to run on a GPU), then you hit about 8 GFLOP/core. So yes, the new six-core Nehalems can hit about 50 GLOP/s. Some benchmarks put it slightly higher at about 55. Nvidia are claiming an 8x increase in DP throughput which puts them at about 616 GLOP/s for DP. It's still a ten-fold increase in performance.

      When you say that Intel can double cores and vector word length in their next family are you confusing Westmere (which is a tick generation) with it's successor (which will be a tock)?

      --
      Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
    13. Re:AWESOME by cheesybagel · · Score: 1

      When you say that Intel can double cores and vector word length in their next family are you confusing Westmere (which is a tick generation) with it's successor (which will be a tock)?

      I am talking about Sandy Bridge since that is the one with AVX. They have demoed systems with it already at IDF. It is supposed to do 256 DP GFLOPS. Intel is also supposedly going to implement Fused Multiply Add sometime after that which will double performance again.

    14. Re:AWESOME by hattig · · Score: 1

      The AMD HD5000 series is IEEE754-2008 compliant. Fermi will be similarly compliant when it arrives in three or four months. Both support FMA for both 64 and 32 bit operations - FMA has even more numerical accuracy than MAD because it is a fused operation with no intermediate storage/loss of accuracy.

      Fermi will have more DP power than HD5000, but less SP.

      Fermi also has some extra benefits - ECC throughout the memory hierarchy for example.

      The RWT article (linked from the end of the Tech Report article) is also a good read.

    15. Re:AWESOME by hattig · · Score: 1

      Well with Fermi starting at around 700 GFLOPS (double precision) in early 2010, and AMD's HD5890 probably achieving 600 GLFOPS, there's still a big benefit for using the $400 - $500 graphics card over three $1000 CPUs.

      Both ATI and NVIDIA's cards already support FMA for single and double precision calculations, and thus gain some numerical accuracy over the CPU that doesn't support it!

  4. Re:Honestly, at this point... by stocke2 · · Score: 1

    only for certain operations

    --
    A Smith & Wesson beats four aces -- Murphy's Law of Poker
  5. Re:Honestly, at this point... by whatajoke · · Score: 1

    What are you trying to say? And why did you get modded insightful for an inane comment?

  6. Re:Honestly, at this point... by Eddi3 · · Score: 1

    Completely OT, but your sig just made my day.

  7. This word "detail"... by Joce640k · · Score: 2, Funny

    ...I'm not sure it means what you think it means.

    --
    No sig today...
  8. Re:Honestly, at this point... by Shikaku · · Score: 3, Funny

    Yes.

  9. How did this get modded "insightful"? by Joce640k · · Score: 3, Interesting

    "Ignorant" would be a better rating - there's a lot of compute power but it's in the middle of a very different architecture to an x86 CPU. Not usable for running an OS.

    --
    No sig today...
    1. Re:How did this get modded "insightful"? by springbox · · Score: 1

      You accidentally the "ing."

    2. Re:How did this get modded "insightful"? by PC+and+Sony+Fanboy · · Score: 1

      but it's in the middle of a very different architecture to an x86 CPU

      OMG! It's not x86? It's USELESS!!!!!....

    3. Re:How did this get modded "insightful"? by Anonymous Coward · · Score: 1, Informative

      GP was foolish to assume people would know what hir was talking about but:

      GPUs are SIMD machines (Single Instruction Multiple Data), they process large quantities of numbers in parallel this is what makes them "fast" despite their low clock speed (compared to a CPU). They also have massive pipelines to decode as many instructions as possible simultaneously. All this makes them very powerful except for one major problem: branches. GPUs stall with major latency on branches.

      If you can write general purpose software and operating system code that rarely uses Goto, Ternary, If, For, Do, While or Switch statements in C then you could pull it off, however, such a subset of C minus those constructs would not be Turing Complete so it'd be damn hard.

      And then, even if you did succeed it would still be slower than the CPU code since not all workloads are compatible with SIMD, it only works on parallel streams; workloads that consist of multiple unrelated units (eg. pixels) that need to have the exact same operation performed on them whilst not depending on the results from the other pixels as part of those operations.

      This is the major benefit of a CPU vs the GPU, the CPU can handle branch dense code with lots of interdependencies without too much stalling, the GPU cannot.

      ---

      Of course, Intel's Larrabee GPU may change all this but that remains to be seen until it hits the market

    4. Re:How did this get modded "insightful"? by imakemusic · · Score: 1

      The whole "ing"?

      --
      Brain surgery - it's not rocket science!
    5. Re:How did this get modded "insightful"? by Anubis350 · · Score: 1

      It's fairly common usage in HPC circles, vis. NCSA to add 47 teraflops of compute power with new heterogeneous system

      --
      "goodbye and hello, as always" ~Prince Corwin, from Zelazny's Amber series
  10. Re:Honestly, at this point... by Ethanol-fueled · · Score: 1

    Probably waiting for a future in which computers are just big FGPAs and SoCs defined entirely by their firmware, with peripheral jacks soldered to their pins.

  11. Re:Honestly, at this point... by jhulst · · Score: 2, Insightful

    Sure, they have lots of power, but only when used for parallel tasks. Each individual core is considerably slower than a normal CPU core and much more limited in what it can do.

  12. So... by fuzzyfuzzyfungus · · Score: 1, Interesting

    Will they also be announcing support for an underfill material that doesn't cause the chip to die after a fairly short period of normal use? And, if they do, will they be lying about it?

    1. Re:So... by PC+and+Sony+Fanboy · · Score: 2, Funny

      OF COURSE, who do you think they are? Apple, Sony, HP and Microsoft?

      Nvidia just makes the cards. It isn't their fault if they're not installed, cooled or properly read bedtime stories after use.

  13. Games before hardware by DigiShaman · · Score: 4, Insightful

    Back in the day up till the year 2000, I used to upgrade my PC four times a year. The point was to always improve multi-tasking and obtain faster frame rates with higher detail in games that I already have. Since then however, the hardware has always been "good enough" for general computing and playing even the latest/popular games. The only time I'm compelled to upgrade my computer (mainly the video card) is if there's a game out that I love.

    Honestly, the only game I'm looking forward to is Diablo3. Even then, my nVidia 8800GT card should be more than sufficient. If not, it would be games like these that will send me over to Newegg to make a purchase. Given the lack of games compounded with hardware that's already decent in the market, I'm willing to bet it's got Intel, AMD, and nVidia scared. Who really wants/need bleeding edge technology anymore? Am I wrong thinking the desire for better video card technology has plateaued in the last few years?

    --
    Life is not for the lazy.
    1. Re:Games before hardware by cjfs · · Score: 1

      Who really wants/need bleeding edge technology anymore?

      Numbers...must...go...higher...

    2. Re:Games before hardware by Ethanol-fueled · · Score: 1

      Honestly, the only game I'm looking forward to is Diablo3. Even then, my nVidia 8800GT card should be more than sufficient. If not, it would be games like these that will send me over to Newegg to make a purchase.

      Recent games are just re-hashes of stale material (or otherwise lacking in storytelling) with emphasis on eye candy to make people continue to buy the latest graphics hardware. The true "good enough"-ers probably don't give a shit about Space Marines and Zombies 10.

    3. Re:Games before hardware by jpmorgan · · Score: 5, Interesting

      Notice the features being marketed: concurrent CUDA kernels, high performance IEEE double-precision floating point performance, multi-level caching and expanded shared memory, high performance atomic global memory operations. NVIDIA doesn't care about you anymore. Excepting a small hardcore, gamers are either playing graphically trivial MMOs (*cough*WoW*cough*) or have moved to consoles.

      They won't want to sell you this chip for a hundred bucks, they want to sell it to the HPC world for a couple thousand bucks (or more... some of NVIDIA's current Tesla products are 5 figures). The only gamers they're really interested in these days are on mobile platforms, using Tegra.

    4. Re:Games before hardware by icebraining · · Score: 1

      You do realise your card is less than two years old? In the year 2000 they were releasing the Geforce 2 Ultra, I bet you can't play Doom3 in it.

      Graphics are still very far from "realistic", and till then graphic cards will continue to evolve; the 8800GT may seem more than enough for now, but it won't seem that way some years from now, Unless, of course, OnLive takes over the PC gaming market.

      --
      Dilbert RSS feed
    5. Re:Games before hardware by LordBullGod · · Score: 1

      I have to agree with you. It seems to me that the power of GPUs over a few years has gotten better, but for what (from a gamers view)? I have not found to many games (that is what I use CPU/GPU power for anyway) that really need the amount of power that is offered right now. Yes, ok, for the nice cinematic movies and cut scenes, but REAL game play. I still run a crossfire rig using two year old cards (2900XTs), and there has not been to much to challenge it lately. Yea, yea, Crysis makes it work, but it SHOULD. Where is the programming in games to support the GPU power that is available? No I don't program, no I don't develop, but really, where is the game play? I had also heard a lot about physics being supported directly on the GPU....bla bla bla, I don't see much of that either. Great, the cards support it, where is the programming other then in tech demos? I'm with you Dig, I used to upgrade all the time (2x /yr), but right now, I just don't see the need. Are the manufacutrers using us PC folks to test waht will and won't work for consoles? I feel that way, and i won't play into it. I'm good for now, but upgrades for me are now 1-2 years apart, not bi-annually. Just my view. -BG-

    6. Re:Games before hardware by Pulzar · · Score: 5, Interesting

      People have been saying that forever now. I think only the first 2 generations of 3D cards were greeted by universal enthusiasm, while everything else since had a number of "who needs that power to run game X" crowd. The truth is, yes, you can run a lot of games with old cards, but you can run them better with newer cards. So, it's just a matter of preference when it comes to the usual gaming.

      AMD/ATI is at least doing something fun with all this new power. Since you can run the latest games in 5400x2000 resolutions with high frame rate, why not hook up three monitors to one Radeon 58xx card and play it like this? That wasn't something you could do with an older card.

      Similarly, using some of the new video converter apps that make use of a GPU can cut down transcoding from many hours to one hour or less... you can convert your blu-ray movie to a portable video format much easier and quicker. Again, something you couldn't do with an old card, and something that was only somewhat useful in previous generation.

      In summary, I think the *need* for more power is less pressing than it used to, but there's still more and more you can do with new cards.

      --
      Never underestimate the bandwidth of a 747 filled with CD-ROMs.
    7. Re:Games before hardware by Nemyst · · Score: 1

      The problem is consoles: with releases showing up on a vast array of systems with wildly different capabilities and most games now coming out on consoles first or at the same time as on the PC, it would make no sense for developers to create a game which would be too complex/heavy to be ran on a substantial portion of machines (read: Xbox 360s and PS3s, not even counting the Wii). Thus, games get stale as "old" hardware doesn't get upgraded.

      This generation is noticeably different in that consoles now have similar capabilities to PCs and that there no more is that differentiation between PC games and console games, whereas with earlier generations you had a lot of "PC only" and "consoles only" games which were specifically tailored for each medium.

    8. Re:Games before hardware by AHuxley · · Score: 1

      Perfect for the 640p xbox and ps3 generation.
      Real gamers still dream of more than Sony or MS allows you to have ;)

      --
      Domestic spying is now "Benign Information Gathering"
    9. Re:Games before hardware by Heir+Of+The+Mess · · Score: 3, Insightful

      Since then however, the hardware has always been "good enough"

      That's because most games are now being written for consoles and then being ported to PC, so the graphics requirements are based on what's in an X-Box 360. Unfortunately consoles are on something like a 5 year cycle. People are now buying a game console + a cheap PC for their other stuff for cheaper than the ol gaming rig. Makes sense in a way.

      --
      Australian running a company that does C# / C++ / Java / SQL / Python / Mathematica
    10. Re:Games before hardware by VocationalZero · · Score: 1

      Notice the features being marketed: concurrent CUDA kernels, high performance IEEE double-precision floating point performance, multi-level caching and expanded shared memory, high performance atomic global memory operations.

      You say this like it is somehow a bad thing.

      NVIDIA doesn't care about you anymore.

      Are you always this melodramatic, or maybe you work for ATI or something?

      Excepting a small hardcore, gamers are either playing graphically trivial MMOs (*cough*WoW*cough*) or have moved to consoles.

      Really? So I now qualify as "hardcore" because I casually play games like Unreal 3 and Arkham Asylum? You are so very wrong. The PC game market has taken a hit along with everything else (except short selling) because of the economy, but it is still a billion dollar industry thats really only to get bigger.

    11. Re:Games before hardware by plague911 · · Score: 1
      Yes and no. There is a danger coming to both Intel nvidea and AMD/ATI within the next 20 years. Computing tech of all kinds will be old hat.... for the average consumer. The human eye can only see so many pixels the human nervous system only has certain response time. Individuals who only use a computer for day to day activities like web/email and a movie or two probably have or just about have hit the point where they dont care any more. There are however those of us who still just "like" computers and the cool new tech. So far there is still a big enough change between generation to make it interesting. For example i cant wait to be able to build a system with a "affodable" SSD within the next year, or id think it would be way sweet to have a 6-core HT so 12 effective core chip so my simulations get done way quick (Im a college student so i do simulations in matlab on my own computer ).

      The most interesting thing i however see coming out of this. Nvidia taking on Intell in the super computing field. I honestly see Nvedia getting pushed out of the market if they stay where they are. Intel is trying to get into the gpu market. and Both Amd/Intel have more resources than Nvidia

    12. Re:Games before hardware by hairyfeet · · Score: 1

      You're not wrong. Most of my customers just rave about how "blazing fast" the new dual and quad AMDs I've been building them are, and I sure as hell ain't been building with the top o' the line chips or GPUs. For most of the tasks folks do with their computers the chips have gotten well past 'good enough" for a couple of years now.

      That is why after being a lifelong Intel+Nvidia guy I have switched over to AMD builds, even going so far as to eat my own dogfood and building an AMD 7550 dual box for myself. Even in high def these new dual and quad chips paired with the new onboard GPUs give just an incredible bang for the buck, and I was playing Far Cry and Bioshock on the 780v board for a couple of months before breaking down and buying a cheap card for the thing. We truly live in great times as far as processing power goes, when I can build my customers really nice dual cores for less than $550 and quads for less than $750 and still make a nice profit.

      And having support for all the major codecs "out of the box" on today's IGPs (DivX, MPEG2-4, H263/264) just adds the wonderful icing to the delicious cake. Hell as I sat up a customer's wireless network the other day he had to call his family in just to show them how the new quad I built him could do LoTR high def while running his browser, messenger, and email all at the same time with nary a hiccup. I just had to smirk at all the "ooohs" and "aaahs" he was getting as he showed off his new PC. And of course that was just the low end AMD quad paired with an Nvidia IGP (an 8400 I believe) and I have a few customers that like I did are gaming on their AMD IGP until they get around to having me pick up them a card, and the framerates are quite smooth when paired with 4Gb of decent RAM.

      So I'd have to say you hit the nail firmly on the head. While there are still guys out there that are into the whole epeen "my box is teh elite!" stuff, even the lowest of the low end now seems to have surpassed "good enough" by a pretty good margin for what most folks are doing with them. that is one of the reasons I switched to AMD, as the lower price allows me to give them more RAM and HDD, which IMHO matters more today than whether or not you have the latest and greatest in GPU or CPU.

      --
      ACs don't waste your time replying, your posts are never seen by me.
    13. Re:Games before hardware by PC+and+Sony+Fanboy · · Score: 3, Funny

      Young coders are too lazy and brainwashed by MS and Sony to think anymore. You finally have the bandwidth and cpu and gpu to do something and your stuck dreaming at 640p. Hack the cards with Linux and dream big. Take computing back from the DRM, locked down junk MS and Sony code down to. You have the OS, now get some graphics freedom too.

      I hope you don't contribute to the trunk, your code is 2x as long as it should be, and only half as effective!

    14. Re:Games before hardware by PC+and+Sony+Fanboy · · Score: 1

      There is a danger coming to both Intel nvidea and AMD/ATI within the next 20 years.

      NEWS FLASH... When you're 2x as old as you are right now, the world will have changed.

      Get off my lawn.

    15. Re:Games before hardware by PC+and+Sony+Fanboy · · Score: 1

      even going so far as to eat my own dogfood

      You do realize this implies you work for ATI/AMD? If not, perhaps you should do a quick check on what dogfood means.

      According to your article, you're obviously price sensitive, so why would you *ever* pay full price for a product (from your competitor)?

      Clearly you need to inform yourself before attempting to inform others.

    16. Re:Games before hardware by VoltageX · · Score: 1

      I haven't seen one GPGPU video encoder that doesn't suck - yet.
      Even the one NVIDIA is touting, "Badaboom" has a Fischer-Price interface and produces some of the worst looking H264 video I've ever seen.

      --
      "Anonymous could not immediately be reached for further comment." - International Business Times
    17. Re:Games before hardware by mwvdlee · · Score: 1

      NVIDIA doesn't care about you anymore.

      They never did.
      Nor does any other company.
      Companies care about your money.
      Name me one company that actually cares more about you than your money.

      --
      Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
    18. Re:Games before hardware by eggnoglatte · · Score: 1

      Or maybe you are just getting old/mature, and are no longer drooling over toys for toys' sake? Just saying...

    19. Re:Games before hardware by wagnerrp · · Score: 1

      The problem is that expensive big name titles are typically the only ones that can afford to push for high end graphics, only now they are limited by what the consoles are capable of, to make them cross platform. As a result, modern high end graphics cards are overkill on everything but the largest 2560x1600 monitors. ATI seems to have anticipated this, which is why they released their new card with support for 6 monitors, and some capacity for spanning with current games. Multi-monitor systems are becoming more common, and multi-monitor gaming may be the only reason (until the next console bump) to need high end graphics on a PC.

    20. Re:Games before hardware by Kjella · · Score: 1

      even going so far as to eat my own dogfood

      You do realize this implies you work for ATI/AMD? If not, perhaps you should do a quick check on what dogfood means.

      Umm... no, it means to use whatever you're selling and recommending. To take a car analogy (surprise) if you sell Fords for a living and drive a Toyota you're not eating your own dogfood. The implied message is that you'd go to a different store and buy a Toyota, even though you don't work for Ford. Same if you're selling clothes you'd never wear yourself, even though you're a retailer and not working for the brands you carry. So if he was recommending AMD systems to everyone but using Intel systems himself, I'd say he's using the expression absolutely correctly. Of course you shouldn't pull the concept to absurdity, but it's certainly broader than you imply.

      --
      Live today, because you never know what tomorrow brings
    21. Re:Games before hardware by RMH101 · · Score: 1

      Not quite. To eat your own dogfood means that you use something that you or your company have developed, which isn't quite ready for general consumption - it's in beta, it's buggy, the GUI's not quite there and you're using it before it's available on sale or to the public.

    22. Re:Games before hardware by ioshhdflwuegfh · · Score: 1

      Most of my customers just rave about how "blazing fast" the new dual and quad AMDs I've been building them are, [...]

      That surely puts a whole new perspective on the next generation of GPUs.

    23. Re:Games before hardware by BikeHelmet · · Score: 1

      The desire for better CPUs is somewhat plateauing, too. I can't recall seeing a game with min reqs higher than a 2ghz Athlon X2 or 3ghz Pentium D. That's despite much faster dual and quad cores becoming the norm.

      For a while, I actually played Left4Dead on my old Athlon XP 2400+ and 7800GS. Got about 30fps with VSYNC on, and settings set somewhat low. But now I've got an Athlon II X2, and 8800GS (which cost $30 brand new, one year ago!)

      I don't have any upgrade plans at the moment. If a game runs too slowly, I'll hunt around for a good deal - but for now, my desire to upgrade is gone.

    24. Re:Games before hardware by jpmorgan · · Score: 1

      HPC has a very, very long tail. The market for systems in the $50k-$200k range, and beyond, is enormous. Companies like Exxon and Saudi-Aramco have an insatiable desire for computing power and more money than god. Everything is done in simulation these days.

      I mean, look at Agilent. Agilent has twice NVIDIA's revenue and they make testing and lab equipment. As for Roadrunner itself, IBM makes a lot more money on the day-to-day HPC business than they do on the occasional hypercomputer for the US government.

    25. Re:Games before hardware by smallfries · · Score: 1

      That's right. After conquering the competitive but profitable mass-market for their products, where they can make a killing in low-margin high-volume sales, they are going after a tiny niche....

      It would certainly have nothing to do with competition from Larabee and the general realisation that as the GPU becomes more general purpose Games will seek to offload more calculations onto them. For graphics it's rare to hit a case that needs double-precision (it happens in HDR), but when you move your physics code on there you don't want nasty rounding errors to destroy the consistency of your virtual world.

      --
      Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
    26. Re:Games before hardware by L4t3r4lu5 · · Score: 1

      why not hook up three monitors to one Radeon 58xx card and play it like this?

      Because not even in the publicity shot could they get that dirty great inch gap from between the top and bottom tiers of screens. The horizontal looks ok (but you lose definition in the resin overlay between the horizontal monitors), but that joint right in the middle of where you're looking would be very similar to constantly having a piece of masking tape over the middle of your current monitor.

      Why not just output to a high-def TV?

      N.B. Strategy games do not scale well to high-def TVs. The resolution is great, but for PC gaming you're usually around 1-2m away from the screen; Consoles could be well over 4m with wireless controllers. It becomes extremely hard to see what's going on, or read dialog (having played Empire Total War on a 1080p screen).

      --
      Finally had enough. Come see us over at https://soylentnews.org/
    27. Re:Games before hardware by Waccoon · · Score: 1

      Plus, it brings down the cost on the low end. I love getting massive power for $100.

      More enthusiast cards, please!

    28. Re:Games before hardware by hairyfeet · · Score: 1

      And what exactly out there is high end besides Crysis, which frankly wasn't very good? If you want me to drop some other names I played the FEAR series from 1 through PM, MOH:Airborne, hell there really haven't been many good games out for PC lately, thanks to "multiplatform" being a code word for a badly ported X360 or PS3 game.

      The point was for the games the average folks have and are playing, from FEAR to Sims whatever number they are up to now, the new IGPs from Nvidia and especially ATI pack a surprising amount of kick. Maybe it is because I'm an old Greybeard who remembers the days of when if you didn't shell out crazy money for a gamer board you were getting stuck with an "Intel IGP O' Suck" like the I810 on the circa 2000 1.1Ghz I'm typing this on, where trying to do anything more with it than drawing the screen is pretty much pointless. The distance we have come as far as "bang for the buck" is really quite amazing if you think about it.

      With the newer IGPs you get not only decent framerates in most of the popular with average folks games, but you get hardware acceleration for all the major codecs folks use which makes video watching smooth as butter even while doing some serious multitasking. All this in a chip that doesn't suck tons of juice or turn your PC into a space heater. When I started out, in the days of the 486DX and the Pentium 1, if you didn't get a soundcard as well as an offboard GPU you really weren't gonna have a nice experience. Now with the exception with laptops, which still seem stuck in the "Intel IGP O' Suck" even the lowest end boxes I'm building give customers everything they want and more. I'm happy, they're happy, it's a win/win.

      --
      ACs don't waste your time replying, your posts are never seen by me.
    29. Re:Games before hardware by Kjella · · Score: 1

      Answers.com has this:

      An expression describing the act of a company using its own products for day-to-day operations. (...) This slang was popularized during the dotcom craze when some companies did not use their own products and thus could "not even eat their own dog food".

      In other words, to NOT eat your own dog food means you SELL it but you're not using it yourself. If you are using it too, you are eating your own dog food. I'm not sure when that expression changed, if ever, to mean that you're not selling it yet. I guess your use is more a result of the other, because there's so much focus on "eating your own dogfood" you end up eating crap for marketing purposes.

      --
      Live today, because you never know what tomorrow brings
    30. Re:Games before hardware by gad_zuki! · · Score: 1

      >so the graphics requirements are based on what's in an X-Box 360.

      I dont think thats such a limiting factor. Lets say they develop the xbox game first, instead of the PC version. They settle on 1080i for resolution and only a certain level of quality for textures. They also tone down the physics and AI to a level it doesnt slow down the xbox cpus.

      Okay, now when you port the PC version, you let the user select the resolution he likes and you up the textures to max and ungimp the physics and AI. Its not that hard. Any company that wants to produce a good PC game from an Xbox start is able to. The real question here isnt the technical limitations, which are easy enough to get past, but if the business wants to produce a quality port. If the PC market isnt big enough then they have little incentive to make a decent port and will just outsource the port to some shitty porting company and PC users will just have to deal with it.

    31. Re:Games before hardware by linhares · · Score: 1

      AT&T?

    32. Re:Games before hardware by Kjella · · Score: 1

      They've been saying it forever, but it's starting to come true. Let me quote you from Anandtech's HD 5850 review, benchmarks starting at page 3:

      "Crysis Warhead: Warhead is still the single most demanding game in our arsenal, with cards continuing to struggle to put out a playable frame rate with everything turned up."
      "Far Cry 2: Thankfully it's not nearly as punishing as Crysis, and it's possible to achieve a good framerate even with all the settings at their highest."
      "Battleforge: [Has no real one-sentence summary, but 30+ fps with everything max'ed]"
      "World of Warcraft: (...) As WoW is known to be CPU limited, we can turn it up to rather silly settings and still get a playable framerate on just about anything."
      "HAWX is another game that's not particularly GPU-bound, which means we can turn in some high numbers."
      "Dawn of War II is our other RTS benchmark. It's among the more challenging games in our collection, leading to there being a definite cutoff for playability."
      "Resident Evil 5: (...) As is often the case with console ports, it's not particularly GPU starved, and can crank out high numbers on just about anything"
      "Batman: Arkham Asylum: (...) Since we can't use AA, the name of the game is still "runaway performance", with the 5850 bringing in 88fps."
      "Left 4 Dead: As the Source engine is CPU limited, this is once again going to be a collection of ridiculously high frame rates."

      These are all comments at 2560x1600 with everything cranked up to max. And yet half the branchmarks they could come up with are basicly irrelevant even with everyhting cranked to max. If you play at anything less, like I do at 1920x1200 then most of the other half too. There's Crysis but cards are leaping ahead of the games, what we lack for realism isn't more power but lots and lots of programming of details for all those shaders - which is expensive as hell. I'm saying that even with the hardware, it might not pay off to invest that money in the software.

      --
      Live today, because you never know what tomorrow brings
    33. Re:Games before hardware by Sycon · · Score: 1

      Unfortunately consoles are on something like a 5 year cycle.

      Actually, they're trying to get consoles on a larger development cycle. The Xbox 360 came out in 2005 and they're not expecting to replace it until 2014-2015. The PS2 has already experienced a 9 year life and its still selling strong.

    34. Re:Games before hardware by fastest+fascist · · Score: 1

      I bet you also played games a lot more back then. Maybe you just grew up?

    35. Re:Games before hardware by Pulzar · · Score: 1

      Why not just output to a high-def TV?

      HDTV displays are big, but the resolution is still only 1920x1080. The resolution you get from 3 monitors is much higher, and you get a very different aspect ration... so you can show a lot more content on the sides. You increase your field of vision in the game vs. just having a big display.

      --
      Never underestimate the bandwidth of a 747 filled with CD-ROMs.
    36. Re:Games before hardware by NoOneInParticular · · Score: 1

      He's been building AMD boxes for his customers. Now he built an AMD box for himself. Why is this not eating his own dogfood? Following your logic, only farmers can eat their own dogfood, not dogfood-makers.

  14. Re:Honestly, at this point... by Lord+Crc · · Score: 1

    Why bother buying a computer motherboard, cpu and case? Maybe the case to store it in, but you could make a full fledged computer with just a graphics card they are so powerful.

    GPU's vs CPU's is a bit like having 5000 highly trained monkeys vs 5 highly trained people. If your task is easy enough for the GPU, it'll do it blazingly fast. On the other hand, for some tasks the CPU is still the better option.

  15. Re:Honestly, at this point... by blackraven14250 · · Score: 1

    It's really funny, because this kinda happens, in that things keep getting integrated into the CPU. It's really just that we keep adding more new stuff outside the CPU that keeps literally everything (except power regulation) from being one chip.

  16. GPU to network by soldack · · Score: 1

    I wonder when a GPU will be able to directly access a network of some sort. Right now, you would need glue code on the CPU to link multiple GPUs in different systems together. I imagine that some HPC applications would run quite well with 100 GPUs spread over 25-50 machines with a QDR InfiniBand link between them.

    --
    -- soldack
    1. Re:GPU to network by ceoyoyo · · Score: 1

      Probably not, without major architectural changes. Currently when you want something to run fast on a GPU you really want to have an algorithm where you load everything up and then just let it run.

      You could potentially improve that by cutting out the CPU, PCI, etc., but then you're not really talking about a graphics card anymore and you might as well just market it as a stream processor or a competitor to Cell blades.

    2. Re:GPU to network by symbolset · · Score: 1

      Intel has moved the northbridge and the GPU onto the processor. I understand the southbridge is next. The keyword is "SOC" or System on a Chip. They're not inventing this here -- others have done it long ago. The southbridge is the part that talks to the physical port controller (PHY) for the network. PHYs will remain discrete components because of their radically different power requirements - until we go optical. But Intel hasn't promised us multiple 10Gbps optical ports on consumer equipment until next year, and 100Gbps per port could be 10 years out.

      Larrabee is supposed to be a massively parallel cache-coherent x86 architecture with 24/32/48 cores on one chip.

      But of course for serious work you could put four of them in one 2U system and use your fancy Infiniband to interconnect your 192-core nodes. You'll need some insane power and cooling to fill a rack with 22 of those but at 172TFlops per rack (If the thing hits 2GHz with 32 cores), maybe it's worth it: you're in the top 20 supercomputers in the world, in one rack. I'd probably go with a shared PCIe bus myself, though, and tie together 8-packs of servers in a tiered arrangement, but I'd put five in each server that wasn't a PCIe hub. I'm cheap that way - I'll go asymmetrical to save a few bucks. On the upside, there's space in there for 172TB of local SAS storage as well. Come to think of it, I might try and fit in some SSD.

      The Larrabee add-in cards had better come equipped with external DC power connectors and external water or compressed air connectors so they can fit in a single-wide slot. There's no way the PSUs or thermal solutions in servers are going to handle those Watts - unless it comes in significantly under the expected TDP.

      --
      Help stamp out iliturcy.
  17. Another article here by Vigile · · Score: 3, Informative

    http://www.pcper.com/article.php?aid=789

    Just for a second glance.

  18. More than just graphics by mathimus1863 · · Score: 5, Informative

    I work at a physics lab, and demand for these newer NVIDIA cards are exploding due to general-purpose GPU programming. With a little bit of creativity and experience, many computational problems can be parallelized, and then run on the multiple GPU cores with fantastic speedup. In our case, we got a simulation from 2s/frame to 12ms/frame. It's not trivial though, and the guy in our group who got good at it... he found himself on 7 different projects simultaneously as everyone was craving this technology. He eventually left b/c of the stress. Now everyone and their mother either wants to learn how to do GPGPU, or recruit someone who does. This is why I bought NVIDIA stock (and they have doubled since I bought it).

    But this technology isn't straightforward. Someone asked why not replace your CPU with it? Well for one, GPUs didn't use to be able to do ANY floating or double-precision calculations. You couldn't even program calculations directly -- you had to figure out how to represent your problem as texel- and polygon-operations so that you could trick your GPU into doing non-GPU calculations for you. With each new card released, NVIDIA is making strides to accommodate those who want GPGPU, and for everyone I know those advances couldn't come fast enough.

    1. Re:More than just graphics by mathimus1863 · · Score: 1

      You don't need to "trick" the GPU anymore, but it is still lacking tons of functionality. I remember a coworker complaining that he got some double-precision operations working, but everything crashed if he used any for-loops. Additionally, there are different bit-representations for variables on CPU vs GPU which can complicate things to hell when trying to get the CPU and GPU to cooperate with each other. Yes, CUDA is like C or C++, but it's not the same, yet. Hence, new releases like this get people excited.

      Personally, I've found it to require too much time-investment to learn how to do GPGPU programming, but new cards with updated functionality help flatten the learning curve.

    2. Re:More than just graphics by six11 · · Score: 1

      I'm curious to know how you go about writing code for GPUs. I've been thrown into a project recently that involves programming multicore architectures, so I've been reading about StreamIt (from MIT). It looks really interesting. But they don't mention GPUs in particular (just multicores), probably because the current batch of GPUs don't have a lot of candy that CPUs have (like floats).

    3. Re:More than just graphics by Anonymous Coward · · Score: 2, Informative

      Start looking at OpenCL as soon as possible if you want to learn gpgpu, cuda is nice but opencl is portable between vendors and hardware types :)

    4. Re:More than just graphics by LucidLion · · Score: 1

      If you have a compatible nVidia graphics card (anything within the last 3 years), you should look into CUDA (http://www.nvidia.com/object/cuda_get.html). It's got a steep learning curve at first, but it's quite awesome. OpenCL is another option designed to be more cross-platform. Either way, you basically write some code in C (well...C with a few extensions and limitations) and it executes in parallel on the GPU. The current batch of nVidia GPUs support floats with both single precision and double precision, although there is a significant performance penalty with double precision for most applications. With the announcement nVidia made today, though, there will soon be new cards with much less of a performance penalty for double precision.

    5. Re:More than just graphics by drgould · · Score: 1

      But this technology isn't straightforward. Someone asked why not replace your CPU with it?

      Because GPUs have no memory management unit, so they can't run any modern multitasking operating system worth beans; Windows, Linux, Mac OS, whatever.

      And why should they? That's not what they're designed for.

    6. Re:More than just graphics by Mad+Merlin · · Score: 1

      GPUs (of the past) are basically just massively parallel floating point units. I think the OP meant to say that they lacked integer operations and double precision floats in the beginning.

      --
      Game! - Where the stick is mightier than the sword!
    7. Re:More than just graphics by Belisar · · Score: 1

      Actually, GPUs started out doing all rendering in fixed point arithmetic, i.e. the equivalent to integers. That worked fine for rasterizing and shading for quite a while.

      Then, they started doing limited-precision floating point support (with some 16-bit 'half' floats and other non-standard-conforming weirdness). Only later did they actually go on to support full IEEE floats. The current (as in, can buy them now) generation added IEEE double for both manufacturers, but of course performance is about an eighth or so of single precision.

    8. Re:More than just graphics by ioshhdflwuegfh · · Score: 1

      I work at a physics lab, and demand for these newer NVIDIA cards are exploding due to general-purpose GPU programming. With a little bit of creativity and experience, many computational problems can be parallelized, and then run on the multiple GPU cores with fantastic speedup. In our case, we got a simulation from 2s/frame to 12ms/frame. It's not trivial though, and the guy in our group who got good at it... he found himself on 7 different projects simultaneously as everyone was craving this technology. He eventually left b/c of the stress.

      That sounds like a physics lab alright!

    9. Re:More than just graphics by smallfries · · Score: 1

      Go and have a look at GPGPU. There's tons of material on there about techniques, some tutorials and a busy forum.

      --
      Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
    10. Re:More than just graphics by makomk · · Score: 1

      Hmmm? Pretty sure that the last/current-gen ATI cards can do double precision floating point at reasonable speed, though it's technically not 100% IEEE-compliant.

    11. Re:More than just graphics by hattig · · Score: 1

      Fermi supports IEEE-754.

      It also supports C++ and Fortran programming directly - probably with some caveats but nothing major.

      Yeah, it started off with tricking the GPU, but soon after NVIDIA and ATI released CUDA and CTM, and now they're supporting a wealth of different APIs (NVIDIA will soon be ahead with Fermi considering the C, C++ and Fortran support, and the developer tools). AMD will probably have a new architecture late next year too (RV870 is a tweaked, enhanced RV770, which was a tweaked, enhanced R600, a good path for them to take, but they can't do it forever).

  19. Can also be useful in graphics by Sycraft-fu · · Score: 2, Insightful

    It depends on what you are doing, but when you get something that involves a lot of successive operations, even 32-bit FP can end up not being enough precision. You get truncation errors and those add up to visible artifacts. This could also become more true as displays start to take higher precision input and even more true if we start getting high dynamic range displays (like something that can do ultra-bright when asked) that themselves take floating point data.

    1. Re:Can also be useful in graphics by TheRaven64 · · Score: 1

      You need a lot of rounding errors for that to be the case. 16-bit floats per channel give more range than the human eye can perceive. If your 32-bit floats are causing artefacts then you must have lost more than half of your precision to rounding errors. It's possible that this is the case for geometry, but usually floating point errors from geometry come from not scaling properly (if you do everything a long way from the origin then you are wasting a lot of the floating point value's precision).

      --
      I am TheRaven on Soylent News
    2. Re:Can also be useful in graphics by MrResistor · · Score: 1

      I think you'd be surprised how quickly rounding errors can compound. You might not see it much when doing graphics, but for scientific apps it's a huge problem.

      --
      Under capitalism man exploits man. Under communism it's the other way around.
    3. Re:Can also be useful in graphics by TheRaven64 · · Score: 1

      For a scientific app, sure, but the grandparent specifically stated graphics.

      --
      I am TheRaven on Soylent News
    4. Re:Can also be useful in graphics by MrResistor · · Score: 1

      So he did, I missed that.

      I don't know enough about graphics to know how often floats are used, but you'll have the same issues anytime you have numbers that are close to zero.

      --
      Under capitalism man exploits man. Under communism it's the other way around.
  20. Embedded x86? by Doc+Ruby · · Score: 1, Interesting

    What I'd like to see is nVidia embed a decent x86 CPU, (maybe like a P4/2.4GHz) right on the chip with their superfast graphics chips. I'd like a media PC which isn't processing apps so much as it's processing media streams, pic-in-pic, DVR, audio. Flip the script of the fat Intel CPUs with "integrated" graphics, for the media apps that really need the DSP more than the ALU/CLU.

    Gimme a $200 PC that can do 1080p HD while DVR another channel/download, and Intel and AMD will get a real shakeup.

    --

    --
    make install -not war

    1. Re:Embedded x86? by deathguppie · · Score: 1

      All I know is I want to be able to move, normal map, and texture 10M Triangles in real time. When they can do that.. I'll buy beer for everyone!

      --
      once more into the breach
    2. Re:Embedded x86? by Sponge+Bath · · Score: 1

      Your proposal sounds similar to the IBM/Sony cell architecture: one general purpose processor core with a collection of math crunching cores. The enhanced double precision FP in this latest Nvidia chip also maps the progression of cell with the PowerXCell 8i over the original cell processor.

    3. Re:Embedded x86? by Doc+Ruby · · Score: 1

      Indeed, I have two (original/80GB) PS3s exclusively for running Linux on their Cells. The development of that niche has been very disappointing, especially SPU apps. And now Sony has discontinued "OtherOS" on new PS3s (they claim they won't close it on legacy models with new firmware, but who knows).

      I'd love to see a $200 Linux PC with a Cell, even if it had "only" 2-6 (or 7, like the PS3) SPUs, and no hypervisor, but maybe a separate GPU (or at least a framebuffer/RAMDAC, but why not a $40 nVidia GPU?). The Cell shouldn't be so exotic and rare, and it certainly shouldn't cost $5000 for a workstation just to get 8 SPUs when it's really that bus that's the magic, and 2-6 SPUs are so fast, and such a CPU should cost under $80.

      --

      --
      make install -not war

    4. Re:Embedded x86? by Anonymous Coward · · Score: 2, Informative

      What I'd like to see is nVidia embed a decent x86 CPU,

      They did, its called Tegra. Except its not using the x86 hog, but way more efficent ARM architecture

    5. Re:Embedded x86? by Doc+Ruby · · Score: 2, Interesting

      That's better than nothing. But I want all the x86 packages, especially the Windows AV codecs. That requires an x86.

      Though that requirement suggests an architecture of ARM CPU for OS/apps, little x86 coprocessor for codecs, and MPP GPU cores doing the DSP/rendering. If Linux could handle that kind of "heterogenous multicore" chip, it would really kill Windows 7. Especially on "embedded" media appliances.

      --

      --
      make install -not war

  21. nice die shot by skarhand · · Score: 1

    As opposed to TFA, this article includes a nice die shot, for those that care.

  22. Re:Honestly, at this point... by symbolset · · Score: 1

    It's called Larrabee. Coming soon.

    --
    Help stamp out iliturcy.
  23. Re:Honestly, at this point... by PC+and+Sony+Fanboy · · Score: 1

    with a UID > 1.6m you clearly know how the mods function.

    Yeah, but he's been lurking since day 1. or was it day 2. It's so hard to remember when ... HEY YOU, get of my damn lawn!

  24. Re:linux users can suck horse cock by PC+and+Sony+Fanboy · · Score: 1

    AC posters suck

    Fixed that for ya.

    I'd have posted AC, but the refresh time is too long.

  25. Re:Honestly, at this point... by whatajoke · · Score: 1

    The anonymous coward posts earlier were actually read by mods and brought to the attention of others. But lately I realised that the anonymous coward posts were no longer floated upwards by the mods. So I had to create an account just so my posts will have atleast 1 point.

    These days if you want to be modded up, ask the same questions that were asked when a similar article appeared earlier, or just roll out endless memes.

    Maybe I am deluded about the times when redundant posts were really modded redundant.

  26. This is the easy way to go by raftpeople · · Score: 1

    I had my code up and running quickly. It takes a little more time to re-arrange the algorithm and data to get optimal efficiency and to learn about the strengths and weaknesses of the computing model, but still relatively easy for any competent C coder.

  27. Re:Intel had this 10 years ago. by symbolset · · Score: 1

    Intel was buying optical processor technologies a quarter century ago. I don't see any threats to their dominance in the next quarter century. They're not dumb. They could launch a 100GHz photonic processor if they needed to. They just don't need that advancement yet. I can't think of a better reason to buy AMD processors than that: Intel is not going to give us the good stuff until their dominance is threatened.

    --
    Help stamp out iliturcy.
  28. Re:Honestly, at this point... by mwvdlee · · Score: 1

    It's more like comparing a single William Shakespeare with 5000 monkeys that each memorized a small part of Hamlet; The monkeys will be able to write Hamlet in a few seconds, but only Shakespeare is able to write anything other than Hamlet.

    Anybody else got any contrived analogies? Something with cars perhaps?

    --
    Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
  29. Re:Honestly, at this point... by TheThiefMaster · · Score: 1

    Bus vs racecar? You'll transport 50 people faster in a bus than the racecar, but the bus isn't winning any speed races and doesn't corner as quickly.

  30. Re:Honestly, at this point... by ioshhdflwuegfh · · Score: 1

    Why bother buying a computer motherboard, cpu and case? Maybe the case to store it in, but you could make a full fledged computer with just a graphics card they are so powerful.

    Why bother with GPU and just go with a CPU they're so powerful?

  31. Re:Honestly, at this point... by smallfries · · Score: 1

    Maybe I am deluded about the times when redundant posts were really modded redundant.

    Accept certain inalienable truths, redundancy will rise, mods will
    philander, you too will get old, and when you do you'll fantasize
    that when you were young redundancy was reasonable, mods were
    noble and children respected their elders.

    And trust me on the sunscreen.

    --
    Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
  32. Re:Honestly, at this point... by smallfries · · Score: 2, Funny

    That is an astoundingly bad analogy.

    What about it's like having a regiment of 5000 soldiers vs 5 ninjas. If the task can be accomplished by rote then the regiment will win on sheer manpower, but it requires adaptability then the ninjas will triumph.

    Substitute pirates for ninjas for an instant paradox.

    --
    Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
  33. Not out until Q1 2010 by physburn · · Score: 1
    The Fermi (great name, was it after Erico Fermi the italian nuclear pioneer), want be out until next year, and early on it will be in the $400 top range cards, more that what most of us spend). So ATI has the lead for next 4 months and the christmas sales. The Fermi might be quicker than the current 5870 Radeon, but although ATI aren't ready for a new archicture or process bump. Chip Tweaking will probably get a usual 20% boost for later versions of the Radeon. ATI will be a lead of a bit. This a just as well as AMDs CPU are falling behind a lot. So it will that the graphic lead to keep AMD near profit.

    ---

    Graphics Card Feed @ Feed Distiller

  34. Emacs of the graphics cards by AniVisual · · Score: 2, Funny

    Hi thar. We gave you some useful hardware to support general purpose calculations in your graphics accelerator so you can compute while you compute.

    The only thing we can't support is decent graphics in games without resorting to special, NVIDIA-specific patches.

  35. I'm Nvidia's worst nightmare. by u64 · · Score: 1

    I dont need that much fps anymore.

    What i would like is a Geforce8800-performance card for 10$.
    And it should idle at 1W !

    That would be a must-have. Normally i dont pay for 3D cards, i
    just wait for rich people to get boored with thier cards. Then
    i ask: Are you gonna finnish that?

    (I do that with mobile phones too)

  36. Re:Honestly, at this point... by parlancex · · Score: 1

    For CUDA core is actually comparable to a CPU core in terms of raw processing power, what it lacks is branch prediction and a significant and intelligent cache control mechanism. As far as being more limited in what it can do? No. CUDA cores have a full floating point, integer, logical operation and branching instruction set. They can do anything a CPU core can.