Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
AMD Cloud Graphics Technology

AMD Reveals Radeon Sky Series For Cloud Gaming, Previews Radeon HD 7990 53

Posted by Soulskill from the future-of-gaming-whether-you-like-it-or-not dept.
MojoKid writes "AMD made a number of interesting announcements today at the Game Developers Conference, currently taking place in San Francisco. AMD revealed their 'Radeon Sky' series of graphics products targeted at cloud gaming and virtualized computing applications. The company also showed off the dual-GPU powered AMD Radeon HD 7990, and extended the 'Never Settle: Reloaded' gaming bundle program to include BioShock Infinite. AMD revealed three Radeon Sky Series cards, two based on the Tahiti GPU and another based on Pitcairn. The top of the line Radeon Sky 900 is powered by two Tahiti GPUs linked to 6GB of memory (3GB per GPU). The Sky 700 is powered by a single Tahiti GPU and the Sky 500 is based on Pitcairn. All of the cards are passively cooled and are designed for cloud gaming / computing servers. The upcoming high-end, consumer targeted Radeon HD 7990 was also previewed, but few details were given. Devon Nekechuk, Product Manager of AMD Graphics, did say the triple-fan setup was whisper quiet. We think it's safe to assume the card features 6GB of memory and clocks are in-line with current Radeon HD 7970 GHz Edition cards."
This discussion has been archived. No new comments can be posted.

AMD Reveals Radeon Sky Series For Cloud Gaming, Previews Radeon HD 7990 More

AMD Reveals Radeon Sky Series For Cloud Gaming, Previews Radeon HD 7990

Comments Filter:

  • yeah whatever... (Score:5, Funny)

    by fredan ( 54788 ) on Wednesday March 27, 2013 @04:20PM (#43295921) Homepage Journal
    how many megahashes do they produce?

    • Re: (Score:2)

      by Tynin ( 634655 )
      Since it is primarily the number of stream processors that drive the hash rate, my 7870 with 1280 stream processors does ~380 Mhash/s, so this new Sky 900 card with 3584 processors should be able to push ~1000 Mhash/s, considering they have near identical clock rates.

  • NVIDIA..? "Sky" (Score:3)

    by Lashat ( 1041424 ) on Wednesday March 27, 2013 @04:22PM (#43295965)

    Will NVIDIA release a repsonse product or do they already have a "cloud" offering that is not announced?

    anyone...

  • The 7990 is bigger than some motherboards. (Score:4, Funny)

    by RubberChainsaw ( 669667 ) on Wednesday March 27, 2013 @04:29PM (#43296079)
    It finally happened, add-in graphics cards are bigger than the rest of the computer.

  • But will it blend? (Score:3)

    by DigitAl56K ( 805623 ) on Wednesday March 27, 2013 @04:48PM (#43296309)

    Seems Blender's Cycles renderer is still having problems with AMD GPUs, and I see the finger pointed in AMD's direction. It would be nice to have some more hardware choices, but NVIDIA seems to be the only options at this point in time.

    • Seems Blender's Cycles renderer is still having problems with AMD GPUs, and I see the finger pointed in AMD's direction.

      In what specifically? I have to support nV and AMD graphics cards and get a lot of eng samples from both and when writing to the standard i don't find AMD cards to be any more buggy.

  • I know its just a reference design but putting three fans, even two fans, on a video card is stupid. In all likelyhood, that more than doubles the failure rate vs a single fan cooling setup...

    • I know its just a reference design but putting three fans, even two fans, on a video card is stupid. In all likelyhood, that more than doubles the failure rate vs a single fan cooling setup...

      That's unlikely. It would require that fan failures make up much more than 50% of total failures (I'm too lazy to do the math, I'm guessing at least 66%) and it assumes that there is no redundancy - that when you lose one fan, the other two can't speed up and cover the air-flow requirements at a cost of increased fan noise.

      • Out of the dozens of video card failures I've seen at work, only one wasn't caused by a cooling/fan problem (capacitor blew out, but this could possibly be traced to insufficient cooling as well). I'd say that 95% of the time (if not 99%) moving parts are the root cause of modern PC failures, i.e. fans and disks.

    • Re:3 fans on the 7990 is stupid (Score:5, Interesting)

      by Type44Q ( 1233630 ) on Wednesday March 27, 2013 @05:42PM (#43296969)

      I know its just a reference design but putting three fans, even two fans, on a video card is stupid. In all likelyhood, that more than doubles the failure rate vs a single fan cooling setup...

      All other things being equal, sure. However, which do you think would be more reliable, three fans at X rpm or one fan at 3X rpm? :)

      • Re: (Score:1)

        by Anonymous Coward

        That would depend very much on X.

      • However, which do you think would be more reliable, three fans at X rpm or one fan at 3X rpm?

        One fan at 3X... for a sufficiently small X.

        • Re: (Score:2)

          by GauteL ( 29207 )

          However, which do you think would be more reliable, three fans at X rpm or one fan at 3X rpm?

          One fan at 3X... for a sufficiently small X.

          This is an enthusiasts card, pushing the boundaries of the current generation AMD GPUs to squeeze out more performance. I think it is safe to say that X will not be sufficiently small. Or more accurately; 3X will not be sufficiently small.

          I would venture to guess that fan reliability scales super-linearly with fan speed and that we're talking about a speed for which the failure rate of one fan of speed X is smaller than the failure rate of any of 3 fans at 3X.

          I just don't think the AMD engineers are total i

          • This is an enthusiasts card, pushing the boundaries of the current generation AMD GPUs to squeeze out more performance. I think it is safe to say that X will not be sufficiently small. Or more accurately; 3X will not be sufficiently small.

            I would venture to guess that fan reliability scales super-linearly with fan speed and that we're talking about a speed for which the failure rate of one fan of speed X is smaller than the failure rate of any of 3 fans at 3X.

            I just don't think the AMD engineers are total idiots. Obviously I have no evidence, but I prefer not to assume they are.

            I see that you took my innocent little joke, and went with it! Well, OK then, I'll go one further and declare that my statement is true for X=0.

            Note that, logically, when I said "sufficiently small" I didn't talk about the needs of the graphic card's cooling. I was only looking at the fans' reliability.

    • If one of three fans die, you have time to save, shut down, replace.
      If one of one fans die, you hope the card will automagically shut itself down before permanent damage is done, that you have a replacement fan or card on hand, or that your mobo has a working vid out.

      Rather reminds me of a question from Chris Crawford's Balance of Power, after you've inadvertently started World War III: "What will you do now, smart person?"

      • Do you run special software to monitor your GPU fans at all times? I don't, nor do I want to. I just expect that if the fans fail other components will shut down before they suffer thermal damage (not always the case though, especially in GPUs /coughnvidiacough). Not to mention that if the fans are not hot-swappable, redundancy is pretty pointless because you still have to shutdown and get out a screwdriver. In this case, it looks like you may need more than a screwdrive to replace any of those GPU fans (as

        • It's not that I don't get what you're saying; different strokes, and all that. And I have no great liking for extra parts, either; just enough to get the job done as I prefer.

          Yes, I run psensors; right now it's displaying CPU temp in the icon in Unity launcher - I could change that to GPU. What I have set are some values in BIOS for "if all hell breaks loose", and psensors is set to notify me if certain temps are reached by either CPU or GPU. But since the machine is always on that wouldn't matter if I'm

  • I wish ATI would focus more energy on making their product stable, reliable and unable to be corrupted from innocent programs on the same computer. Right now, ATI and Nvidia drive the industry, and they couldn't give a damn about the BSODs they create!

    • Are you in the right year? Let me bring you up to speed, AMD bought ATI, ATI is gone. As far as stability, I can't say I've had any problem with AMDs drivers for my 6670 on Windows or Linux. Where as I have had all sorts of stability problems with Nvidia on Linux, which is odd since most people say Nvidia is the best for Linux, even though all the forums are full of question on how to resolve issues with Nvidia. Seems backwards to me. o0

  • Why is it that the fastest AMD based passive cards are based on the 7770 and yet for the "cloud" market they've got the equivalent of 7950 available? My recent build used a 650ti with super quiet fans because there is no current generation passively cooled card capable of 1080p gaming.

    • Re: (Score:2)

      by Endo13 ( 1000782 )

      Probably because the "cloud" parts are basically the same stuff, just clocked way down to run cooler? Just a guess.

    • Re: (Score:2)

      by Trongy ( 64652 )
      Passive cooling isn't the right terminology. CPU heat sinks in most severs these days don't have fans attached either. They rely on the much more powerful case fans in servers and the better optimised airflow. These GPU cards will work the same way. The servers will not be quiet.

      • Re: (Score:2)

        by afidel ( 530433 )

        Sure, but I can move a lot of CFM silently with 140mm fans compared to the little 60-80mm fans on the video cards.

  • Just as has been the story with ATI (now AMD) for more than a decade, it simply does not matter what kind of hardware they produce if they can't write a driver that is solid enough for things like gaming and GPGPU. No one is going to be satisfied with buggy GL, screen tears, etc., and things like that wreak absolute havoc on GPGPU solutions.

    I have tried ATI cards several times over the years only to be repeatedly disappointed to the point of returning them. Returns are so common that Newegg, who does not

  • How is the pci-e bus setup on the dual-GPU cards?.

    each GPU get's 8 lanes? pci-e switch? dual core?

    • Effectively its quite similar to a dual core CPU, except since the calculations are all so similar the workload is evenly split by the card & drivers.

  • The 7970 was great because it was only $500 for 2TFLOPS SP, 1 TFLOP DP, but the downside was it took 3 slots. The 7990 looks like it takes up only 2 slots. That means an ATX-sized motherboard like the Gigabyte GA-Z77X-UP7 [gigabyte.com] can handle three and possibly four (with case modification) such cards in tandem.

  • So the word "cloud" appears 11 times in the article, and in each case, is used in a context in which it comes off as a buzz-word or in an SEO-optimized fashion.

    What I'd like to know is, how does it fit into the "cloud" paradigm?

    • Does it somehow better handle streamed texture and shader content, thus making my "cloud gaming experience" feel more fluid?
    • Is the chipset designed to integrate with next-gen tablets and smartphones, which would typically be "cloud gaming experience"-devices?
    • Will it better handle ne

Slashdot Top Deals

A morsel of genuine history is a thing so rare as to be always valuable. -- Thomas Jefferson

Close