Quantcast
Problems with Distributed Computing - Page 4
Page 4 of 4 FirstFirst ... 234
Results 61 to 68 of 68

Thread: Problems with Distributed Computing

  1. Well... not unless you're trying to do raytracing...
    When you're raytracing, the pixel level is the first stage.
    Well you are quite right, when you hit the button in Maya or 3D studio this is what happens when you say render the scene. I was talking about sth else.
    Lets leave the raytracing for the GS there are lot of parameters before raytracing, lots of lists to be processed and send for display, if you think about the progress of a scene there is quite an advantage in a distributed architecture.
    You need to set up the world, kinematics, AI, etc, in a game before the "Photorealistic Result". and there are a lot of processes for this.
    And also think about the progress of a scene.
    I have also emphasized on the GPU or whatever someone wants to call it, it needs to be fast and capable for photorealism.
    Thats what I expect to see: gaming and photorealism meeting.
    But anyway this thread is for distributed computiing not for GPUs... ;-)

    Alex.

  2. Carefully reading the patent.
    The GPU is going to be inside the chip package. The chip is gonna look propably like configuration 6 in the patent... Now what will be inside the Pixel Engine remains unknown.
    Fig. 27 is also enlighting.

    Alex.

  3. Actually two chips with a bus...
    Yeap it is a super enhacement of the old now Emotion Engine in PS2.

    Alex.

  4. just for fun....

    http://tinypic.com/dd83q

    5 Dreamcast's combining their processing power through distributed computing. Total power is greater than that of the Gamecube on the left

  5. Re: ....

    Quote Originally Posted by Deadmeat
    This thread is so stupid.

    At 60 FPS, the screen refresh cycle is less than 1.6 ms. In other word, you have to prepare, send, compute, and retrieve, and render the display list to external APU in another box in 1.6 ms assuming one frame deferred rendering.

    Obviously, you cannot obtain that kind of responsiveness from an off-the-box ethernet connection to another CELL processor in another Sony branded box.

    So the distributed peer-to-peer processing model is worthless for gaming due to high latency. Gaming performance on PSX3 is restricted to the number of APU packed onto the die itself.
    Obviously your mind is limited, with advances such as Internet2, FTTH, and various other caching technologies, this could be a reality. Its a matter of time, before the ideas in thread become a reality.
    -Long Dead-'Editor-in-Chief' of PSINext.com

  6. #66
    just for fun....

    http://tinypic.com/dd83q

    5 Dreamcast's combining their processing power through distributed computing. Total power is greater than that of the Gamecube on the left
    I don't get the picture :?. Perhaps my monitor is just too dark to see anything.

    Side note: I can't seem to Quote megadrive's post using the quick "Quote" button. It simply brings me to the page with the Dreamcast pics :roll:. Weird.
    "Please understand that in cases where the OS itself is acting in an unstable fashion, this software could behave in unpredictable ways. Unfortunately, this happens to be the case at all times with Microsoft OS's." - Japanese PC game disclaimers!

  7. #67
    Obviously your mind is limited, with advances such as Internet2, FTTH, and various other caching technologies, this could be a reality.
    I don't see how FTTH and Internet2 are supposed to bring latencies and ping times down by over an order of magnitude. Besides which, FTTH and EFM are all things that the consumers want, but the telecom companies do NOT want to do, and go out of their way to lobby against. Why? Because they want to milk the DSL/Cable cash cow. If you have the technology to increase bandwidth by 35x, then the best thing to do from a business standpoint is to slowly increase what you provide by 2-fold every few years and to show all your customers that you're constantly improving services. Sure you can give a guy in Palo Alto a 100 Mbit connection, but will it bring the ping times to a guy in Copenhagen down from 110 ms all the way to 5-7 ms?

    Either way, don't pay much mind to Deadmeat. He's a zealot who can't divide correctly -- (It's over 16 ms, not less than 1.6). And he keeps talking about distributed rendering, and there's so much more to a game than rendering. There's plenty of things that don't need to be updated on every frame, but either way, I don't think it's feasible over the internet, and probably won't be within our lifetimes. Telecom companies would never bring that kind of backbone to the US so soon.
    Cell phones have changed mankind. Finally, men have something they can flip out and argue "mine is smaller than yours."

  8. #68
    Quote Originally Posted by cpiasminc
    Obviously your mind is limited, with advances such as Internet2, FTTH, and various other caching technologies, this could be a reality.
    I don't see how FTTH and Internet2 are supposed to bring latencies and ping times down by over an order of magnitude. Besides which, FTTH and EFM are all things that the consumers want, but the telecom companies do NOT want to do, and go out of their way to lobby against. Why? Because they want to milk the DSL/Cable cash cow. If you have the technology to increase bandwidth by 35x, then the best thing to do from a business standpoint is to slowly increase what you provide by 2-fold every few years and to show all your customers that you're constantly improving services. Sure you can give a guy in Palo Alto a 100 Mbit connection, but will it bring the ping times to a guy in Copenhagen down from 110 ms all the way to 5-7 ms?

    Either way, don't pay much mind to Deadmeat. He's a zealot who can't divide correctly -- (It's over 16 ms, not less than 1.6). And he keeps talking about distributed rendering, and there's so much more to a game than rendering. There's plenty of things that don't need to be updated on every frame, but either way, I don't think it's feasible over the internet, and probably won't be within our lifetimes. Telecom companies would never bring that kind of backbone to the US so soon.
    Well that isnt totally true about FTTH, Verizon is losing more money on DSL nowadays, thats why they are proactived about introducing FTTP around the USA. Already FTTP deploying is underway in parts of Southern CA, and other states. Do a serach on Verizon FTTP rollout. They plan to pass 4-5 million homes just next year with FTTP. I'll probably get FTTP in my city in 1-2 years from now. http://fios.3v3bizich.com/

    FIOS is still new, Verizon needs to train people to install the fiber and what not, already people in Keller Texas have FTTP running in there homes.

    http://www22.verizon.com/ForYourHome...ckagePrice.asp

    http://www.dslreports.com/forum/rema...9999~start=220

    FTTP is the best thing since slice bread

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

User Tag List

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •