Page 1 of 26 12311 ... LastLast
Results 1 to 10 of 251

Thread: Nvidia 'Shows Off' Fermi

  1. #1
    640k who needs more?
    Join Date
    Oct 2009
    Posts
    649

    Nvidia 'Shows Off' Fermi

    Without numbers.

    Considering how long they've been working on that benchmark, I would call that a bad sign.

    we know they've had a 'working' version of the benchmark running since november 17th '09

    http://www.brightsideofnews.com/news...benchmark.aspx

    And not to pat my own back, but myself and others in this thread http://www.semiaccurate.com/forums/s...Unigine+Heaven

    managed to spot some problems with the benchmarks image.

    Now they are showing the benchmark, but without scores. I won't say that means it underperforms the Radeons, but that absolutely means the scores don't live up in any remote way to the hype that is Fermi.

    Either they are having serious issues with tesselation like I suggested, the whole damn chip is borked, the driver team is incompetant, or as others suggested, the cards pictured in that story simply weren't Fermi. I'm not really willing to believe the driver team has come to suck so hard over at the goblin's camp, so I firmly believe this is a hardware thing of some kind.

  2. #2
    640k who needs more?
    Join Date
    Sep 2009
    Posts
    743
    Quote Originally Posted by JHFolsom View Post
    Considering how long they've been working on that benchmark, I would call that a bad sign.
    Now they are showing the benchmark, but without scores. I won't say that means it underperforms the Radeons, but that absolutely means the scores don't live up in any remote way to the hype that is Fermi.
    According to some of the reports I have read the NV reps at CES are telling everyone that GF100's score in the benchmark is "already higher than AMD's fastest GPU". Note the use of "GPU" rather than "card". My guess is that NV scores a bit higher than a 5870 and still slower than a 5970 but they aren't willing to disclose by how much.

  3. #3
    I just posted this in the other FERMI thread we must have been posting at the EXACT same time.

  4. #4
    640k who needs more?
    Join Date
    Oct 2009
    Posts
    649
    Lol, looks that way, forum says our posts were within 3 minutes of each other, meaning one of us was probably typing while the other was hitting the submit button, so do we want to merge that into this or this into that or leave em sepearate?

  5. #5
    Senior Member
    Join Date
    Jun 2009
    Posts
    2,626
    Quote Originally Posted by JAE View Post
    According to some of the reports I have read the NV reps at CES are telling everyone that GF100's score in the benchmark is "already higher than AMD's fastest GPU". Note the use of "GPU" rather than "card". My guess is that NV scores a bit higher than a 5870 and still slower than a 5970 but they aren't willing to disclose by how much.
    At what resolution? Were the resolutions at Popular settings? Max Settings? Low settings? Playable settings? What other settings (AA, and other junk)? Color depth? CPU?

    Heck, some sites were able to get Intel GPU run faster than an ATi low end GPU. But the game was not Playable.

    Let's wait for the benchmarks to come out. And let's see what the ATi response (if there is going to be one) is going to be.

    Edit:
    I like this part of the article "Reported bin splits of the A3 silicon are not looking promising for a hard launch."
    Also, maybe the high-end card, the one that is can hardly be manufactured is the only one that out-performs.
    Disclaimer: I don't doubt that Fermi is faster... The question is: Will it cook my bacon?
    -V
    My views and opinions do not represent those of my employer

  6. #6
    640k who needs more?
    Join Date
    Oct 2009
    Posts
    649
    At this point, I am not sure a response is necessary aside from watching the price/demand curve.

    Fermi is so late, that working on the next process and architecture is the only proper way forward.

  7. #7
    Quote Originally Posted by JAE View Post
    According to some of the reports I have read the NV reps at CES are telling everyone that GF100's score in the benchmark is "already higher than AMD's fastest GPU". Note the use of "GPU" rather than "card". My guess is that NV scores a bit higher than a 5870 and still slower than a 5970 but they aren't willing to disclose by how much.

    After coming to the party this late, NV would be falling all over themselves to show you numbers proving that it was worth the wait.

    If they don't show their numbers now, they don't have the numbers they want.

  8. #8
    Senior Member
    Join Date
    Jun 2009
    Posts
    2,626
    Quote Originally Posted by JHFolsom View Post
    ...snip... Fermi is so late, that working on the next process and architecture is the only proper way forward.
    Is it the proper way forward? If NVDA can't make a chip with 3billion transistors at the .40 nm process, what makes you think they can create a +3billion transistor chip at .32 or .28 nm process? Most likely the chip would be 6 billion transistors since the numbers have been near doubling each generation.

    As Charlie stated, the design if flawed. And a more reasonable idea woud be for NVDA to retreat from the consumer market this generation and stick to the HPC market were margins are higher, and they have limited competition.

    But what is most likely: They will keep a limited consumer presence with Fermi to keep the Hardcore customer base, and the industry contacts/relationships going, and go all out with the HPC market with telsa (or whatever you call it). Outside of the supercomputer loss, rumor has it NVDA is making its way into gov computers.

    Edit: From very credible IT sources I have, ATi has had the superior product for the past two years, however the only thing that is holding ATi back from great contracts is Drivers and Support.
    -V
    My views and opinions do not represent those of my employer

  9. #9
    640k who needs more?
    Join Date
    Oct 2009
    Posts
    649
    Sorry, i was referring to the proper way forward for ATi.

    Fermi is so late that ATi should forget about reacting to it other than on the price area.

  10. #10
    Senior Member
    Join Date
    Jun 2009
    Posts
    2,626
    Quote Originally Posted by JHFolsom View Post
    Sorry, i was referring to the proper way forward for ATi.

    Fermi is so late that ATi should forget about reacting to it other than on the price area.
    I couldn't agree more, thank you for the clarification.
    -V
    My views and opinions do not represent those of my employer

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
WordPress Appliance - Powered by TurnKey Linux