Announcement

Collapse
No announcement yet.

The Fermi!

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Well usually the highest priced cards will come first and till xmas the low price ones

    Comment


    • #17
      Originally posted by Ex-Cyber View Post
      Have they actually said anything about when Fermi-based GeForce cards will actually show up in quantity, or is that still in rumor-land? The latest stuff I can find says March optimistically and April/May pessimistically.
      No, I believe all that has been said is Q1 2010, they don't have an exact day let alone an exact month. Anything beyond that is just a guess.

      Originally posted by chithanh View Post
      It is possible that NVidia launches Fermi for USD 500 but they would have to eat a loss for each unit sold, if the estimates for board costs are remotely accurate.

      Regarding power consumption that is a good point. Reports that there will be special Fermi-certified cases and PSUs have appeared some time ago, and various tech sites saying that it will be the most power hungry NVidia chip of all time.
      I have one issue with the enormous amount of power this chip uses, if the unofficial benchmarks are even remotely accurate that I have seen, all I can say is I expected more out of a card using 280ish watts. Of course the amount of speculation and bias out there is pretty ridiculous so the benchmarks could all be BS.

      Theres people who want this card to fail horribly and theres people who want it to completely destroy ATI/AMD. Me personally, I don't care either way, I just want to see if it lives up to the hype. I am probably not going to buy the card, I'm pretty happy with my 9600 GT still, but definitely interesting to watch.

      Comment


      • #18
        Just want to point out that the Fermi architecture was originally designed for gpgpu stuff, and as such targeted a slightly different audience than AMD's latest offerings.
        Of course, I can't say what nvidia plan on doing with it now, only time will tell with that.

        Comment


        • #19
          the latest "official" release date I know was "certainly 2009". I'm not sure if the release dates around march are official or just rumors.

          I find it a bit difficult to get excited over a video we've seen on evergreen before. It's not even a real-world benchmark because it's so tesselation-heavy: nvidia's architecture will use the shaders for tesselation while ATI has dedicated tesselation units. Heavy tesselation will decrease the available shader power on nvidia cards, ATI cards are unaffected.

          So yeah, nothing impressive yet. Let's wait until march.

          Originally posted by Kano View Post
          Well usually the highest priced cards will come first and till xmas the low price ones
          so... a year late?

          Comment


          • #20
            It is late, but there is exactly 1 DX11 benchmark out there that uses tesselation, and thats not even 3dmark just an engine demo that was never used in a real game. Of course it is not that ideal when want to buy a new pc and put in a gfx card that you would have to replace to see all possible features. Till the more affordable cards will be available it will certainly take more time, if it is 1 year late i don't know, but at the end of the year you could expect first "real" DX11 games, which may not look 99% the same as DX10 ones. Then you could also buy a new monitor/tv to watch bluray in 3d. At least vp4 should support h264 mvc which all cheap g210/gt220/gt240 have got. Has ati announced something for their chips yet? I doubt so as "UVD2" is not even fast enough for more than h264 l4.1...

            Comment


            • #21
              Originally posted by Kano View Post
              Then you could also buy a new monitor/tv to watch bluray in 3d. At least vp4 should support h264 mvc which all cheap g210/gt220/gt240 have got. Has ati announced something for their chips yet? I doubt so as "UVD2" is not even fast enough for more than h264 l4.1...
              1) H.264 MVC support is in GT215 chips only, i.e. G240 with VP5, not in G210 for example.

              2) There is no HW decoder generally "fast enough" for more than L4.1. All current HW decoders (including NVIDIA) are designed for H.264 High-profile @ L4.1. Higher levels may be supported, but with compromises in terms of speed or features. NVIDIA won't advertise L5.1 because real-time constraints can't be fully guaranteed.

              Comment


              • #22
                Even though DX11 isn't widely used yet (how could it be, when DX11-capable cards only became widely available this month? (Widely referring to plentiful and cheap mid-range parts)

                On the other hand, even without DX11 the 5xx0-series pretty much own everything else on the market right now. Previous gen cards are already a hard sell and getting worse as 5xx0 prices are starting to go down.

                Fermi right now is set for a spectular FX 5800 Ultra-level failure: late, expensive, power-hungry and potentially underperforming in actual games. Nvidia's slides say that Fermi wins 5870 by 60% in Unigine/tesselation, which is impressive - until you realize that they are comparing a potential $600 card (Fermi) to a $350 card. In other words (speculation):

                (a) there's a good chance that Ati's 5970 is very close or faster than Fermi.
                (b) the difference is likely to be lower than 60% in actual games (this is a tech demo showcasing a very specific part of the pipeline - actual games tend to be more balanced in performance)
                (c) Fermi may be launched at a lower price than previous Nvidia high-end parts (typically $600-$700). With a little luck, it Fermi might go down to ~$500 - still pricey, but a much better value proposition given the 60% (max) performance delta.

                One of the links posted here mentions a late March launch date. Come on Nvidia, we need some competition here!

                Comment


                • #23
                  @gbeauche

                  If there are VP5 cards then this should be marked here.

                  http://en.wikipedia.org/wiki/VDPAU

                  or

                  ftp://download.nvidia.com/XFree86/Li...DME/README.txt

                  VP2 = VDPAU FEATURE SET A
                  VP3 = VDPAU FEATURE SET B
                  VP4 = VDPAU FEATURE SET C

                  So VP5 would be a still unknown VDPAU FEATURE SET D in theory. I do not believe in marketing speach of nvidia, all those mentioned cards usually have got the same engine.

                  Comment


                  • #24
                    Originally posted by gbeauche View Post
                    1) H.264 MVC support is in GT215 chips only, i.e. G240 with VP5, not in G210 for example.

                    2) There is no HW decoder generally "fast enough" for more than L4.1. All current HW decoders (including NVIDIA) are designed for H.264 High-profile @ L4.1. Higher levels may be supported, but with compromises in terms of speed or features. NVIDIA won't advertise L5.1 because real-time constraints can't be fully guaranteed.

                    Just for your reference:

                    Code:
                        NVIDIA GPU product                    Device PCI ID      VDPAU features
                        ----------------------------------    ---------------    ---------------
                        GeForce GT 220                        0x0A20             C
                        GeForce 210                           0x0A23             C
                        GeForce GT 230M                       0x0A2A             C
                        GeForce GT 240M                       0x0A34             C
                        GeForce G210                          0x0A60             C
                        GeForce 205                           0x0A62             C
                        GeForce 310                           0x0A63             C
                        GeForce 210                           0x0A65             C
                        GeForce 310                           0x0A66             C
                        GeForce G210M                         0x0A74             C
                        GeForce GT 240                        0x0CA3             C
                        GeForce GTS 260M                      0x0CA8             C
                        GeForce GTS 250M                      0x0CA9             C
                        Quadro FX 380 LP                      0x0A78             C
                    As mentioned above all these cards carry the same PureVideo engine.

                    Comment


                    • #25
                      Originally posted by deanjo View Post
                      As mentioned above all these cards carry the same PureVideo engine.
                      Or the required VDPAU features implementing H.264 MVC are not available yet, so be the best matching feature set is still marked as "C" and the rest is not advertised... NVIDIA doesn't do paper launch so they will report things when proper drivers are available.

                      Comment


                      • #26
                        Originally posted by gbeauche View Post
                        Or the required VDPAU features implementing H.264 MVC are not available yet, so be the best matching feature set is still marked as "C" and the rest is not advertised... NVIDIA doesn't do paper launch so they will report things when proper drivers are available.
                        Well when feature set c cards debuted Nvidia announced it's capabilities before vdpau even had support for those extra capabilities.

                        Comment


                        • #27
                          All and all I don't think we would see MVC support in vdpau anyways as there still isn't a "legal" way of playing back bluray in linux.

                          Comment


                          • #28
                            Well i have got an old asus v7700 deluxe with shutter glasses, but no connector to use with new vga cards. So the stereo glasses have been introduced very long ago with crt monitors. Currently the 120 hz tft full hd monitors are only "new" thing. The alternative way reduces the effective resolution by 50% and is cheaper however as only pol filter glasses are needed. I see no reason why h264 mvc would not be possible to decode correctly on linux as m2ts. The same way you can use to watch youtube videos too...

                            Comment


                            • #29
                              Originally posted by Kano View Post
                              Well i have got an old asus v7700 deluxe with shutter glasses, but no connector to use with new vga cards. So the stereo glasses have been introduced very long ago with crt monitors. Currently the 120 hz tft full hd monitors are only "new" thing. The alternative way reduces the effective resolution by 50% and is cheaper however as only pol filter glasses are needed. I see no reason why h264 mvc would not be possible to decode correctly on linux as m2ts. The same way you can use to watch youtube videos too...
                              BTW as a side note, a 120Hz monitor or LCD TV really isn't required. Many plasma TV's are capable of doing 3D. The new Nvidia glasses work fine on my Samsung plasma.

                              Comment


                              • #30
                                Originally posted by kyzz View Post
                                I have one issue with the enormous amount of power this chip uses, if the unofficial benchmarks are even remotely accurate that I have seen, all I can say is I expected more out of a card using 280ish watts.
                                Sadly, I wasn't: from the benchmarks I've seen, ATI's shaders have been substantially more efficient than Nvidia's for some time, presumably because ATI are still making graphics chips whereas Nvidia have been trying to produce vector processing chips and hope they could still do graphics. I can understand Nvidia wanting to find new markets for their hardware, but they seem to have put those new markets above their traditional market where they make most of their money.

                                To me this does really look like another 'Geforce FX' if the articles I've seen are correct; which is unfortunate because I was planning to build another gaming/video editing Windows PC later this year and Avid don't like ATI cards.

                                Comment

                                Working...
                                X