Announcement

Collapse
No announcement yet.

Buy GTS2 or wait for G800 ??

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

  • #61
    I haven't tested it with a GeForce2, but SOF is pretty darn slow on a GeForce DDR at 1200x1024, 32-bit. I've tried with the Guillemot's "pre-overclocked" DDR board, a PIII clocked at 800 MHz, and Raven's "Timing" demo. I got an average framerate of 34.5, and that's a little pokey for a shooter. The map used is almost entirely snow, so I doubt it's the most "texture-packed" part of the game.

    The game is based on the Quake 2 engine and looks it, no matter what the resolution. I would think the game would play a lot better at 1024x768, 32-bit, which gives framerate average of over 20 FPS better.

    I'm not saying the G400 would do better. I'm sure it wouldn't. Maybe the GeForce2 will do better, but I doubt 1280x1024/32-bit is going to be the "sweet spot."

    Paul
    paulcs@flashcom.net

    Comment


    • #62
      Roughly in response to Mr. Peanut but also to Cypro:

      I do lots with my machine (in order of time spent (roughly)):

      Web/browsing, Photoshop, Dreamweaver/Flash, Pagemaker, Quake3/UT, Word, DVD.

      I play a LAN game every few weeks (at my house to I don't have to lug my 19 incher everywhere!) and enjoy Q3/UT, but for the other 5 days of the week, and non-LAN weekends it's not Q3 (though once I get a fast connection it may well be ) - so I'd rather have a Matrox card in my machine.

      I've got a TNT2 just now which I picked up cheap at an clearance auction - it's a generic-ish card which nobody else knew what it was (and I got to test it to see if it worked first too). I payed a fiver (£5~$8) for it and saw other people pay ten times as much for broken Voodoo2's! I'd use my G200 for 2D if it was a 16Mb one but I do like 1600x1200 @ 32bit so it has to be the TNT2 - even with the poor quality - it's still acceptable.

      Back to topic though, and I'd like to have a G400 but it's not worth the £100~$150 or so over my TNT2 but the G450 might be and the G800 would be even nicer. And after all, the guy was comparing the G800 against the GTS2 - the G800 may not be faster that the GTS2, but it may well be faster that the GeForce3 (or with NVIDIA's release schedule GeForce10 ) - what is certain though, I've said it before and I'll say it again (though with the length of this post you won't want me to be saying anything twice ) - the G800 will be doubly excellent at everything the G4X0 is excellent at - and I'll be wanting a G800 when it comes out - Q3 performance winner or not!

      Paul.

      PS: I see you've fixed your glaring error Mr.Peanut hehe

      [This message has been edited by Pace (edited 06 May 2000).]
      Meet Jasmine.
      flickr.com/photos/pace3000

      Comment


      • #63
        You mean about the G400 having the 800mega pixel fill rate.

        hehe

        I don't understand how Paul was having slow downs at that res for Soldier of Fortune. I for one play Unreal Tournament and Soldier of Fortune at 1184x968 (messed up res, I know) but at 1280x1024 my screen does not look right.... Anyhow. You can't use Time demos to give you scores on FPS games. Because not everyone plays with 40,000 people at a time.. usually it is like 10 people and have never suffered any slowdown what so ever... What CPU are you using Paul ? ?

        Comment


        • #64
          Oh I see, you're using the highest fps not the average fps, I'd love to see the lowest it drops to
          and how about when FSAA is enabled, oh wait, the Geforce crashes 75% of the time with FSAA, never mind
          Canadian Royal Matrox Police
          Check the sys!

          Comment


          • #65
            Mr. Peanut, I've tested the board with a CPU clocked at 800 MHz.

            You *can* test shooters with timedemos. It is very common to test shooters with timedemos. I used the "Timings" demo, which, from what I understand, was created and used by Raven to test the game on a variety of platforms. It does not simulate a "death match" scenario, but instead simulates single player game play. Again, it is not very demanding. This is not at all like the Q2 Crusher demo.

            The "Timing" demo doesn't scale particularly well from 800x600 to 1024x768/16-bit. The results are pretty much the same. You do see a drop-off in FPS from 1024x768 to 1280x1024, and you should expect one.

            The decreases in framerate are much more substantial after increasing resolution when the color depth is set to 32-bit. If you don't see a slowdown, your CPU may be the bottleneck. You are, to a certain extent, measuring the capabilities/limitations of your CPU and not your GeForce.

            Paul
            paulcs@flashcom.net

            [This message has been edited by paulcs (edited 07 May 2000).]

            Comment


            • #66
              To Cypro,

              If you decide to go with the GeForce2, I would wait a couple of weeks. Right now, the only GeForce2 available is the Elsa board, and there are serious questions amongst nVidia users about just what was Elsa's role in the board's manufacter.

              A number of people with the board suspect it was manufactured by a company named VisionTek, and that Elsa rebadged the board to get them out before anyone else did. These boards are currently priced at anything from $350 to $400 US. If people's suspicions are true, you'll be paying top dollar for an Elsa and getting a VisionTek.

              To make matters worse, there are *rumors* that Elsa ignored nVidia's street date and shipped the boards with the "leaked" 5.16 Detonators instead of the intend 5.18 shipping drivers, and that nVidia is extremely irritated with their once favorite OEM partner.

              Nothing has been substantiated yet, but I'd still sit on my hands until this is cleared up.

              Paul
              paulcs@flashcom.net

              Comment

              Working...
              X