Announcement

Collapse
No announcement yet.

New Matrox Chip

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

  • New Matrox Chip

    Does anybody know when the new matrox chip (replacing the G400) is going to be released or is likely to be released?

  • #2
    Late Spring/early Summer time I would guess. Are you in a hurry to upgrade?

    [This message has been edited by RoGuE (edited 01 February 2000).]

    Comment


    • #3
      Four words:

      NO, only Matrox knows.

      ------
      Btw, why do you want to know??
      ------
      And: You want a new Gx00 out every 1/2 month like a Geforce???
      ------
      Jorden.

      ------------------
      Got bugs?? Mail Maggi...

      Jordâ„¢

      Comment


      • #4
        LOL, nice one Jorden.

        1/2 a month? have n*idia slowed down then

        Comment


        • #5
          You mean it was every 1/2 a week??
          Jordâ„¢

          Comment


          • #6
            This may violate my NDA, but I am so excited I can't keep this to myself.

            I've been secretely beta testing drivers for Matrox on a prototype board, based on their next chip. They marketing guys haven't named it yet, but intenally it's been using the code name J'Blome.

            The final configurations haven't been decided upon, but single and multi-processor units are being planned. My proto board has 16 J'Blome chips. They are .18 micron, and run at 800 MHz. Each chip contains 4MB of full speed L2 cache, to go with 32KB L1 cache.

            There are 37 million AlGaAs transistors with gold interconnects on each chip. The number is so high because in addition to texture and lighting, they have included a number of "secret" features, that they will be able to enable in random driver updates, and market it as an entierly new product.

            My proto board also has 128 MB of RIMMS. For each processor.

            Power consumption is a bit of a problem. Matrox plans to provide an external power solution, to take some burden off the case PS. My proto board has a dongle that connects to an external box, that takes 230V 3-phase and converts it to +12V and +5V DC. The box is rated for 60 Amps in each channel, so it is about a 1000W PS.

            The trouble right now is that the graphics acceleration is so far ahead of the CPU, that most of the time the card is waiting for the computer to do game physics calculations. To supplement the computer, my proto board has two Flip-Chip sockets, and a special chipset designed for Matrox my Micron. The idea being that this card will actually off load all game physics to the board, freeing up your computer's processor to run Windows.

            As it stands, my protoboard, outfitted with two Coppermine PIII 700's, is turning out 917 fps in Q3A in demo001, but bear in mind these are, like, alpha drivers! Can't wait to see what it'll do when it hits production next week. Oops, I shouldn't have said that.

            Later,
            Charles
            System: P4 2.4, 512k 533FSB, Giga-Byte GA-8PE667 Ultra, 1024MB Corsair XMS PC333, Maxtor D740x 60GB, Turtle Beach Santa Cruz, PCPower&Cooling Silencer 400.

            Capture Drives (for now): IBM 36LZX 9.1, Quantum Atlas 10KII 9.1 on Adaptec 29160

            Comment


            • #7
              Me thinks a little bullshit flying around here.
              By the way I've just got G800+ it's a hungry beast eats Voodoo cards and nridia chipsets for breakfast lunch and supper then goes down to the pub and beats up the Bouncers.
              Chief Lemon Buyer no more Linux sucks but not as much
              Weather nut and sad git.

              My Weather Page

              Comment


              • #8
                And it produces sooo much heat that you can only run it in Antartica, with the box outside and the case open. And even then, it has the nasty side affect of melting the polar ice caps and flooding the world.

                All I want to do is 3D gaming at 1600x1200x32 with butt kickingly highly detailed textures. Is that too much to ask? More speed...more power...more polygons...more fill rate!!! Bwwahaahahaahahhah!

                Matrox's product rollout is not as often as others, like nVidia, but when they come out, they're worth it. Seeing that I got my G400 7 months ago, they've probably got their next generation part taped out, sampled, and on prototype boards. Give them another few months to develop drivers and QA the thing, and then ramp up fabrication.

                Comment


                • #9
                  strange thing moreau, i have the exact same setup, except i only have one p3 700, and i'm getting 1200 fps in quake 3. sounds like a problem with windows 2k.....

                  hehehehehehehehehehehe, matrox is gonna kill us for telling everybody about our alpha cards.....

                  keep it fun!

                  Comment


                  • #10
                    Thundrchez: I'm no EE, but i didn't think they still used tape... !).

                    nosuchluck: my problem is, my Asus P3B is all filled up with PCI Raid Controllers. I had to put my SBLIve in PCI-1, so it's sharing an interrupt with the J'Blome card. I'll have to see if I can do better...
                    System: P4 2.4, 512k 533FSB, Giga-Byte GA-8PE667 Ultra, 1024MB Corsair XMS PC333, Maxtor D740x 60GB, Turtle Beach Santa Cruz, PCPower&Cooling Silencer 400.

                    Capture Drives (for now): IBM 36LZX 9.1, Quantum Atlas 10KII 9.1 on Adaptec 29160

                    Comment


                    • #11
                      Moreau that was too much! LOL
                      Celeron 566@877 1.8V, 256meg generic PC-100 RAM (running at CAS2) Abit BH6, G400 16meg DH@150/200, Western Digital Expert 18gig, Ricoh mp7040A(morphed to mp7060A) Pioneer 6X DVD slot load, Motorola Cable Modem w/DEC ethernet card, Soundblaster Live Value Ver. 2, Viewsonic GT 775

                      Comment


                      • #12
                        For who's sick of Matrox's g400





                        [This message has been edited by andrei (edited 01 February 2000).]

                        Comment


                        • #13
                          moreau:

                          tape-out = semiconductor speak for finish the chip design and send the data to the reticle manufacturers. OK, I can't make a sentence without using technical terms. In the old days, we used to send them the data on tape. Nowadays we use the internet and ftp it directly to them. Regardless, we still make an 8mm tape backup of the design for our own archives.

                          reticle = a mask used in semiconductor fabrication. Once you get the reticles for all of the layers in the design, you give them to the fab and they make the part.

                          Comment


                          • #14
                            Well Thunder, you're not quite as antiquated as some of my professors.
                            For them, taping out was when they applied the tape to form the dies for etching
                            Gigabyte P35-DS3L with a Q6600, 2GB Kingston HyperX (after *3* bad pairs of Crucial Ballistix 1066), Galaxy 8800GT 512MB, SB X-Fi, some drives, and a Dell 2005fpw. Running WinXP.

                            Comment


                            • #15
                              Compared to some of the guys at the office, I'm the new kid on the block. They've given history lessons about how back in the ol' days they got out their xacto knives and were just a cutting away. We've still got some of those tucked in a corner.

                              Comment

                              Working...
                              X