Matrox: Launch date official - May 14th.

they sure take all the publicity they can get... Now we know the time too.

Quote from Matrox newsletter that I received about 5 minutes ago:
----------------

From: Matrox In-the-Loop newsletter <newsletter@matrox.com>

***************************************************************************


Who: Matrox Graphics

What: The biggest launch since the “Millennium"

Where: http://www.matrox.com/mga/

When: Tuesday, May 14, 2002 9:00:00 AM EST

Why: It’s what you’ve been waiting for


Tell your friends.

 

Finally a solid date . Beyond that , no comment

 

all i want to know is how much of the leaked information is true....because it seems all impressive

one problem about these upcoming hardware rumors is that i don't know to belive it or not...i seem to remember when cursh 12 or the now called nforce was to be released , rumor had it to be a mobile chipset with 4.2 gig bandwidth and dual ddr . Well that turned out to be half true and i suspect who ever's leaking ou information they're adding in features off the top of their heads just for the fun of it

 

The unveiling of this new chip is exciting news to be sure. Still my enthusiasm for Parhelia and all things matrox is tempered by three things:

1. Time to market. Of course I cant wait to see what the 14th brings, but if this is just a paper launch or tech overview, and actual cards wont see shelves until late summer/fall my excitement is going to drop a few notches. A fall timeframe will very likely mean matrox only has a competitive product, and not a product to measure others by. I suppose Nvidia could drop the ball on NV30 or ATI could fumble the R300, but frankly I dont see that happening. And of course there is absolutely NOTHING wrong with MAtrox re-entering the 3d consumer marketplace with 'merely' a 'competitive' product, but I would SO like to see some one rock the industry. I think a Parhelia release, ohhhh, NOW could do that. =)

2. Drivers. Nuff said.

Well ok, here's a little more. Frankly while demoing Doom on Parhelia would be neat it wouldnt hold alot of stock with me. too many vendors these days tune their drivers explicitly for ids software engines. I'll wait and see and how they do over a WIDE variety of games and apps before judging.

3. Product turnover. Is matrox committed to being a 'player' or is this a one trick pony that they will now ride for the next 4 years?

 

Johnny Rotten: afaik, this is completely new core just like G200 was. my suspicion is that they use this core to build their next chip next year. Just like G200 was refreshed to G400. (and was supposedly refreshed once more to G800, but never made to final in the planned form. Just G550 came out which is heavily stripped down version.)

 

Parhelion: the Sun in the middle, two smaller lights on each side. Well there you have it - your 21 incher in the middle, flanked by the two 17 inchers that used to collect dust in the garage.

Heh, first post from a long time reader.

 

Does anyone have any idea what the T & L output might be because I have no idea on how to factor T&L ouput ?
The rumoured spec's dictate of Parhelia having 4 vertex shaders & a core frequency of 350 Mhz so maybe that would be of help to anyone who can factor this.

 

hmm I expect Matrix more than matrox

Matrox have hardware Matrix?

 

I don't think they will hit 350 MHz.

With the DX8.1 instruction set, you need 4 instructions for a full transformation. So with 4 vertex shaders, you can do one transformation per clock.

Assuming 250 MHz, this translates to a peak vertex transformation rate of 250 MVertices/second. This normaly does not directly transfer to triangle throughput, as your triangle setup would have to keep up with that rate. So, basicly, you can't calculate the triangle throughput without knowing the performance of the triangle setup. Further, assuming longer vertex programs than a transformation, superscalarity comes into play making it even impossible to calculate the throughput without knowing the internal details of the shaders as well as the instructions used in the vertex program.

 

I think you guys may be expecting a bit to much from this chip, at least performance wise. I wouldn't be suprised if it offers GF4 performance...

 

double post

 

I have to say, that judging from the statements I've heard from Haig the technical support manager at Matrox Graphics, I think there are gonna be a few surprised faces around here (and everywhere else).

From what I've heard, the GF4 will be no competition. But the card will be very expensive of course. I'm all set to get the "super" gamers version when the cards hit the streets, so it's gonna be interesting.

I also have to say, that I'm almost just as exited about the rumored quality of 2D as I am about the rumored 3D speed of the chip/card. I can't wait!

 

Well - assuming that some rumours are true, it's a pleasent surprise to see what they did in AA and xxx mapping department. Colour space is also a (little) step in right direction, IMO.

 

I don't believe the 2D quality will be a major step forward from G400. I expect the hardware video acceleration to be a leap, tho...

 

I don't believe the 2D quality will be a major step forward from G400. I expect the hardware video acceleration to be a leap, tho...

 

I want to believe, that S3 days are well over and we don't have to be concerned about picture or video quality. However, I wonder, if this card will be that appealing to ehm... Audigy owners...

Looks like I'm talking rubbish... it's late... need to sleep.

 

So what are your thoughts on this recent article on The Register? Is that spec sheet for real? http://www.theregister.co.uk/content/54/25238.html

Regards / ushac

 

Oh well it's all old news to me.......but maybe that's because I was the very first one who leaked info on Parhelia well at least the non believers should apologize I think which they didn't do yet.......

Anyway you can all be glad what I told is now here.......and Gf4 is not an issue it's yesterdays card and don't comes even close

 

A GF4 Ti will outperform Parhelia in a lot of games due to driver optimizations. Matrox themselves have admitted to that. In fact, unless AA and/or aniso are enabled, the GF4s will probably be faster in everything, which means reviewers who bench predominantly without those two settings enabled are going to be Parhelia in a very bad, perhaps even unfair, light.

 

Driver optimisations is like human life.

A old person has a great future behind them.
A young person has a great future ahead of them.