The russians cant listen to NDAs either.

[otlg22]

Apologies to all the 256 registers are apparently for teh vertex shaders now that I look at it again.. no idea how many are available to the PS units.

[DGhost]

Unfortunatly, until it gets released (or more info posted) we will not know for certain what the core is capable of. The PS core may possibly be capable of handling PS2.0. it may not. or it could be faked in software. or a lot of things. same with bandwidth saving features... we don't know until the NDA is lifted and Haig +others are free to answer questions *grin*

[otlg22]

Page 65 of that MS presentation really makes me think it would be more difficult to not support PS 2.0 if you support VS 2.0.. it appears you could save a lot of time and effort (and reuse functional blocks on the die) by supporting 2.0 of each.. oh well a couple of days to wait and see.. no biggie..

[Rimfaxe]

I am drooling over here.


Now, the only answer I want is, when can I get my hands on one? What price will I have to pay. (or, how mad will my GF be)

Damn, it may cost a lot but who cares!

[nggalai]

Hi otlg22,

it appears you could save a lot of time and effort (and reuse functional blocks on the die) by supporting 2.0 of each.. oh well a couple of days to wait and see.. no biggie..

Exactly--no biggie. I couldn't care less for full, official DX9 support. The features interesting for me such as displacement mapping and Vertex Shader 2.0 are supported (so enough material to play around with), and my bet is it will take something like a year or two at the very least until we'll see true DX9 titles (i.e. games) on the shelves. So I won't really worry about PS2.0 being supported or not--until those features find wide support, the Parhelia will be in my backup computer.

The only thing I AM worrying about, right now, are drivers. I won't get the board right from the start, but wait for this forum to light up like a christmas tree with people's posts complaining about slow-ass drivers and incompatibilities . . . or rather, hope that won't happen. *keeps his fingers crossed*

ta,
.rb

[MrBigglesworth]

So about the HDTV bit......what does this mean in usability terms?

Would I only need to go buy an HDTV tuner and pipe it into the video card and then have HDTV on my monitor?

[roadie]

i am sure everything will be sorted out when matrox change to a .13m process. haig has said that it is possible to mass produce .13 now so in the future we will see a full dx9 parhelia running even quicker

[[GDI]Raptor]

Yes, I Matrox is probably going to release a "refresh" of their chip in the end of this year, start of 2003. I this refr4esh we'll probably se full DX9 compabillety, 0,13micron process, more pipelines?? and other optimalisations.

The main purpose with the release now must be to show everybody that they are capeble of doing it.... I don't belive we'll se anny nV30 before september/october, same with R300

[TdB]

this card wil compare to their future cards, like the g200 compared to the rest of the gxxx series, they will most likely use this core as a foundation for future cards for the next 2-3 years.

And it looks like this core will be up to it.

like someone else said:"imagine the parhelia II max".

[otlg22]

nggalai:

I actually enjoy times like these.. before you can get your hands on a part, figuring out what it can do and how it does it.. kinda like prior to the Athlon release when the new was a buzz.. that time I had some extra info though

[nggalai]

Hi otlg22,

sure thing, much the same here--I love founded speculations. After all, that's why I'm posting here.

It certainly has been a very interesting week-end. I doubt the NV30 oder R300 launches will be able to double that . . .

have a nice day,
.rb

[BuddMan]

I'm sure that the NV30 and the R300 will at least be able to comabt Matrox's new card. For one, I know the R300 will be fully DX9 compliant as they have been working closely with Mircrosoft. The NV30, I'm not sure, but I've never been an nVidia fan anyway

But I hope that Matrox finally comes out on top for at least a little while. I don't think I'll be upgrading my video card until I see the .13 micron chips out yet though.

[UncleSam]

Parhelia cant capable to ete PS 2.0. They requre much more textures, and another execution architecture, and floating point calculations. Hardware implementation of that is MUCH bugger than that we have at parhelia now.

we also cant share execute blocks for pixel and vertex shaders.

parhelia capable run only 1.3 and (!) they run only 5 stages per pixel pipe, i/e/ it needs to use 2 pipes together to run 1.3 shader that have more than 5 arithmetical commands!

only 4 different texture samples can be used in shader.

for 2.0 we need up to 16 samples! and 32 (64) stages per each conveyer!

__________________
/Alexander Medvedev,
Technical Director,
iXBT.com (www.ixbt.com),
Digit-Life (www.digit-life.com).
And low life mole who has no shame.

[P5ycho]

check it out, it's the ixbt "screw-that-NDA" guy right here at our MURC

I believe that choices had to be made to keep the chip from becoming too big, so they chose to keep the pixel shaders less complex, when the 0.13micron upgradetakes place you can bet on full ps2.0 compatibility.

[BuddMan]

Well, if those specs from the russian site are true, then he's right. But the way they have this setup leaves them a lot of room to upgrade later on in the future. I wouldn't be surprised to see the next version be a fully DX9 capable card.