Oh no, GeForceFX to appear on street in April quarter

[UtwigMU]

Article on C-net

The first order of business for NVIDIA is to get NV30 off the ground. Though we picked up product specifics at the launch on Monday, we had a chance to get more detail on the timing of the ramp and remaining challenges yesterday. As expected, NVDIA has chosen not to use low-k dielectric on the NV-30, a decision which involves some performance compromises but also lowers the risk of any manufacturing miscues. With the part now qualified, waiting on system integrators to finish boards is the issue – that clearly is not going to happen until the April quarter.

Management was not willing to answer questions about whether additional versions of NV3X have taped out, although the intention to hit the April quarter was reiterated during our conversation.

[GNEP]

Lifted straight from a recent Merill Lynch report. (me got Bloomberg terminal on desk )

[GT98]

boy Nvidia is seriously stepping on their dick lately...nice to see

[thop]

looking good for ati.

[Greebe]

Originally posted by GT98
boy Nvidia is seriously stepping on their dick lately...nice to see

Just trying to keep in step with Kyle

[UtwigMU]

Pitou taped out. (?)

Low versions of FX did not. (?) Looks like FX won't hold great performance lead over R97000 (benchmarks suplied by nV, that was the Ultra part, 128bit controller ...)

Since Ati will hold crown till April, there won't be massive price drops and I think they'll wait with R300/DDR2 untill FX come.

If Matrox makes it and Parhelia2 is competitive in performance and launched in May-June ... If.

[Indiana]

And I don't really believe the first Geforce FX cards will really ship with 500MHz. More like 450MHz first, then sometime afterwards - when the manufacturing process is better, - the 500MHz parts.
NVidia has a long "tradition" with their final cards not reaching the first announced (hyped) speeds - this goes way back up to the original TNT series which came with 90MHz instead of the announced 125MHz....

This really would give Matrox some time to at least announce a strong ParheliaII card. If it is taking another half of a year untill it actually hits the market this is not so important nowadays...

[Admiral]

Sparkle is rummored to begin delivering their GF FX cards to German retailers in January.

Sparkle: GeForceFX im Januar ab 500 Euro

[DGhost]

Originally posted by Indiana
And I don't really believe the first Geforce FX cards will really ship with 500MHz. More like 450MHz first, then sometime afterwards - when the manufacturing process is better, - the 500MHz parts.
NVidia has a long "tradition" with their final cards not reaching the first announced (hyped) speeds - this goes way back up to the original TNT series which came with 90MHz instead of the announced 125MHz....

This really would give Matrox some time to at least announce a strong ParheliaII card. If it is taking another half of a year untill it actually hits the market this is not so important nowadays...

Keep in mind ATI had this same legacy until the 9700 launch... there are a lot of things that NVidia is doing "right" with this launch so far, certainly a lot better than what Matrox did with the Parhelia... I wouldn't be so quick to issue blanket statements this time around...

[Indiana]

If announcing a card in the first half of November knowing that it will be available at the earliest in end January more like February (some sources alreaady quoting April!! for larger delivering) is doing a launch right, then maybe yes.

And the Clock speeds on the Radeon8500 and the 9700 not only did deliver the rumours but were even raised for the actual cards.
O.K., this is one field where Matrox did actually act wrong. If they'd told the projected clock-speeds of the Parhelia earlier nobody would've expected a Nvidia-crusher. But then maybe the "projected"/planned speed of the Parhelia actually was much higher and was lowered shortly before release of mass production cards due to problems (e.g. like the banding issue and heat / power consumption...)

[Wombat]

If they'd told the projected clock-speeds of the Parhelia earlier nobody would've expected a Nvidia-crusher.

And clock speeds mean <I>what</I> exactly? Nothing.

[Indiana]

With the current similarities in gfx-chip designs (esp. between ATI and NVidia) clock-speeds can give a rough idea on performance - O.K., maybe very rough, but still nobody would've expected a Matrox 220MHz part to beat >300MHz ATI/NVidia chips.

And even if the Athlon(XP) vs P4 performance backs you up, still even that performance-per-MHz superiority of the Athlon would not be enough to close a gap that large.

[UtwigMU]

Originally posted by Indiana
With the current similarities in gfx-chip designs (esp. between ATI and NVidia) clock-speeds can give a rough idea on performance - O.K., maybe very rough, but still nobody would've expected a Matrox 220MHz part to beat >300MHz ATI/NVidia chips.

Since DX and OpenGL determine number of textures per pass, which dictates pipelines, pixelshaders and vertexshaders design and implementation, which influences GPU design.

Since the best (implementing on a given process) solutions are based on the same criteria, GPU design is similiar (more than CPU and P4 (P68) is radically different than Ppro/2/3 (P6) and Athlon (K7)

It's also similiar with cars: Airflow is calculated by same formulae and today's cars look much more alike than in the past.

[DGhost]

A 2 month gap between launching a product and having it on shelves is not uncommon in this industry. Look at Intel. Hell, look at AMD if you wanna see this. it took ATI 2 months between the announcement of the 8500 to when reviewers had them.

it used to be with desktop processors that clockspeed was a good indication of things... that being said its not nearly as important now as it used to be, with different architectures and different programs.

same thing is gonna happen with graphics. the GeForce3 is horribly underpowered as far as shader capabilites... as far as i can figure the 9700 is far underpowered when it comes to shader capabilities, but unfortunately i can't find any concrete numbers as to how many shader processors there are per pipeline... clock speed is gonna start becoming unimportant, as a large chuck of the next generation of graphics will be waiting on the pixel shader units... the cards that can execute more will offer a lot more performance in the next several years..

[Indiana]

We will see.
I don't think that the tactics of the NVidia marketing throwing all kinds of dust on their released benchmarks is NOT a very good sign towards the performance of the GF FX. If the Geforce FX would really trounce the R300, I'm sure they would provide exact details for the benchmarks - so all the websites could test the (available) R9700 with those same settings and tell the world how much faster the NV30 will be, but they deliberately chose to do not.
E.g. the QuakeIII bench: they mention (quite detailed) the system CPU, mainboard, FSB, RAM. But they don't mention the individual detail settings used nor even the used timedemo...
The other test-details are even more vague: take e.g. the very fast looking 1280x1024, 4xAA, 8x aniso Nature demo score of a whopping 49fps, what they "forgot" to tell is which of their countless AA modi they used. If it was Quincunx or some similar thing, the score is not really impressive anymore.

So a real comparision can only start when real production cards are out and I guess that's the intended effect of this maneouver.