Official 9800pro specs "leaked" by FIC

RussSchultz said:
MuFu said:
At about 600MHz things will start getting *really* nasty and they'll need a long run in with those Cadence design tools they just aquired.
Eh? The Cadence tools they switched to were ASIC design tools, not PCB design tools.

Nah - I think it's the full suite they have access to now (which includes Specctra etc...). They were using Power PCB before.

MuFu.
 
MuFu said:
RussSchultz said:
MuFu said:
At about 600MHz things will start getting *really* nasty and they'll need a long run in with those Cadence design tools they just aquired.
Eh? The Cadence tools they switched to were ASIC design tools, not PCB design tools.

Nah - I think it's the full suite they have access to now (which includes Specctra etc...). They were using Power PCB before.

MuFu.

"The agreement is initially for three years and includes a broad range of Cadence technology"

http://www.cadence.com/company/pr/012003_ATI.html
 
The only area where ATI might not be able to compete is in the $99 low end market. With Nvidia having a directx 9.0 card vs ATI's budget DirectX 8 cards.
 
Well according to xbitlabs the Radeon 9800 will be launched tommorow.

The upcoming RADEON 9800 and RADEON 9800 PRO that will be launched on the 5th or on the 6th, both feature dual 400MHz RAMDAC, 165MHz TMDS transmitter, 8 pipeline architecture, 256-bit memory bus, 512Kb of serial flash ROM, AGP 8x/4x/2x compatibility, integrated TV-out controller (up to 1024x768) and HYDRAVISION technology to support for simultaneous dual displays and DirectX 9.0/OpenGL support.

It was said that the RADEON 9800 will be clocked at 325/620MHz for core and 128MB of DDR SDRAM memory. The RADEON 9800 PRO will probably support 256MB of DDR memory and will be clocked at 400/920MHz. Some sources indicated that there will be 375/750MHz versions of the product, while the 400/920MHz clocked graphics cards will either be limited edition or will be available later. To tell you the truth, the latter is more likely, as the VPU itself supports DDR-2M memory, while no memory maker announced production of such chips. On the other hand, Hynix announced 1000MHz DDR SDRAM memory products in late 2002 (see this news-story), thus, such chips should be available by now and ATI may utilise them in order to create the most-powerful graphics card on the market. Remember that this information should still be considered as unofficial.

What also seems to be important is that the RADEON 9800 and the RADEON 9800 PRO will be made using copper 0.15 micron technology and will have a number of advantages over the previous-generation RADEON 9700 family.

According to our sources, the most important innovation of the RADEON 9800 is the F-buffer technology that supports pixel shader programs with unlimited number of instructions. We will compare this feature of the RADEON 9800 as well as the GeForce FX 5800, which also boasts with infinite number of pixel shaders instructions, once we have both in our lab. Another pair of innovations among the SMARTSHADER 2.1 architecture is Multiple Render Target and shadow volume rendering acceleration support.

Another couple of improvements are said to be introduced under SMOOTHVISION 2.1 brand-name. ATI is going to again optimize its FSAA and anisotropic filtering patterns to offer you even more speed and quality.

There may be other minor differences between the RADEON 9700 and RADEON 9800 VPUs, but I think you will have a great opportunity to learn more about them in one of our upcoming articles.

http://www.xbitlabs.com/news/story.html?id=1046790346
 
So whats up with the 512 kb Flash?
Those Flashes I have read about isn't exactly fast compared to DDR RAM (rather big latencies). And it’s the first time I have seen a Flash as a feature for a graphic card. (Or have I just missed it?)
 
Tokelil said:
So whats up with the 512 kb Flash?
Those Flashes I have read about isn't exactly fast compared to DDR RAM (rather big latencies). And it’s the first time I have seen a Flash as a feature for a graphic card. (Or have I just missed it?)
I bet this is just the standard 64KB graphic card bios, nothing really revolutionary ;) .
 
UncleSam DL iXBT said:
xlabs didnt say any new simply digest known leaks :)

Actually, this is the first leak I've heard that mentioned changes to AA and Aniso algorithms.

And I think somewhere nVidia claimed that the FX architecture was "essentially unlimited" in terms of instructions, and they more or less "arbitrarily" imposed instruction limit counts on the products. (For example, the QuadroFX apparently imposes twice the instruction limit as the GeForceFX variant).
 
mczak said:
Tokelil said:
So whats up with the 512 kb Flash?
Those Flashes I have read about isn't exactly fast compared to DDR RAM (rather big latencies). And it?s the first time I have seen a Flash as a feature for a graphic card. (Or have I just missed it?)
I bet this is just the standard 64KB graphic card bios, nothing really revolutionary ;) .

As I said before, that was a PR document.
Imagine the agony of a PR-bunny who can only add "512 kb of Flash memory". Damn, that must have hurt. Which is odd - it would make sense if the R350 was simply a faster, bugfixed spin of the same core and this was the only additional feature bullet they could think of. But that doesn't quite jive with FIC offering both the A97 and A98 at the same clocks at the same time. Unless, of course, they have R300 inventory or are contractually obliged to sell a certain quantity.

Hmm.

Entropy
 
Or maybe ATI has made too many R300's and needs to "dump" them, so their cost to FIC will be considerably less than the R350.
 
Well that page has been pulled down. So either the information is incorrect and FIC is correcting it (one wonders why they didn't just make the simple HTML change if that were the case), or it is indeed correct, but ATI didn't want the beans to be spilled before their announcement this week.

I'm willing to bet on the latter.. ;)
 
Joe DeFuria said:
Actually, this is the first leak I've heard that mentioned changes to AA and Aniso algorithms.

somethimes better to read news posts all to end even at theinquire ;)

Whats they say two+ days ago:

Both have improved adaptive anisotropic filtering, improved memory controller, six times multisampling and up to 6:1 frame buffer lossless colour compression when in MSAA mode. Both have improved HyperZ III+ technology with 8x depth buffer compression.
 
Veridian3 said:
TB,

Care to post an edited ati2mtag.sys so we can all try it? Or is this functionality only in beta drivers that you cant pass on?

Thanks

Stu

Works only with the latest beta, I hope ATi releases new drivers soon then I will retest this issue and post an article an the mod. files.

Thomas
 
Back
Top