Specs for r300 and rv250
- Thread starter Tygrus_Artesiaoa
- Start date
Wooh, looks like the cat's out of the bag doesn't it. Also the inquirer (that last bastion of truth!) is the running the following story:
In fact, Nvidia is playing it rather safe – its next graphics foray will use 150 nanometer technology, while sources tell us that its bitter rival, Canadian firm ATI, will risk it for a frisket and go the 130 nanometer full hog for the R300.
I've posted this in another thread but it' seems relevant here too.
In fact, Nvidia is playing it rather safe – its next graphics foray will use 150 nanometer technology, while sources tell us that its bitter rival, Canadian firm ATI, will risk it for a frisket and go the 130 nanometer full hog for the R300.
I've posted this in another thread but it' seems relevant here too.
We do? Forgive me, but I could find no reference to that on there at all.
MuFu said:
Ok, the slides seem credible. The only thing about them is that they are either old (and the dates have been changed) or are a really bad marketing presentation that misses a few key points.
R300 DOES have a 256-bit memory bus. You can look at the termination resistor sets on the Computex board's PCB and pretty much determine that yourself if you are an EE. The trace density is also pretty ludicrous on that board if you look at close-up shots of the memory.
Again... that 350MHz figure is only possible if they have secured 0.13u. I'm still quite doubtful. The details about MPEG code/decoding are confused as well. Apparently, most of the work will still be done using software algorithms. This is very likely for RV250 and pretty likely for R300. Overlay and iDCT assist will be in hardware as before though. Rage Theater 2 is actually not that different from Rage Theater 1, the main difference being that RT2 has audio DACs and supports component (YPrPb) capture.
MuFu.
pascal said:
Aren't MCM's only used for multiple-chip packages though?
I can tell you now that boards based on the first revision drew about 54W, which is 6W over the AGP spec from three rails. The second revision is within spec but I still don't know whether this is because of a process change or refinements in the silicon. If the 350MHz figure seen in those slides is accurate (and more importantly, up to date) then no doubt ATi are confident in having 0.13u fabrication available for launch. Can't really conclude too much from them asking tb to remove the slides though.
MuFu.
P.S. That is an ATi slide template, whether the information on it is correct or not, BTW. I doubt someone could have guessed the layout so accurately.
Ascended Saiyan
Newcomer
If it isn't any trouble can anyone pm me the powerpoint slide,thanks.
I don't think they'd use 400MHz SDR, it wouldn't make much sense when DDR is almost as cheap. Though whether the 400MHz units will be economically viable by launch may be in doubt. I put my money on a 300MHz or 350MHz DDR setup for launch.
Oh and the inquirer have changed their story about ATi using .13u and nVidia using .15u. They've swapped them back around again.
Oh and the inquirer have changed their story about ATi using .13u and nVidia using .15u. They've swapped them back around again.
Doomtrooper
Veteran
Old slides..very old 
Doomtrooper said:Old slides..very old
I agree, they may be old slides. But in that case why were ATi so eager to get them removed from all the sites hosting them?
Doomtrooper
Veteran
Don't forget ATI visits this board, and some have bankers hours
Joe DeFuria
Legend
To prevenet the spread of misinformation....or to prevent the spread of accurate information prematurely.
Typedef Enum
Regular
Did anybody comment on the dates appearing on the slides? The date actually says "June 18, 2002." It was a little difficult to see, but I zoomed in on it...
