ATi same mistake as nVidia?

I was just wondering since the RV350 is .13 and R350 is .15, is it possible that ATi messed up like nVidia did with the .13 process? and is now just clocking the RV350 lower so it doesnt heat up so much and selling it as the slower card to the R350 which is, for the most part, a R300?

The R300 is already proven so with a lil modyfing they could just use it for the R350 since their "NV30" didnt turn out good either?

:?
 
I'm betting that the RV350 will be the budget card that comes out around the same time the R400 (which will probably also be 0.13 micron) comes out. By that time, perhaps the manufacturing process will have matured enough so as to not cause the problems that have plagued NVidia. The R350 and the RV350 should be quite different chips.
 
Less complicated 0.13 micron process products, including graphics chips, already exist and are being sold. I believe that, for example, the Xabre is 0.13 micron.

By virtue of being less complicated than the nv30, the RV350 experiencing the same issues is not likely.

At the same time, by testing data gained from the RV350, ATI seems likely to gain atleast some information that could be useful for R400 implementation (this will be a complicated chip, and the real question is what issues that chip will encounteer).

Aside from the above, TSMC is also likely to be improving their various 0.13 micron processes and gaining manufacturing data as time progresses, and the later a part is released, the more likely these improvements would contribute to reducing issues.

So, I'd tend to answer "no" to your initial question. I'd also agree that the rv350 has been said to be due to be announced very soon (CeBit, I believe).
 
RV chips are value chips, not flagship chips. The RV350 was always planned as the value replacement for the DX9 series of products... close to the R300 in performance, better than the 9500 products in price. Therefore, .13u is a tool to reduce die size and cost, not a tool to increase transistor count and performance.

RV350 is nothing at all like the NV30 was.
 
demalion said:
Aside from the above, TSMC is also likely to be improving their various 0.13 micron processes and gaining manufacturing data as time progresses, and the later a part is released, the more likely these improvements would contribute to reducing issues.

Do we know for sure that the RV350 is on TSMC's 0.13um process though?
 
Hmmm, maybe it was at TSMC, if we trust the inquirer...... :rolleyes: ...(always wanted to use that icon :) ) This is a Dec 10 2002 article though....

http://www.theinquirer.net/?article=6661

"THE INQUIRER has received several tips recently regarding ATI's R350 and RV350.

The first reiterates that ATI's R350 is being built on a 0.15 micron process. ATI's RV350 mainstream part is however being built on a 0.13 micron process.

What is new is that both the R350 and the RV350 have already taped out with their production partner TSMC, which did not do quite as good a job with their friend Nvidia."
 
ATi already has working RV350 and M10 chips back, and has had for a little while now. 13u went pretty flawless for them except for a couple little issues.
 
Another point: M10 is either exactly or just-about-exactly identical to RV350. Finer process reduces power consumption as well as die size...
 
ATI is using the RV350 as their .13u testbed of sorts, to acclimate themselves with the smaller process. Hopefully, the experience they have gained designing the RV350 will help them with their first "real" .13u card, the R400.
 
Hellbinder[CE said:
]ATi already has working RV350 and M10 chips back, and has had for a little while now. 13u went pretty flawless for them except for a couple little issues.

Would it be safe to say, that the RV350 wouldn't come with a fan ;)

Boy, that would be something :D
 
If Nvidia had major problems with a 125 million transitor chip what will they have when they go up to 185 million and beyond? Like the R400. How close is ATI in taping out the R400? Anyone know?
 
R300 taped out in Feb 2002. If R400 is their summer/fall product then you would expect a similar pattern.
 
DaveBaumann said:
R300 taped out in Feb 2002. If R400 is their summer/fall product then you would expect a similar pattern.

Would love for this to be true, would even love for someone to leak out some specs.. :p
 
mat said:
demonic said:
Would it be safe to say, that the RV350 wouldn't come with a fan ;)

Boy, that would be something :D

well there are already passive cooled Radeon 9700 running at 325/310 available: http://cwsoft.at/images/products/5/3A/c3d/9700prosil_b.jpg

from the shop: Connect 3D Radeon 9700 Pro Silent Edition, 128 MB DDR, DVI/TV-Out, AGP mit passiver Heat-Pipe-Kühlung

Yes, I know of the zalman h/s. I know a m8 that has recently cooled it R9700 Pro with a zalman :)

But, what I mean't. Is if ATi *released* the Rv350, with just a heatsink. Just imagine the picture it would send to everyone?

In the blue corner we have ATi, with just a heatsink.
In the red corner we have Nvidia, with the dustbuster!!

Now you get what I mean :LOL:
 
I was surprised last week when I got a new R9000 from Saphire. While my old GF2 MX 200 was using a small fan (that was full of dust when I removed the card) the R9000 at double the frequency was just using a somewhat large heatsing (more or less the size of the old PIII heatsink).

So maybe if in the RV350 they remove some stuff or if they use a lower clock there will be a fanless versions.
 
demonic said:
But, what I mean't. Is if ATi *released* the Rv350, with just a heatsink. Just imagine the picture it would send to everyone?

In the blue corner we have ATi, with just a heatsink.
In the red corner we have Nvidia, with the dustbuster!!
What picture? The FX Ultra is a high-end part, one that won't see much use. The RV350 is a mainstream part, meant to compete against nVidia's MX line. The products are utterly different.
 
RoOoBo said:
I was surprised last week when I got a new R9000 from Saphire. While my old GF2 MX 200 was using a small fan (that was full of dust when I removed the card) the R9000 at double the frequency was just using a somewhat large heatsing (more or less the size of the old PIII heatsink).
Well, the GeForce2 MX parts should actually have been capable of running without any heatsink whatsoever. The power draw was actually quite low. I believe most shipped with passive cooling in the end, however.
 
Back
Top