Interesting Info from COO of ATI

Would be great, if someone could interview Dave Orton again. To see if he is *still* excited about the R400.

I mean, now knowing what the NV30 is, and speculating what a refresh part it could be in July-Sept time.

What are his thoughts are on the R400.

I guess you have tossed out the entire history of Nvidia refresh products eh? Since when does Nvidia make sweping changes to its refresh product? The Nv30 has *Nothing* over the R300 core without its extreme core clock. I dont think the more advanced pixel shaders ammount to a hill of beans in the real scheme of things. While capable of executing more instructions etc... It seems highly unlikely that they will be at playable rates for anything over what Standard DX9 offers. Especially for the next 12-24 months. I am not saying that whatever improvements to the pixel shaders ATi makes is all that important either. However you do seem to have a misplaced belief that the Nv30 is somehow significantly more advanced than the R300...... Realistically what a Nv30 refresh part would do is add another Vertex shader or two, and increase the core clock again.

The R400 is ATi's next Generational jump. Not a half-way refresh. thats the R350. Nvidia will be throwing their Refresh part against ATi's Generation jump.

EDIT:

I keep reading people say that the R350 is just a faster clocked R300 with DDR-II. This is not the pattern of ATI at all. It is the R350 Folks not the R300A. It has significant additions to the R300. Nor is it going to use DDR-II. Just DDR.
 
Bjorn said:
It is likely that their mid-range is simply not doing the volume as of yet because the 9500 + 9500pro have just recently been added to that market, thus making the 9000pro their only real mid-range product.

The R9500 Pro was available for preorder a couple of weeks ago but now it's removed. The R9000 is available for preorder, preliminary available on the 16'th according to the store.

This is in Sweden though, dunno about the situation in other countries.

Anyway, it's easy to understand that they're not making any money on products that people can't buy (at this moment at least).

I think that right now the US market coupled with the Asian market will soak up much of the demand for these cards. You can be sure that the Swedish market isn't as lucrative as a result of the Sweedish governments high tax rates on technology thus making it a lower priority in terms of marketing. High taxes make pricing the card competitively while still maintaining decent margins, difficult.
 
Hellbinder[CE said:
]I keep reading people say that the R350 is just a faster clocked R300 with DDR-II. This is not the pattern of ATI at all. It is the R350 Folks not the R300A. It has significant additions to the R300. Nor is it going to use DDR-II. Just DDR.

How do you know? ATI's never released a refresh part.
 
Chalnoth said:
How do you know? ATI's never released a refresh part.

Wasn't the RV200 (Radeon 7500) a 'refreshed' version of the orginal Radeon Core? Process change, new memory interface and speed bumped. Absolutely a refreshed part if you use Nvidias definition.
 
The RV250 is a refresh of the R200 core as well. But the difference is that the RV250 is the R200 core cut back by some nearly 20 mil transistors, AFAIK it is about a 40 mil transistor part making it a more considerably cheaper chip to produce resulting in an extremely competitive design. But this is an RV part the V supposedly standing for value and the part is ment for the low end market. It is a "refresh" if you like but not a part that has higher performance then the original core. Personally I see it as the new low end bar replacing the MX series cards and quiet an improvement on the lower end of the market as a whole.
 
Hellbinder,

I posted the A similar post LAST NIGHT, and provided a link.. and not one person replied..

So its news today... but not yesterday eh?

Easy fella. It might not be because no one likes you....don't be so offended. ;) This thread was ALSO started last night, and based on the time stamps, earlier than yours. (I don't recall seeing your post last night when I made mine...)

Chalnoth,

I would be. This will be ATI's first real refresh part.

This is what I don't get about you, Chalnoth. The fact that this is basically ATI's first "high-end refresh" part, should immediately tell you that you shouldn't really be too surprised about ANYTHING about it. We have no track record to reflect on.

And uh, trying to relate what ATI's refresh might be, based on to nvdidia's early refresh track record is, well, absurd IMO.

In any case, you make what are IMO contradictory statements. You say you don't expect R350 to be like the GeForce3-->GeForce4, but do expect it to be like GeForce256-->GeForce2.

IMO, those two situations are almost exactly the same. The GeForce4 is no more than a tweaked GeForce3...effectively, just a faster GeForce3 with rather minimal architectural changes. Just like the GeForce2 GTS over the GeForce256.
 
Sabastian said:
Mr.huang said:
R400, hopefully it have a "16"pipeline and higher clock GDDRIII.

MMMM *cough* that's a bit much isn't it? *cough* :D
My very reliable sources tlel me it will have thea GDDRIII on a nice fat 1024 bit bus, running at 1.6 Ghz for an unprecedented 240GB/s of bandwidth.
Effectve bandwith will have to be measure in TB/s
 
Chalnoth...

All Ati chip designations have differing Architectures to some degree. This can be followed WAY back through their history. Trying to push the idea that the R350 is just a R300 with a faster core is totally silly. a slightly Tweaked faster clocked R300 would not get a 1/2 generation Chip identifier. It would be R300A or R305 or something like that.

While i have no idea what the specific changes and enhancements are, I do know that It offers several New features that are apparently going to shock a lot of people, Its going to include some features that people were not expecting till the next generation.

This information is not that of the Alexsox variety. Just like I have been telling you that the R350 was a real product and would be introduced Q1, likely FEB for several wekks now.

(BTW, in case you did not know i have been a Staff member of Area3D with NitroGL, Bacon, Nathan, and until recently MultiGL for a few months now.... not that my *info* comes from any of them)
 
Mr.huang said:
R400, hopefully it have a "16"pipeline and higher clock GDDRIII.
Basically, you're suggesting an arrangement of 8x2... while i could definetly agree with you there (that does sound reasonable), i'd like to point out that we can no longer refer to TMUS as we always used to... it's no longer just that, whether it's 8x2 or 8x1, that doesn't make much difference anymore, as it's becoming less and less important nowdays...

As a proof to my claim, here is a quote from Digit-Life's GeForce FX article:
If earlier a TMU was a unit including both a mechanism of texture fetch and filtering and an interpolator of texture coordinates, now there are no such TMUs. Now we have two separate units - a set of interpolators and a set of units fulfilling texture fetching by request of a shader. Multiple effects need access to different textures according to equal coordinates or, for example, multiple fetches from one texture according to dynamically calculated coordinates, that is why such approach seems to be justified.
 
Althornin said:
My very reliable sources tlel me it will have thea GDDRIII on a nice fat 1024 bit bus, running at 1.6 Ghz for an unprecedented 240GB/s of bandwidth.
Effectve bandwith will have to be measure in TB/s

My hotmail source substantiates your source. 8)
 
Althornin said:
Sabastian said:
Mr.huang said:
R400, hopefully it have a "16"pipeline and higher clock GDDRIII.

MMMM *cough* that's a bit much isn't it? *cough* :D
My very reliable sources tlel me it will have thea GDDRIII on a nice fat 1024 bit bus, running at 1.6 Ghz for an unprecedented 240GB/s of bandwidth.
Effectve bandwith will have to be measure in TB/s

How about 0.0000000000000000001*10^(-23) micron tech with 20 GB eDRAM having 100 TBit bus?

sounds nice? yeah. why not drop out the old pipeline design and make MultiScalar 512 Bit accurancy Pixel RayTracer with huge parallerlism capable pushing more than 2 fully 24 Super Sampled anti-aliased tera pixels per second on polar cordinate iMax compliant format?

why not?
- propably MS wouldn't support it.

Cider gets on my head... need something to eat...
 
If you cut out a few things that Nappe1 said you can probably get quite good picture on what R400 might be ;).
 
I picked up some info yesterday night... in a club ;) (if you were over there you know what I'm talking about ;)): it seems R350 is just an overclocked R300 with no ddr2.
Dunno but previously I expected ddr2 - well, we'll see... ;)
 
T2k said:
I picked up some info yesterday night... in a club ;) (if you were over there you know what I'm talking about ;)): it seems R350 is just an overclocked R300 with no ddr2.
Dunno but previously I expected ddr2 - well, we'll see... ;)


This would be in line with what I have suspected all along--what's the point in making a significant jump in the spring, and then another significant jump in late summer--when you don't have to? Or, as someone else suggested, R350 could be a value part for notebooks (that might or might not fit in with what was said about R350.) I guess I have to concede at this point that we do have some decent evidence that "R350" exists (not that I doubted you, HB, I was just looking for some confirmation from ATI and this would *appear* to be that--except for the nature of the R350, of course.)

But let's say yields have improved and pipelines get tweaked, etc., and you come up with a better cooling solution (not like nv30's--just a lot more efficient than at present)--I don't think 400MHz-425MHz is out of the question for R300. The whole point of the 256-bit bus is to be able to put off the move to DDR II until you have the GPU power to fuel such a bus and you've given the II technology a bit of time to mature. If they stick to DDR but move to 2.2-2.3ns DDR, the combination would significantly enhance current R300 performance (enough to easily steal nv30's thunder--assuming nv30 actually hits 500MHz in quantity), and then save your .13 micron ~500MHz R400 beastie, coupled with your ~40 gig/sec DDRII/256-bit bus for the late-summer/fall release--at which time that product will easily keep up with and probably surpass whatever nVidia might wish to do with its nv30 fall refresh.

But I'm not placing bets...;) ATI surprised the heck out of me with R300--won't be making glib projections about them again...
 
R350 has nothing to do with the mobile segment... that would be the chip under the M10 codename. They may share tech, but they'll be completely different chips.
 
Back
Top