R4xx will break Moore's law

[FUDie]

Plus, if the 9500 was so successful, why the 9600?

Easy one. First, the 9600 is faster than the 9500, making it more competive. Second, to meet demand.

-FUDie

Or "cheaper to make, yet sell for the same price".

In other words, a separate die makes more economic sense than ressurrecting yield cast offs.

But again, its all speculation on our part as to whether it worked out well for them or not.

Thanks, Russ, for completely missing my point.

Let's say vendor X wanted to by Y 9500s, but ATI only had Z < Y chips to sell. Problem here, no? The 9600 fills this gap because you don't have to rely on a possibly limited supply of crippled R300 parts. Also, ATI selling the 9600 doesn't prevent them from continuing to sell crippled R300 (or even R350) parts and I believe they are still doing so today.

Each broken R300 chip that is sold as a 9500 is money they (ATI) wouldn't have had if they just threw the chip away.

-FUDie

 

[KimB]

it's obviously really, that R500 would be 400~500 million transistors.

...

ATI, roughly, doubles its chip complexity each generation.

Process technology advancements are slowing down, and will continue to do so until a fundamentally new technology takes hold. Pay attention to those, not past history of transistor counts. I doubt ATI is going to be increasing the size of their chips by the amounts your suggesting in the next couple of years.

 

[StealthHawk]

You'd consider the NV35 a respin of the NV30? The thing contains more transistors, a 256-bit bus, twice the floating point operation performance, ultra shadow technology, and maybe HCT if you want to stretch things a bit, and you consider that a mere respin? R300->R350 is a respin. The R350's differences over the R300 are merely superficial. Sure it contains the shadow technology like the NV35, and it has the "f-buffer", but that's IT. Who's to say the F-buffer wasn't there in the first place and just didn't work right? The NV38 should be nothing more than a tweak and fix of the NV35 to get it to run at higher clockspeeds and possibly get better yields. It's nothing like the extreme differences between the NV35 and the NV30.

Um, since we're treading into unchartered territory here with the conspiracies about R350 fixing broken silicon on R300, what makes it any less likely than NV35 fixed some broken features of NV30 :?

This whole "respin" debating seems ridiculous. Now we're counting the number of added features needed for something not to be considered a respin? Sheesh.

 

[Arun]

What? Now we get respins with added features? Bah... I always thought that if there were any added features or whatever ( and that is, added, not activated because it was disabled originally ) , it wasn't a respin.

This makes me remember the tape-out debates we had a year ago about the NV30


Uttar

 

[PaulS]

Which reminds me - When did the NV30 properly tape out?

 

[KimB]

Which reminds me - When did the NV30 properly tape out?

If I remember correctly, it was in late September.

 

[Rugor]

The 9500/9500 Pro was all about contesting the market segment. ATI needed something they could put in at that price point, and they released the 9500 series while waiting for the 9600's to come on line. Remember, the 9600 Non-Pro is faster than the 9500 Non-Pro, but the 9600 Pro is slower than the 9500 Pro.

Of course the 9500 Pro was the real issue. They were selling fully functional R300 chips on cards for 2/3 the price of the 9700. That hurt the bottom line but they needed it to take that part of the market.

 

[KimB]

Of course the 9500 Pro was the real issue. They were selling fully functional R300 chips on cards for 2/3 the price of the 9700. That hurt the bottom line but they needed it to take that part of the market.

I'm not so sure. I think that the reason for the 9500 Pro was specifically to sell more chips. Improved economies of scale may have made quite a lot of sense financially.

 

[StealthHawk]

The 9500/9500 Pro was all about contesting the market segment. ATI needed something they could put in at that price point, and they released the 9500 series while waiting for the 9600's to come on line. Remember, the 9600 Non-Pro is faster than the 9500 Non-Pro, but the 9600 Pro is slower than the 9500 Pro.

Isn't the 9600 non-pro slower than the 9500 non-pro? From what I remember from reviews, the r9600 was quite a bit slower than the r9500, while the r9600Pro was close to the performance level of the r9500Pro.

 

[Rugor]

What I have been able to find shows the 9600 to be slightly better overall than the 9500 Non-Pro due to the faster core and more efficient memory controller that makes up for the reduced bandwidth.

 

[Pete]

Hmm, I thought the 9500 was faster than the 9600 in B3D's review.

 

[KimB]

I think I'll take Wavey's line and say, read the reviews here:

http://www.beyond3d.com/reviews/ati/rv350/index.php?p=8

Looks like in most benchmarks, the order is (from slowest to fastest):
9600
9500
9600 Pro
9500 Pro

The 9600 beat out the 9500 in the UT Botmatch bench, and in the 3DMark2003 "Mother Nature" benchmark (leaving out FSAA/aniso benches, which were convoluted at times...).

 

[Dave Baumann]

(it would also in Splinter Cell because of the improved PS1.1 performance in RV350)