DegustatoR
Veteran
So those are TUs, OK...
Anyone want to make a size comparision between GT200 and RV770 TUs and SPs?.. =)
AMD: R7xx Speculation
- Thread starter Unknown Soldier
- Start date
- Status
Anyone want to make a size comparision between GT200 and RV770 TUs and SPs?.. =)
Sure but why looks it like a perfect 32 TMU result and not something like 34 or 37 TMUs?
BW and interpolation performance should be high enough.
Z-Fill NV AMDAnyone can tell me how it is possible????
http://www.pcgameshardware.de/screenshots/medium/2008/06/1213877482046.PNG
i think quarl texture pack is really full of heavy shaders...or not?
4xAA 100% 100%
8xAA 20% 50%
Subtlesnake
Regular
But we haven't seen any reviews yet.So a future part is announced (9800GTX+) on the day HD4850 debuts.
It seems paper launches are back in vogue and perfectly acceptable to some hardware sites again?
Anyone can tell me how it is possible????
i think quarl texture pack is really full of heavy shaders...or not?
That looks very impressive considering with Quarls texture pack you easily get up to an average of 700-900MB of VRAM usage. I would have thought the GTX280 with its 1GB of VRAM would take the lead. Neverthless if not mistaken this pack comes with lots of parallaxed mapped textures.
It really depends how you define competition. I was meaning in absolute performance terms. i.e. R300 was the fastest GPU out there bar none. The 4800 series certainly aren't that so if you want the fastest then the 4800 loses to GT200.
Price of course is a completely different matter but depending on how pricing ends up working out, in many people eyes the 4870's primary competition could be the 260. AMD would prefer us to think its competing with the 9800GTX, and beating it at a higher price but its possible NV could say exactly the same thing for GT260 vs 4870.
Again though, it all depends on price at launch.
For the record btw, I don't want people to think i'm being defensive of NV here, i'm made up that AMD are competitive again and i'm hoping the R700 can destroy the 280 at a lower price. I just think the popular opinion is weighing a bit to heavily against NV atm given the actual numbers that we're seeing.
Mintmaster
Veteran
Note that the same page lists 40 TUs.
Also, the way ATI calculated 80 was 16 bilinear (each needing 4 point samples) and 16 point samples, for a total of 16x4+16 = 80. In the diagrams, RV770 no longer has the point samplers. 160 is derived from 40x4, not 32x4+32.
Bouncing Zabaglione Bros.
Legend
So a future part is announced (9800GTX+) on the day HD4850 debuts.
It seems paper launches are back in vogue and perfectly acceptable to some hardware sites again?
These may be Reviewers Edition ie not really very available but good for grabbing review headlines.
In view of Nvidia's recent attempt to sort out its rambling and confusing product line (minimum price restrictions, etc), it seems that the recent price drops to compete with the 4000 series and now a GTX+ is just going to make things even worse. It might even cannibalise sales of the cards above it.
Not only also notice the 8800Ultra being 3 times faster than the 9800GTX. Thus presumably 9800GTX gets hit by its too small memory where it falls off way faster than the AMD cards. Using 8xAA doesn't help for that, neither (and that's probably the reason GTX280 is "slow")?
A lot of credit goes to AMD for throwing two curveballs at Nvidia in succession. :smile:
The guys in console forum will have something to talk about.Lol, I love the first point on that slide
leoneazzurro
Regular
In DX9 against NV30-35 it was certainly.
In DX8 the competition was a lot more fierce.
Megadrive1988
Veteran
Radeon 9700 series launched in 2002, not 2001
Radeon 9700 series launched in 2002, not 2001
Classic!
So does that mean the slides are fake or that someone at AMD really fracked up?
AMD Genuine™.Radeon 9700 series launched in 2002, not 2001
Radeon 9700 series launched in 2002, not 2001
Heh. Nobody left from ATI to catch the error?
Bouncing Zabaglione Bros.
Legend
I think a lot of people may not be considering the high end Nvidia cards just because of crazy pricing. Sure, some people will pay for the very top limit, but a lot will just say that $500+ is too much and will only be looking at the tiers below that. You can build the rest of a decent PC for less than that. The 280 becomes a non-issue, much in the same way a Ferrari becomes a non-issue when you are shopping for your car.
When you start looking around at the more reasonable $200-300 mark, then the 4000 series starts to really bite. When you look at a top end Nvidia graphics card as a percentage of your total PC build cost, and look at what it gives you in performance compared to a 4000 card, the numbers are very hard to swallow unless you have a lot of cash to burn.
There's no doubt that Nvidia can still point to a card above the ATI's range and say "we're still faster", but at a price that is simply not viable when you can spend half on 48xx and get nearly the same performance.
I think Nvidia is going to have to drop prices and take the pain, and ATI still has more room to combat that with a price cut of it's own. From a business point of view, I really think that ATI has "worked smarter" this time around, while Nvidia has "worked harder".
It's lucky that Nvidia has been making big piles of money over the last couple of years while ATI misfired, because it will help them through what I expect to be a sticky period while they transition to a similar multicore strategy behind a competitor that seems to be well on the way.
I did a fast comparison between the posted specs of RV770 from amdzone and RV670 from AMD.
To my surprise, the specs of RV770 are missing:
1) Ring Bus Memory Controller
2) Fully distributed design with 512-bit internal ring bus for memory reads and writes
3) Temporal anti-aliasing
4) Bicubic filtering
5) Inverse telecine (2:2 and 3:2 pull-down correction)
6) Bad edit correction
Also, RV770 includes this:
while on RV670 it looks like this:
To my surprise, the specs of RV770 are missing:
1) Ring Bus Memory Controller
2) Fully distributed design with 512-bit internal ring bus for memory reads and writes
3) Temporal anti-aliasing
4) Bicubic filtering
5) Inverse telecine (2:2 and 3:2 pull-down correction)
6) Bad edit correction
Also, RV770 includes this:
while on RV670 it looks like this:
Heh. Nobody left from ATI to catch the error?
For crying out loud, they can't even work a spell check, much less fact-check.
Even with a compelling product, AMD is still managing to bungle things with its NDA and launch date vacillations.
Cripes, if they had the perfect card able to manage 100 FPS in Crysis on full settings, they'd probably accidentally pack the retail boxes with dynamite.
- Status
