ATI Radeon HD 4000 Series discussion

[perkam]

Charlie is saying that the R700 should beat the GeForce 280.
http://www.theinquirer.net/gb/inquir...0-280-revealed

I hope what that says is not true and that the GTX 280 is a monster of a card.

As an ATI fan, I wouldn't mind Nvidia to continue to have the performance crown so that ATI competes on price

Perkam

[Calmatory]

Charlie is saying that the R700 should beat the GeForce 280.
http://www.theinquirer.net/gb/inquir...0-280-revealed

That Anti-nV talk just hurts my brain. Though, at least it is something against the regular anti-ATI talk, but doesn't really help here.

It's all FUD! GT200 > R700 in terms of raw performance(so 3DMark benchers are happy), but performance/price ATI wins.

[Papu]

I hope what that says is not true and that the GTX 280 is a monster of a card.

As an ATI fan, I wouldn't mind Nvidia to continue to have the performance crown so that ATI competes on price

Perkam

nah i would prefer a better cheaper product so Nv has to step up there game!

[AliG]

Nah I think the r700 will spank gt200 in 3dmark (except maybe Vantage, I think Nvidia had a lot to do with that though), we're comparing 960 shaders (minimum, very low possibility of 1600) and 3dmark is extremely shader intensive. As for in game performance, it'll be close, but the gtx 280 will probably have a slight lead over the 4870x2. As big of a leap the rv770 is over the rv670, ATI is still playing catchup so its hard to say at this point



And guys leave the fanboyism out of the thread please, give an honest guess of what you think will happen, there's no point to have a flamewar ruin a 40 page thread

[Thesavage]

Well if we extrapolate our non existant data ie: 1x GeForce 280 is 3x faster than a 3870 and a RV700XT is neary 2x 3870 then the R700 should just about win, thats my 2 cents.

[AliG]

See the thing is, folding means nothing for in game performance, and also due to the law of deminishing returns, except with high detail I don't think the 4870 will be 2x the 3870

[Macadamia]

nVidia probably used a slower CPU for their folding "benchmark". Ahem. As CUDA isn't as CPU intensive as CAL is- for now.

[ownage]

I hope what that says is not true and that the GTX 280 is a monster of a card.

As an ATI fan, I wouldn't mind Nvidia to continue to have the performance crown so that ATI competes on price

Perkam

As an ATI fan, i didn't mind nVdia had the performance crown and ATI the price/performance crown, but I wouldn't mind if ATI would take the performance crown for a change.. This because nVidia was pretty much on top last 2 years.

although the GTX 280 will be very powerful, i don't see any trouble for ATI as nVidia's top model will consume 236watt with all negative details that will give.
ATI will have some powerful cards for a good price, while there top dualgpu card probably can compete with nVidia's best card.

[Macadamia]

As an ATI fan, i didn't mind nVdia had the performance crown and ATI the price/performance crown, but I wouldn't mind if ATI would take the performance crown for a change.. This because nVidia was pretty much on top last 2 years.

although the GTX 280 will be very powerful, i don't see any trouble for ATI as nVidia's top model will consume 236watt with all negative details that will give.
ATI will have some powerful cards for a good price, while there top dualgpu card probably can compete with nVidia's best card.

R700's TDP is rumoured to be 250W, but 2 dice are easier to cool than one, evident with R680 vs R600.

[Philip_J_Fry]

weeee, i am so psyched for both 4870 and GTX 280, the price of 4870 makes it looks very attractive, can not wait for leaked/rumor bench numbers

[Jowy Atreides]

That Anti-nV talk just hurts my brain. Though, at least it is something against the regular anti-ATI talk, but doesn't really help here.

It's all FUD! GT200 > R700 in terms of raw performance(so 3DMark benchers are happy), but performance/price ATI wins.

Are you thinking of the 4870 or the 4870x2? (which is the R700)
because two 4870s vs one gt280 could very well result in nvida being inferior as said.
I don't think it's biased to make such a guess

Charlie is saying that the R700 should beat the GeForce 280.
http://www.theinquirer.net/gb/inquir...0-280-revealed

Yeah, it should. Whether it will is a guessing game for now

[p2501]

weeee, i am so psyched for both 4870 and GTX 280, the price of 4870 makes it looks very attractive, can not wait for leaked/rumor bench numbers

I can't wait, too. Seeing that the cards will be launched soon, I wonder how long it'll take until we see some comparing numbers for both fah and games.

I want to play that freaky tech-demo called Crysis at last.

[Origin_Unknown]

i'm not impressed by the stock 4870x2 cooler.... single heatpipe?

[Macadamia]

i'm not impressed by the stock 4870x2 cooler.... single heatpipe?

2 heatpipes.


I reckon 4870 vs GTX260 is a much closer fight than most reckon, probably 3870 vs 8800GT v2, but this time around the former card IS attractive enough.

2 4850s face off GTX260 as posted earlier, same prices but probably better performance.

And the big question...

2 4870s vs GTX280. It's hard to estimate, but there might be a probability that BOTH are competent.

[Hornet331]

no way the 4850 is up against the gt260 with the current rumored specs. Imho the direct competition of the 4580 will be the G92b

[Calmatory]

Are you thinking of the 4870 or the 4870x2? (which is the R700)
because two 4870s vs one gt280 could very well result in nvida being inferior as said.
I don't think it's biased to make such a guess



Yeah, it should. Whether it will is a guessing game for now

RV770 is not very powerful compared to GT200. Even if it was some 65% of what GT200 is(and some 130% of what G92 is, 105% what G92b is), two of them would result in 130% yes, but problems with CF and scaling makes R700 to be some 90-95% of GT200. This is my guess. :p

[Jowy Atreides]

RV770 is not very powerful compared to GT200. Even if it was some 65% of what GT200 is(and some 130% of what G92 is, 105% what G92b is), two of them would result in 130% yes, but problems with CF and scaling makes R700 to be some 90-95% of GT200. This is my guess. :p

:p ?

My post was very polite.

[AliG]

RV770 is not very powerful compared to GT200. Even if it was some 65% of what GT200 is(and some 130% of what G92 is, 105% what G92b is), two of them would result in 130% yes, but problems with CF and scaling makes R700 to be some 90-95% of GT200. This is my guess. :p

dude no way g92b will be 20+% faster than g92, its just not going to happen. G92b is a direct shrink, look at the r600 to 3870, did you see any big ipc gains? Samething will happen with g92 to g92b, it'll drive down prices and costs+up yields, but it won't do much for performance. You'd need it clocked at something like 800+mhz stock for it be 20% faster than g92 and that you know as well as everyone else that won't happen unless nvidia plans on having low yields again

[zerazax]

Yeah G92b is 55nm so its an optical shrink - it'll be better priced than at 65nm but performance wise, it won't see anything massive improvement wise

[BenchZowner]

dude no way g92b will be 20+% faster than g92, its just not going to happen. G92b is a direct shrink, look at the r600 to 3870, did you see any big ipc gains? Samething will happen with g92 to g92b, it'll drive down prices and costs+up yields, but it won't do much for performance. You'd need it clocked at something like 800+mhz stock for it be 20% faster than g92 and that you know as well as everyone else that won't happen unless nvidia plans on having low yields again

I'm not firing up my calculator, but let's say that G92b will arrive clocked at ( I'm not saying things that originate from any source, this is pure speculation, not even well thought )

Core 800MHz
SPs 1998MHz
Mem GDDR5-3200

Which is feasible considering that a 65nm 8800GT/8800GTS512/9800GTX can run those clocks ( 780 to 800MHz core depending on your luck ) ( except the RAM obviously ) with a decent air cooler ( HR03-GT ).

There's also a chance ( not that much, but still possible ) that nVIDIA has touted the SPs and the MUL is working "properly" on G92b ( it didn't work most of the time on G8x/G92/G94 ), which results in even better numbers.

We just have to wait a tad to find out ( for the info, the card won't be out before August more than likely )

[Calmatory]

dude no way g92b will be 20+% faster than g92, its just not going to happen. G92b is a direct shrink, look at the r600 to 3870, did you see any big ipc gains? Samething will happen with g92 to g92b, it'll drive down prices and costs+up yields, but it won't do much for performance. You'd need it clocked at something like 800+mhz stock for it be 20% faster than g92 and that you know as well as everyone else that won't happen unless nvidia plans on having low yields again

Yeah. I just threw the G92/G92b numbers without thinking really. If G92b is (only)G92 @ 55nm, then with what is nV going to compete agasint RV770? ...haven't heard about low/mid range chips based on GT200. :|

Well, anyway I think that R700 stays somewhere between GTX 260 and GTX 280, leaning towards GTX 280. Though, I am not very sure about how big difference there is between 260 and 280. Heard about 25%, which is IMO rather high to be true.

[AliG]

Nothing, except maybe the gtx 260, but that's probably designed to compete with a 4850x2 (assuming ati makes one), or perhaps gt200 is so powerful it will compete with the 4870x2 and there won't be any competition for the gtx 280 (like what happened with g80 vs r600, the g80 gtx was so powerful that the gts was in line with the 2900xt)

[AliG]

I'm not firing up my calculator, but let's say that G92b will arrive clocked at ( I'm not saying things that originate from any source, this is pure speculation, not even well thought )

Core 800MHz
SPs 1998MHz
Mem GDDR5-3200

Which is feasible considering that a 65nm 8800GT/8800GTS512/9800GTX can run those clocks ( 780 to 800MHz core depending on your luck ) ( except the RAM obviously ) with a decent air cooler ( HR03-GT ).

There's also a chance ( not that much, but still possible ) that nVIDIA has touted the SPs and the MUL is working "properly" on G92b ( it didn't work most of the time on G8x/G92/G94 ), which results in even better numbers.

We just have to wait a tad to find out ( for the info, the card won't be out before August more than likely )

It is possiblefor the g92 gts to run there stock (I haven't seen any non modded gt though, don't know where you've seen that), but the thing is, g92's yields were bad enough on 65nm, I don't see how they'll be that much better on 55nm that they can run stock 800mhz (I predict probably around 700mhz for the 9900gt and perhaps 750, top of 775 for the 9900gtx). Also why would g92b get gddr5? You said it yourself, the reason why nvidia didn't bother with gddr5 for gt200 (aside from more ROPs, but they could have changed their design to work around that) is because ATI would get just about all the GDDR5 in the channel. So they'd be hurting their yields even more to go with GDDR5, I just don't see it happening.

When's the launch date? Is it before or after gt200?

[Hornet331]

When's the launch date? Is it before or after gt200?

We just have to wait a tad to find out ( for the info, the card won't be out before August more than likely )

...

[Calmatory]

I believe it is somewhere near GT200, a bit later I think. So August it is. :P