I am glad to here this but i have a question, So this wont be a 5870? So that means the high end model doesn't come out till gt300?
Printable View
I am glad to here this but i have a question, So this wont be a 5870? So that means the high end model doesn't come out till gt300?
Hahah I can't see me being signatured, as I am nowhere near A-List Xtreme Forums celebrity to be signatured on here....but what we all need to realise is that all the ATi and nVidia hype is taking the attention away from the Blue Giant.... beware those who awake the sleeping Giant, as I hear that Larrabee is going to bring something different to the table...
nVidia WILL provide the fastest SINGLE GPU solution, however I can see ATi having the fastest MULTI GPU solution (however as we all know MultiGPU = praying to the driver Gods and serving penance with Microstutter..)
John
Sure, Nvidia will probably hold the fastest single-gpu solution once more. By looking at die size (rumors) that should be quite obvious otherwise Nvidia really has failed. But AMD will be performance leader with their X2 and have best bang for buck at most price segments (or at least be the once cashing in with their smaller die strategy).
Exactly, perhaps ATi should (if they really wanted to get the ball rolling), ditch their failing CPU arm and concentrate solely on Graphics Cards. The HD48xx series are fantastic GPU's... especially for the price, I am not too impressed with the HD4890 though as in the UK it is more expensive than a HD4870X2 (which I might be mistaken is still the fastest card out there? albeit a few games the GTX295 has a slight edge?)
Actually I am slightly wrong, the price of the HD4890 has fallen in the UK quite a bit since I last looked, HD4870X2 = £240 and the HD4890 = £232 for the Atomic Sapphire card and £180 for a standard card. :)
Now... are they going to do a HD4890X2 4GB card?
John
Yes, of course.
For the sake of last year's graphics division success, AMD should quit producing an entire x86 platform and scratch their long-term plans of creating CPU+GPU chips (which was the reason they bought ATI in the first place).
This way, Intel may get a 99.9% share of consumer CPUs and we can all live happily ever after with a 3GHz Core i7 for 1000$ each for the next 10 years.
Makes perfect sense.
Not sure why you think things are going to happen in the exact same way they did last generation. RV770 definitely smacked Nvidia in the face but the surprise factor is gone. The same trick won't work twice. Right now they know/think ATI is going small and fast again. The surprise would be if ATI reverses direction and goes big die again.
Also, while there is a correlation between cost and die size it isn't a linear correlation. There are other variables at play - Nvidia and AMD don't pay TSMC the same price per wafer. So that could explain why Nvidia can afford to counter-slash prices even with larger chips. Still they can't enjoy having to do so. But they pretty much have no choice but to go with a larger die given their GPGPU centric architecture. And this talk of MIMD like functionality is just going to make it worse - unless game engines start to make heavy use of more general computing algorithms where stuff like that matters.
That would be a bad surprise.
But ATI seems to be on a roll, and I'm expecting the similar situation to occur once more. If both companies enter the next round (e.g. dx11) with no fundamental shift in strategy, what else of an outcome will you predict?
For the record, nvidia will probably still hold the performance crown as well as consumer recognition. But I have a feeling that reviewers are shifting their perspectives from high end to what dominates the value segment, which is reflecting most consumers desire. So I think ATI will come out on top in the end, building on their previous 4xxx success. Of course, they will probably lose due to really insufficient marketing and not pushing enough to position their products (they need to send ads to tvs, magazines, whatever).
Well, for starters it was what they have been touting for the last year or so. Secondly they did great with RV770 and why not continuing a winning concept? AMD is clearly going for bang for buck in CPUs and GPUs and aiming more on mainstream than ultra high end. With GPUs they can reach high end with the X2s but with CPUs that's simply not possible.
Even though the prices on wafers may quite different it's still safe to say that ATI/AMD has the edge in ability to cut costs.
I wouldn't say that. Why wouldn't we want a bigger RV770? There's nothing inherently good about a small die (note that power consumption of RV770 is pretty much equivalent to the much bigger GT200b).
Sure I expect them to follow the same strategy as well but AMD doesn't operate in a vacuum. Each company's success is contingent on what the competition is doing. So while RV770 looked like a bang/$ winner vs Nvidia's ridiculous GT200 pricing it won't look that way again unless Nvidia makes the same mistake. That would be hard to do considering AMD is shooting first this round. So unless AMD pulls another R300 and totally knocks it out of the park we shouldn't see the gross mispricing on Nvidia's side that made RV770 look so good at launch.
ATi's strategy to create smaller dies instead of large, monolithic ones seems to be working; as such I see no reason for them to change their strategy.
As for power consumption, their load figures (pulled from a survey of TR and hardwarecanucks) shows that they consume less while providing 'similar' performance. This is due to their tauted increased efficiency due to design.
Quite ambiguous. Does this mean that it will occupy the same spot as the 4870/4890, I thought that was considered highend, besides the x2? When I first read this line the first thing that popped into my mind was evergreen x1<x2<x4. But after reading it again it could mean anything.Quote:
The Evergreen card with a dual slot cooling solution is, according to Huddy, neither an entry level nor a high-end product - it is supposed to be part of the product array above 100 USD.
Of course Nvidia won't price their cards as badly as they did with GT200 but what if they have no other choice? They can't keep losing money. AMD have kept the prices on 4*** below those of Nvidia at all times making sure they were the most beneficial bang for buck even though it's been a very close race. Still they've managed to make money meanwhile Nvidia lost $100M wasn't it?
The main key for AMD should still be to win over some OEMs, that's where the cash is. Will be nice to see Q2 results and what impact PhII has.
There is no point to make large chips atm since TSMC 40nm tech process is so screwed, and even for a small chip like RV770 yields are just way too low. On the other hand, CF / SLI scaling for 2 cards is nearly perfect, so making dual GPU cards makes perfect sense to me.
I want to wait for this if this is not rv840(i cant remember the codenames) and its the high end rv870 and not the x2 solution
My interpretation is that it is meant to replace where the 4850 was introduced.
The different market segments/price points/TDP targets AMD/ATi is trying to hit is, by their definitions;
+$500 enthuiast 250w
$300 performance 150w
$150-$200 mainstream 110w
<$100 value (no TDP defined ~ <80w)
then IGP
My thoughts on a 40nm lineup-
RV830 replaces RV770/RV790 in the under $150 range.
Lower RV870 takes the $200-$250 range.
RV870 takes the $300-$400 range.
RV740 keeps the under $100 while RV810 replaces RV730 in the under $70 range.
it would be sick if this gave about 4870 performance, and used 1/5th the power to idle, and a little over half the power load, and cost 150$ (sure for dx10.1 it may not be the best perf/dollar, but it will sure kill it in dx11, when we have a way to benchmark it)
im gonna be happy with my 4850 until i see a dx11 game i like, and just get whatever 200$ buys
Isn't the 4870 already < $150?
Sure that can happen if you assume best case for AMD and worst case for Nvidia.
Yeah it looks like Phenon II is making a run so they can definitely capitalize on that to get AMD GPUs into those systems as well.Quote:
The main key for AMD should still be to win over some OEMs, that's where the cash is. Will be nice to see Q2 results and what impact PhII has.