7970: Another Disappointment from AMD

[TheRealDeal]

Did I imagine it or did AMD not claim a couple of years back that they were focussing on the mainstream / cheaper price point (under say £250) rather than bringing out expensive high end cards? I'm sure at the time it was a case of "er, we can't compete with Nvidia in terms of top end performance so lets market ourselves at the mainstream" but having come out and said that it seems a bit strange for them to be offering cards priced at say £400.

Yea that was on single gpu cards at the time. They also said they would use there dual gpu to compete at the high end. Was around the time of the 3870/4870. Looks like there strategy is changing again.

 

[KNiVES]

Did I imagine it or did AMD not claim a couple of years back that they were focussing on the mainstream / cheaper price point (under say £250) rather than bringing out expensive high end cards? I'm sure at the time it was a case of "er, we can't compete with Nvidia in terms of top end performance so lets market ourselves at the mainstream" but having come out and said that it seems a bit strange for them to be offering cards priced at say £400.

Thing is with a new card like this, GTX580 costs £350, 7970 is a fair bit faster so will be at least £400, question is for how long i.e. at what point will NV cut prices on the high end parts to try and remarket the GTX580 as an 'almost as good, costs considerably less' alternative, which could then force AMD to tweak their pricing a little (depending on where the 7950 fits in).

I could have sworn I heard AMD say that, but with their CPUs and not their graphics cards.

 

[Marine-RX179]

Only a complete retard could look at all the reviews and conclude a faster, quieter and less power hungry card is a disappointment.

And only people with more money than sense would think with the likelihood of price premium of extra £100 or more in price is a non-issue.

If everytime a new gen flagship card launch and stack their price on top of the previous gen top end already overpriced card, rather than forcing the previous top end card down in price, it would take no more than a few years before we see single GPU flagship single GPU card priced at 1K...

What could happen is Nvidia take the same approach and stack their flagship card price on top of the 7970, then rather then having the new gen cards at reasonable price level, what we would have instead is with previous gen cards filling price point £100 all the way up to £300 (or 78xx replacing the price point of 6950 and deliver EXACTLY same performance as a year ago, but with a lower power consumption), and new gen mid-top cards filling from £300 all the way up to £600....

I think the thing is extra performance from new gen cards is sort of something that people take for granted, rather than having to pay a premium for...

Yea it sucks for the consumers' end, but if I was AMD, I would probably do the exact same thing, and care more for making money than the feeling of consumers which I will probably never meet.

 

[TheRealDeal]

My mate paid £500 pound for a 8800 gtx on launch, this is nothing new apart from its coming from amd. If this was a nvidia launch no one would bat an eye. The price is not official yet anyway. Keplar is late and nvidia have nothing to counter this card with atm. When they do i am sure prices will fall. Amd have gave us some great price/performance cards over the years, especially with the 5870/50 which whooped the competition until gtx 4 series came along.

Whats more likely to happen is nvidia launch something faster and fill the price point of the 7970 and amd reduce there prices to compete.

 

[Rroff]

8800GTX solidly walked all over the 7900GTX tho (and had a massive jump in hardware features) - if it had only been 25-30% faster than the 7900GTX people would have been just as much up in arms about the price tag.

 

[drunkenmaster]

Did I imagine it or did AMD not claim a couple of years back that they were focussing on the mainstream / cheaper price point (under say £250) rather than bringing out expensive high end cards? I'm sure at the time it was a case of "er, we can't compete with Nvidia in terms of top end performance so lets market ourselves at the mainstream" but having come out and said that it seems a bit strange for them to be offering cards priced at say £400.

Thing is with a new card like this, GTX580 costs £350, 7970 is a fair bit faster so will be at least £400, question is for how long i.e. at what point will NV cut prices on the high end parts to try and remarket the GTX580 as an 'almost as good, costs considerably less' alternative, which could then force AMD to tweak their pricing a little (depending on where the 7950 fits in).

Hence the reason I was saying high prices weren't likely but certainly possible, and the arguments given were woeful, the argument, of new CEO as of not many months ago is a VERY good argument for potential changes in company gameplans.

I would bet a LOT of money that the reason AMD prices are so high on launch is down to a new CEO making different decisions. Potentially good, potentially bad. AMD's main reason for value was increased turnover and revenue which it achieved with great success.

The thing is, if these cards were £300-350, a huge number of people on this forum and many others would get them, or their £200-250 7950 little brothers. At potentially £500 for a 7970(I've seen reviews claim everything from £420 to 500 now, and I'm still not certain they won't be significantly cheaper post 7950/CES timeframe), very few people will upgrade.

AMD's strategy was essentially to make it price wise, worth upgrading consistently not just every few years. At £400 people are far less likely to jump on the next gen that isn't far better bang for your buck, at £200-300, they are.

If it works out, only time will tell, realistically I'd be down right amazed if the 7970 sold as many as a 5870 or a 6970 , I'd be surprised if it came remotely close. The question is will it make more or less money.

100,000 cards at £300 is £30mil, at £450 you'd be selling 66k to hit the same revenue, the problem is sales in just about any market is exponentially lower as price increases.

 

[TheRealDeal]

8800GTX solidly walked all over the 7900GTX tho (and had a massive jump in hardware features) - if it had only been 25-30% faster than the 7900GTX people would have been just as much up in arms about the price tag.

Not if amd still had a card on the market that was over £400 and will likely stay there. I think this is amd's justification atm. They never done this in the past though so i am still surprised at these supposed prices. It would make more sense to me to hit nvidia in there pockets right now with a price similar to gtx580 1.5gb and put them out of the high end until keplar hits. Doing this may also hit amd in there pockets though but i do think amd at these prices are not being aggresive enough.

£350 is the max i am willing to pay for a highend single gpu card.

 

[Final8y]

There are a few things that need to be set straight with the ravings of both sides.

- Southern Islands IS a huge step in the right direction for AMD. The architecture is efficient, performs well and is massively impressive regardless of what NVIDIA fans say while burying their heads in the sand.

- Here's a shocker for you: AMD is behind the technology curve. Sure Tahiti is able to pull ~20-25% (on average) ahead of the GTX 580 but we have to remember that we're comparing it to a PREVIOUS generation Fermi architecture, not a comparable next generation Kepler part. When you look at it in that context, AMD may still be fighting one hell of an uphill battle in 2012 and need to get product in the channel....NOW.

- I know the above statement sounds like absolute insanity to some around here but it's important to remember that Southern Islands was originally supposed to be introduced on 32nm before the process was cancelled. Think of it as a Tick (New Architecture) to NVIDIA's Tock (Fermi v2) and that would have put Tahiti in a great position. Instead, AMD had to release "Northern Islands" at a low price AFTER NVIDIA's refresh and push their new architecture back to a time when it would be competing in a less dominant position against Kepler.

- Additional performance through "optimizations" for GCN? Quite possibly but I really don't think so since the goal sticks for driver dev are always in movement. Plus, while there may be increases in today's games, will you be playing Crysis 2 6 months from now? How about MW3? Arkham City? No, probably not. You'll be on to the latest and greatest game and while AMD may be upping performance on the games seen in yesterday's reviews, time has proven that it is their support for NEW games that continues to lag behind. So "performance improvements through drivers" be damned (IMO) until they can show proper and immediate support for new releases. Remember, NVIDIA's driver dev. is hardly standing still either so once everything evens out, you won't see any additional separation between the GTX 580 and Tahiti-based cards in upcoming games. And you can quote me on that.

- We can't only think of Tahiti from a purely gaming performance standpoint. AMD has built in enough appealing features like ZeroCore Power that there are more than enough reasons to consider a HD 7970 over any other card currently on the market. From a personal perspective, the HD 7970's power consumption alone has caused me to immediately replace a GTX 580 3GB in my main gaming system. It may be winter up here in Canada but even a few bucks every month saved on my power bill will add up.

- The HD 7970 is priced lower than most GTX 580 3GB cards and slightly above reference GTX 580s. It consistently outperforms both. The fact that people can continue to complain about pricing simply boggles my mind.

http://www.xtremesystems.org/forums...ether-As-One&p=5021240&viewfull=1#post5021240

 

[drunkenmaster]

8800GTX solidly walked all over the 7900GTX tho (and had a massive jump in hardware features) - if it had only been 25-30% faster than the 7900GTX people would have been just as much up in arms about the price tag.

Yes and no,

http://www.anandtech.com/show/1967/15

7900 was $500, the 8800gtx wasn't really any more than that. Nvidia to be fair has been around and about the £350-400 mark for a long time, back them the x1900xtx was a joke in pricing, lesser variants of the top end cards were much better priced though, the xtx was closer to 8800 ultra levels of stupidity.

http://www.anandtech.com/show/2116/22

Fear, its only 40% ahead at 1920 and 50% ahead of a x1900xtx at higher res. Its not exactly mind blowing.

One of the single biggest reasons a 8800gtx spanked a 7900gtx was memory, Nvidia had and still has far worse memory problems and always run into memory limits long before AMD cards. 256mb as standard on all but the silly edition was WAY to little for the 7900gtx.

It was also just over 40% ahead of a xtx in B&W 2, again around 50% ahead in BF2(lol at how long BF3 took).

It was also awesome in oblivion, where both AMD didn't fare so well and that likely approached the memory limit even on AMD cards, and Prey as iirc that was opengl and AMD wasn't great with Opengl back then(on later generations Prey ran better on AMD than Nvidia iirc).

I'm not saying the 8800gtx wasn't good, it was, but like any card, there were low points of +40% over the previous gen, and high points of around 100%.

It was flat out not as good as most people consistently suggest it was, it wasn't double the speed, it wasn't the biggest leap ever, it was relatively standard.

 

[TheRealDeal]

I was looking at this the other day. An old anandtech 9700 pro review v ti4600. The 9700 pro was stupidly much faster than the ti with aa and af. Thats a real jump in performance with 5x the performance of the ti 4600 in certain situations. Those days are gone though.

http://www.anandtech.com/show/970/18

 

[Marine-RX179]

My mate paid £500 pound for a 8800 gtx on launch, this is nothing new apart from its coming from amd.

Except back in 2006~2007 the world wasn't in global economic crisis/recession like it is now...people can't even go into work without worrying about being laid off.

 

[TheRealDeal]

Except back in 2006~2007 the world wasn't in global economic crisis/recession like it is now...people can't even go into work without worrying about being laid off.

I know this but dont you think businesses also feel this pressure. Its no secret that prices rise everywhere in recession why would graphics cards be any different. Costs go up so the prices follow. You should here my woman after shopping for the last few years with her this used to be this price now its this price. For me i will always have a price in my mind of what i can afford at the time if its not the highend then so be it.

 

[Alexrose1uk]

I was looking at this the other day. An old anandtech 9700 pro review v ti4600. The 9700 pro was stupidly much faster than the ti with aa and af. Thats a real jump in performance with 5x the performance of the ti 4600 in certain situations. Those days are gone though.

http://www.anandtech.com/show/970/18

The 9700 was probably one of the biggest jumps forward in the last decade or so unfortunately in that sense. Whilst it was the 9800Pro that picked up the glory, the 9700 was the true father in this sense; the 4/FX series cards from Nvidia were pretty dissapointing, so ATI releasing a real powerhouse new design at this time was like a bolt out of the blue.

These sorts of designs coming every now and then only help everyone really; they move things along, and they also focus the competition!

 

[Rroff]

Yes and no,

http://www.anandtech.com/show/1967/15

7900 was $500, the 8800gtx wasn't really any more than that. Nvidia to be fair has been around and about the £350-400 mark for a long time, back them the x1900xtx was a joke in pricing, lesser variants of the top end cards were much better priced though, the xtx was closer to 8800 ultra levels of stupidity.

http://www.anandtech.com/show/2116/22

Fear, its only 40% ahead at 1920 and 50% ahead of a x1900xtx at higher res. Its not exactly mind blowing.

One of the single biggest reasons a 8800gtx spanked a 7900gtx was memory, Nvidia had and still has far worse memory problems and always run into memory limits long before AMD cards. 256mb as standard on all but the silly edition was WAY to little for the 7900gtx.

It was also just over 40% ahead of a xtx in B&W 2, again around 50% ahead in BF2(lol at how long BF3 took).

It was also awesome in oblivion, where both AMD didn't fare so well and that likely approached the memory limit even on AMD cards, and Prey as iirc that was opengl and AMD wasn't great with Opengl back then(on later generations Prey ran better on AMD than Nvidia iirc).

I'm not saying the 8800gtx wasn't good, it was, but like any card, there were low points of +40% over the previous gen, and high points of around 100%.

It was flat out not as good as most people consistently suggest it was, it wasn't double the speed, it wasn't the biggest leap ever, it was relatively standard.

If you look at a broad range of games tho it comes out about 70% faster than the 7900GTX overall and still more than 40% faster than the 1950XTX which tbh was no mean feat that card was a monster.

 

[drunkenmaster]

Yes, in all fairness the 1950xtx was an awesome card. Though again it was a VERy forward thinking card and half way of the step towards shaders rather than "pipelines" and I think it moved on very well over time in terms of performance.

However the 7900gtx wasn't that bad, non AA results are much closer, not because it was poor at AA< because memory was becoming a serious problem.

If the 7900gtx had 512mb as standard and that was the only card people benched the 8800gtx against, it likely wouldn't have been 70% faster on average.

But my point is people hail the 8800gtx on here as the second coming of the mesiah, an unobtainable, never previously or since seen performance jump from the previous generation. It was good, and it was a great card, but for me it wasn't particularly different to most of the "big" generation jumps.

Technically, the 2900xt was while underperforming, brought SO much more than Nvidia to the table architecturally, ringbuses, tessellation, I'm still a tad unclear if the 2900xt supported dx10.1 unofficially, it didn't at launch because there was no dx10.1, and afaik dx10.1 was almost entirely features AMD already supported that were originally in DX10 before it got gutted.

would have course been nicer if MS let AMD use them but that wasn't going to happen

 

[Rroff]

AFAIK it basically had most of the "DX10.1" functionality partially/fully implemented in hardware... and sitting there pretty much unusable increasing the size of the core, heat output and power requirements.

They got stitched up on that one, but I still insist it was too early to bring those features to the table, theres no way high definition ambient occlusion for instance was ever going to be useable on the 2900XT or even the next card after that.

 

[Alexrose1uk]

I can't really justify the cost of grabbing one of these. Am definately looking forward to see what this does to the lower segments of the market though. If the top end model runs cooler, quieter and slightly faster than the older models, then I can definately see this being good for the lower end if the same technology is carried down the range.

I'll be more interested in what appears in the £100-200 range though; I can't really justify over £250 on a GPU anymore, given the reduced playtime I get.
Kepler MIGHT (healthy dose of scepticism here) be a lot faster, but then who's to say given the smaller core size compared to previous NV cards that AMD won't just cut prices across the board to compete with cost like they did previously on Kepler's release, and release a 7990 at the same time.

 

[cancausecancer]

Fermi was hot, noisey and expensive... but it blew the socks off the GTX280/285 in anything shader heavy and most older games to and topped the latest from AMD

Sure it topped the latest things when it came out but it didn't top AMD's year old 5970 at the time.

I excitedly waited for the Fermi. I checked all the hardware sites daily for months for any piece of info. When the benchmarks came out I was underwhelmed because it didn't perform any faster than the next gen geForce was supposed to.

For me the marketing hype was misleading. I no longer like Huang because of it (I still like nVidia though). Cuda kicks azz but I expected the 580 to be faster than the year old 5970 but not just faster, considerably faster because it was Fermi. Unfortunately it wasn't faster so I bought the 5970 instead. I could have gone with the 580 still but I wanted fastest gaming period.

I'm in love with both the 590 and 6990 but I just ordered the 7970 today. It's not the world's fastest card (6990/590 share that) however all of it's power is in a single core and the stats on it make me feel like I have a Fermi with an AMD logo slapped on the side.

My speculation is the nVidia's 700 series top card will be faster than the 7970 we see today but not too much faster due to AMD maturing their drivers. I'm guessing +0-10% and I can't wait 7-10 months to find out. If it was going to be 20% faster then I would wait but I really can't see them pulling that off. I think 10% will be hard enough.

 

[lrlcr]

100,000 cards at £300 is £30mil, at £450 you'd be selling 66k to hit the same revenue, the problem is sales in just about any market is exponentially lower as price increases.

I think there is a lot of this, early adopters will always be early adopters. People who are prepared to drop £300 at launch are probably not hugely significantly greater than those who'd drop £500. (I know it's a lot of money, but early adopters are tend to be the people who aren't prepared for the prices to stabilise, ie. not penny pinchers) They can afford to take a 30% hit on sales and still net increased profit (assuming similar costs) then slash the price when kepler comes out and then move into volume movement similar to that seen with the 6950/6970 (i'm talking volume not what the pricing does) to remain competitive. They've then made however much more they take on the initial higher price over what they'd have taken at the lower price.

 

[HangTime]

Sales aren't strictly exponentially lower as price increases, because sales aren't solely driven by demand, but also supply. As an example, when you are talking graphics cards you likely wouldn't see an exponential increase in sales between say £10 and £20 (arbitrary numbers, you could maybe come up with slightly more realistic ones), simply because very few manufacturers would actually supply a card at the £10 price point. The demand would be there, but the supply wouldn't.

edit: In other words it depends on what the marginal cost of production is, if a card costs £200 to produce then selling even 50k never mind 66k at £450 would yield bigger profits than selling 100k at £300. If you are selling at a higher price point, you don't need as much revenue.

Fair enough when we are talking gpu industry a lot of the costs will not be marginal costs but rather up front R&D, but ultimately it is profit not revenue that the chipset makers and their partners will be chasing.