• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Polaris architecture – GCN 4.0

Stanners said:
I do not think they will price the top tier low, I do think they will be aggressive with the mid/low level as that is where they can claw the market share back and make money with volume.

Still I would like a cheap high performing Vega :)
Click to expand...

Thing is, there can't be a massive price gap between the 480 and the 490.

It may well be that they are getting rid of the Fury sitting above the usual range of cards (470/480/490), and go back to the 490 Vega being the top single-chip card, with the dual chip cards above that and nothing else (as they used to have).

I don't think the Fury and the Nano sit very well in the product stack. As a short term patch between the end of 28 nm and the start of 14 nm, and a pipe-cleaner for HBM and the interposer, they make sense, but with 14 nm here, I can't see the reason to keep any 28nm chips, and they do seem to be disappearing from stock.
 
Last edited:
Seanspeed said:
????

Nvidia have made a habit out of making the cut-down x70 card a MUCH better value for a while now.
Click to expand...

I don't doubt it but I'm looking at the specs and I can't see it being too close to a 1080.

I could be off by about 5%, considering the 970 has 80% of the floating point performance of the 980 but only 15% slower in games. Best scenario for the 1070 in this case could be 20% behind 1080 or slightly ahead if the 980ti.
Given this and if the 480x results are accurate the 1070 would be about 15-20% ahead of 480 at a higher price point.

At the end of the day you don't always need maximum performance if you can buy something to cover your needs at a lower price. I myself could well have bought a 980ti but decided against it because the Fury pro was cheaper and was more than capable at 1440p.

We will just have too see how well all these newbies are priced but 480x has the potential to be great little 1440p cards at an affordable price.
 
anyone else thinking that if those marks are in anyway an indicator of the performance, then the card needs to be at minimum half as expensive as a 1080, as CF still is below the perf of a single 1080.

So if we look at current pricing say the 1080 is £620, then the Polaris card needs to come in at maximum of say £300 retail, and in all honesty probably needs to be closer to £270 in my opinion to warrant the performance difference and the fact you have to rely on subpar Crossfire drivers etc.

Still, at say £270, one of those cards is a monster in the price per performance stakes. If they are anywhere sub £270, they will have a ton of sales.
 
Lokken86 said:
I don't doubt it but I'm looking at the specs and I can't see it being too close to a 1080.

I could be off by about 5%, considering the 970 has 80% of the floating point performance of the 980 but only 15% slower in games. Best scenario for the 1070 in this case could be 20% behind 1080 or slightly ahead if the 980ti.
Given this and if the 480x results are accurate the 1070 would be about 15-20% ahead of 480 at a higher price point.

At the end of the day you don't always need maximum performance if you can buy something to cover your needs at a lower price. I myself could well have bought a 980ti but decided against it because the Fury pro was cheaper and was more than capable at 1440p.

We will just have too see how well all these newbies are priced but 480x has the potential to be great little 1440p cards at an affordable price.
Click to expand...
Pretty sure most people will buy a Polaris if they are priced correctly and perform close to the 390x, to keep them satisfied until Vega/1080 Ti. Really no point in spending 619£ on a card that won't be worth more than 400£ in a few months.
 
SiDeards73 said:
anyone else thinking that if those marks are in anyway an indicator of the performance, then the card needs to be at minimum half as expensive as a 1080, as CF still is below the perf of a single 1080.

So if we look at current pricing say the 1080 is £620, then the Polaris card needs to come in at maximum of say £300 retail, and in all honesty probably needs to be closer to £270 in my opinion to warrant the performance difference and the fact you have to rely on subpar Crossfire drivers etc.

Still, at say £270, one of those cards is a monster in the price per performance stakes. If they are anywhere sub £270, they will have a ton of sales.
Click to expand...

Should be closer to 230£. ;)
 
Toss3 said:
Should be closer to 230£. ;)
Click to expand...

Should be... but once you add on the Rip off Britain tax and gouging they will be closer to £250-270 i reckon.

If they dropped in at £250 at that level of performance i would trade up my 290 for 2 of them without hesitation, as would many other people i imagine.
 
Well as I mentioned earlier I saw the first of the Nano clearance deals,and it was a massive reduction too,and we are starting to see many retailers running dry of R9 390 and R9 390X stocks too.

This might add some credence to the rumoured performance of Polaris 10.
 
Last edited:
Seanspeed said:
Cuz that's worked out for them the past 4-5 years?

Performance leading cards may not be the most bought cards, but they skew brand perception.
Click to expand...

Hence why I said vega will not be low cost as if it has the performance and the price tag to match it will give AMD the headline in the high end which would benefit the lower cards through perception.
 
CAT-THE-FIFTH said:
Well as I mentioned earlier I saw the first of the Nano clearance deals,and it was a massive reduction too,and we are starting to see many retailers running dry of R9 390 and R9 390X stocks too.

This might add some credence to the rumoured performance of Polaris 10.
Click to expand...

Noticed this also - though it could simply be that Gibbo decided not to order more stock of AMD cards, due to poor sales etc.
 
TheRealDeal said:
Yes but if you look down a little they show you the full 3dmark screen shot. That's where i got the graphics score from.
Click to expand...

They didn't post 3dmark11 screens. Maybe you are confusing FS with 3DM11?
Either way you can check Kitguru, in their review they use graphics scores and 1080 is doing even better there.
 
Toss3 said:
Pretty sure most people will buy a Polaris if they are priced correctly and perform close to the 390x, to keep them satisfied until Vega/1080 Ti. Really no point in spending 619£ on a card that won't be worth more than 400£ in a few months.
Click to expand...

The 980ti is now dead.
eol and whatever else.
a card hyped so much but now replaced by mid range.

Thats Polaris for you. :D
Vega will cut down 1080 also.
seems Polaris have a good ipc improvement for dx11.
 
Toss3 said:
Pretty sure most people will buy a Polaris if they are priced correctly and perform close to the 390x, to keep them satisfied until Vega/1080 Ti. Really no point in spending 619£ on a card that won't be worth more than 400£ in a few months.
Click to expand...
Well it depends on what GP106 is going to be like, too. People keep forgetting about that, thinking AMD will have that niche all to themselves.

And whatever you think about what 'most' people will buy, I doubt 'most' people buy £200+ graphics cards as 6 month stop gaps. Polaris 10 will appeal to people who dont already have 970/390 level cards and are still playing at 1080p. The sort that like to spend £150-200 and not upgrade for the next 3-5 years. Which is still a lot of people. But again, I think having performance leaders are still important for adoption even at these lower tiers. It affects how people view a brand. We've seen for years that AMD have had better value in the low/mid tier levels and still haven't been able to grab any meaningful share percentage back through it. People claim it's because of Nvidia marketing, but I think it's because Nvidia put out cards that perform very well Day 1 and also occupy the 'top dog' spot. It's all great that a 390 can outperform a 970 on 'average' nowadays, but it took a bit too long for that to happen. The damage had been done already. The 970 was received as a god send and the 390 a 'decent card', but merely a rebrand.

AMD need to come out swinging. They absolutely need to. If they want to play this lower tier game again, as they've been doing for a while, they need to do better than they have. Especially if they're going to give up the enthusiast sector completely to Nvidia for the next 6-12 months.

This is why I'm expecting really great pricing. £200 for 390X-beating performance will be great and very hard for Nvidia to match, much less beat. Anymore than that and I'll have to question AMD's grasp on public perception.
 
You must log in or register to reply here.
Back
Top Bottom