AMD Polaris Thread: Radeon RX 480, RX 470 & RX 460 launching June 29th

[3DVagabond]

Are they 'several months' ahead of cut-down GP104 / GP106 (whatever comes next)? Not so sure.

Also he says 'especially', not 'exclusively' notebooks and the mainstream market.

Suppose they have broad availability before nVidia? Does that put them ahead? It's easier to release something if you only have a 5 minute supply.

 

[ShintaiDK]

Suppose they have broad availability before nVidia? Does that put them ahead? It's easier to release something if you only have a 5 minute supply.

I thought people learned after the complete failure of the Polaris claims in terms of release vs Pascal.

GP106 could come before Polaris, just like GP104 and GP100.

 

[ShintaiDK]

It's pretty simple. AMD wants to regain marketshare.

You can't majorly swing marketshare with high-end priced GPUs. Unobtainium to the masses..

Nothing points to AMD will regain market share with their mainstream/low end lineup.

Also a card like the GTX 1070 will most likely be the single best selling card at all. Just like the GTX 970 was.

GP106 from the rumoured specs looks to be better than Polaris 10 in the metrics. so all in all, the dream of regaining a big part of the lost market share and revenue seems to be just that, a dream.

 

[IntelUser2000]

Shintai, do you believe that AMD won't gain marketshare based on the knowledge you have about their products or is it that you don't want them to?

Consumers will decide whether they'll gain it or not. Surely, if Polaris 10/11 is a compelling product(and that includes right pricing), they'll gain some marketshare at least. You are right it won't be at the high end. But at the price point definitely.

Remember that AMD had 50%+ retail marketshare with Athlon 64 chips. You can manipulate markets with bribes and marketing trickery to an extent, but with a good enough product(and that means again, right pricing) it'll speak for itself.

Intel got to where they are in the beginning because they had good products. Of course businesses fall from their "glory" days since people run it, and people mess up things.

 

[ShintaiDK]

I dont see their lineup changing any market share. If anything, they may lose some.

AMDs lineup for the rest of 2016 and part into 2017 essentially evolves around selling mainstream/low end parts that they have to compete with Nvidia about, and doing so at worse perf/watt it seems. While Nvidia gets the higher segments for themselves.

Something went wrong from AMD in terms of the lineup. Either they didn't expect Nvidia to do what they did (Unlikely). Or they simply didn't have the R&D to do better. It was clear that the "Chili event" was a desperate attempt to keep hype and focus, while they would lose the battle later. It made some people believe AMD was 6-9 months ahead.

Again, the products are priced because they perform as such in the metrics. Not due to any other factor.

 

[Bacon1]

I dont see their lineup changing any market share. If anything, they may lose some.

AMDs lineup for the rest of 2016 and part into 2017 essentially evolves around selling mainstream/low end parts that they have to compete with Nvidia about, and doing so at worse perf/watt it seems. While Nvidia gets the higher segments for themselves.

Something went wrong from AMD in terms of the lineup. Either they didn't expect Nvidia to do what they did (Unlikely). Or they simply didn't have the R&D to do better. It was clear that the "Chili event" was a desperate attempt to keep hype and focus, while they would lose the battle later. It made some people believe AMD was 6-9 months ahead.

Again, the products are priced because they perform as such in the metrics. Not due to any other factor.

It is amazing that you can state all of that garbage as facts when you are completely wrong.

AMD has planned to release 390/970 level of performance for cheap since they announced Polaris back in January.

https://www.youtube.com/watch?v=p010lp5uLQA?t=955

We have no idea how much power it takes, only that it has a MAX TDP of 150w (75 board, 75 6 pin), nor do we know what the performance is exactly, only that it is above 970/390. So how can you claim to have any perf/watt #s?

The products are priced that way because, they said that's what the pricing would be 6 months ago.

If Nvidia is ahead so far, why can't they keep tiny amounts of under performing overheating "premium" 1080's in stock anywhere? Why didn't they have more than a few hundred per store? You claim they've sold thousands everywhere but when asked for sources you have never provided any.

Start putting up sources, stop spouting FUD.

 

[kraatus77]

lol dream on, they gained 6% market from nvidia in last 2 quarters and that's with mostly rebrands. just wait and watch how rx480 gains another 10% in next 2 quarters.


meanwhile gp106 is MIA not even a single bench out there. but we already know 29th is the day for rx480.

it's priced low not because they failed but because that's what they promised.
6 months ago they said it will be a mainstream product aim for vr spec in lower price and will launch in mid 2016, 29th june is exactly middle of 2016. what more accuracy you want lol. and its die size + single 6pin suggest it's a low tdp product. just like 115w 7850, which also had 1x6 pin.

 

[flopper]

Start putting up sources, stop spouting FUD.

Paid Nvidia focus group.
you find them in many forums.




Your post is inflammatory to the topic and is not allowed here.


esquared
Anandtech Forum Director

 

[IntelUser2000]

Shintai, there's too much definitive conclusions based on what's unknown. The probability of AMD being successful is lower than Nvidia, but that doesn't mean they are at zero.

Granted, there's a bit of bias on both sides. The market and product will speak. No matter who says, believes, or wishes.

If Nvidia is ahead so far, why can't they keep tiny amounts of under performing overheating "premium" 1080's in stock anywhere? Why didn't they have more than a few hundred per store? You claim they've sold thousands everywhere but when asked for sources you have never provided any.

Who knows? 14/16nm problems, for artificial price inflation(to get 1% more earnings for shareholders), the "I got here first" claim.

On a side note, I believe 14/16nm problems will be just a shadow of what's to come. All these fancy SiGe or CnT technologies will bring ever smaller and harder gains. Moore's Law ends not with a bang(that is, instant end) but a whimper.

 

[Dribble]

Outside of fantasy land it doesn't take a much perception to see AMD are in a bad place right now. They have no Zen or Vega in 2016 (that Vega 2016 was one of the shakiest rumours that was latched onto desperately like it was truth). Not that Zen will help, when it comes out we all know deep down it's got no chance of competing with Intel. It's almost better not releasing it so that hope remains.

Anyway so all they really have to make money is Polaris 10 and 11 this year.

Those gpu's should be more competitive with Nvidia then Zen ever has a chance of being with Intel, but they have been hamstrung by being forced to use GloFo which looks like it's running at a 20%+ clock speed deficit over TSMC. That effectively means they have to sell it a price bracket lower which means lower margins and profits. Add to this the fact they don't really have access to any of the higher margin markets (Pro and HPC) so even if they do have a great product they still don't get make the sort of money Nvidia does.

That to me seems to be the reality of the situation right now.

 

[ShintaiDK]

Agree, they would be much better of using TSMC 16FF+. But the WSA strikes again unfortunately. And its hurting them immensely.

 

[beginner99]

I'll probably get a 480, especially if there's a $229 8 GB version, just because the performance/$ is too good and I'm gaming at 1080 right now, but I hope that AMD uses Polaris to fulfill their wafer obligations from Global Foundries and uses Samsung for Vega because the evidence right now points to GF making a pig's breakfast of another node.

That's exactly the reason AMD goes mainstream first. To gain marketshare and to fulfill WSA. They get wins for consoles and laptops and mainstream desktop. Exactly what they need. High volume products.

What they need less is the performance crown in GPU segment. It would for sure help a lot in terms of brand value but the investment in research and expected return in a low-volume market is regardless of high-margins isn't AMDs first priority. The WSA thing is also why then can live with low margins. Every sold chip means they have to pay less to GF. So there is some kind of hidden margin in every AMD chip.

 

[Adul]

Outside of fantasy land it doesn't take a much perception to see AMD are in a bad place right now. They have no Zen or Vega in 2016 (that Vega 2016 was one of the shakiest rumours that was latched onto desperately like it was truth). Not that Zen will help, when it comes out we all know deep down it's got no chance of competing with Intel. It's almost better not releasing it so that hope remains.

Anyway so all they really have to make money is Polaris 10 and 11 this year.

Those gpu's should be more competitive with Nvidia then Zen ever has a chance of being with Intel, but they have been hamstrung by being forced to use GloFo which looks like it's running at a 20%+ clock speed deficit over TSMC. That effectively means they have to sell it a price bracket lower which means lower margins and profits. Add to this the fact they don't really have access to any of the higher margin markets (Pro and HPC) so even if they do have a great product they still don't get make the sort of money Nvidia does.

That to me seems to be the reality of the situation right now.

This clock speed deficit crap had got to stop. Nvidia purposely went for higher clocks with Pascal, it's designed to hit this higher frequency at the cost of some ipc. I have seen nothing to indicate that amd was going for clock speed with Polaris so far. If it is running at 1266 mhz for a gcn core, that to me seem to be reasonable increase over previous gcn cures. It takes more than a node shrink to hit higher frequencies. You need to design for it a well. I don't think AMD went that route.

Sent from my Nexus 6P using Tapatalk

 

[Qwertilot]

lol dream on, they gained 6% market from nvidia in last 2 quarters and that's with mostly rebrands. just wait and watch how rx480 gains another 10% in next 2 quarters.

Some realism here? The gains were from a historic low, with NV right at the end of their product cycle.

The next quarter AMD are very likely to lose a bit of overall market share, almost whatever the 480 does. A big chunk of AMD's product stack is effectively obsolete. It didn't sell that much compared to NV but it did sell!

Meanwhile NV are also going to have a big spike in sales for the new 1070/80 vs 2 year old 970/80's and one year old 980ti's.

That matters vastly less to AMD than some people here though They're probably at least plausibly happy with Polaris having got those console refresh deals - everything else is sort of free money. If they - unlikely given history - manage to outsell the 1050/60 they'll be very happy.

 

[kraatus77]

Some realism here? The gains were from a historic low, with NV right at the end of their product cycle.

The next quarter AMD are very likely to lose a bit of overall market share, almost whatever the 480 does. A big chunk of AMD's product stack is effectively obsolete. It didn't sell that much compared to NV but it did sell!

Meanwhile NV are also going to have a big spike in sales for the new 1070/80 vs 2 year old 970/80's and one year old 980ti's.

That matters vastly less to AMD than some people here though They're probably at least plausibly happy with Polaris having got those console refresh deals - everything else is sort of free money. If they - unlikely given history - manage to outsell the 1050/60 they'll be very happy.

They gained 6% that's a real fact. period.

and new products are for mass market. 1080/70 aren't. 1080 is already having supply issues (maybe cuz g5x ?).

they hit their low because crap p/w vs nvidia, polaris will have same if not better p/w vs pascal and it is hitting first on mainstream. just wait and watch lol.

can't believe people spin everything against amd even though this is the first time in many years amd is right on track delivering what they promised. but some people can never be impressed by amd so i understand that. :\

 

[ShintaiDK]

they hit their low because crap p/w vs nvidia, polaris will have same if not better p/w vs pascal and it is hitting first on mainstream. just wait and watch lol.

can't believe people spin everything against amd even though this is the first time in many years amd is right on track delivering what they promised. but some people can never be impressed by amd so i understand that. :\

Interest, whats your source on this?

Both RX480 and GTX 1070 is specced for 150W. Yet they are by no means in the same performance class. In other words, Pascal looks to have substantial perf/watt advantage. And GP106 that competes more directly is 90-100W or so according to leaked specs.

You just delivered a spin, didn't you?

 

[Leadbox]

Interest, whats your source on this?

Both RX480 and GTX 1070 is specced for 150W. Yet they are by no means in the same performance class. In other words, Pascal looks to have substantial perf/watt advantage. And GP106 that competes more directly is 90-100W or so according to leaked specs.

You just delivered a spin, didn't you?

They do not define TDP the same, so best to wait on actual numbers no?

 

[Qwertilot]

I'm not spinning anything

They are indeed releasing what they planned, and if they've done it well - we've got no real idea as yet - they might do all right with it. Arguably they've already done quite well out of it before they even sell a single card to consumers.

Just have to be realistic about what what they've done might do for them in terms of market share. There will be absolutely direct NV competition against the 480.

It would be a huge (probably unrealistically optimistic given the sales in the previous generation) success for AMD to get half of this part of the market vs NV.

Even if they did manage that though, I think they'd still drop overall market share.

 

[kraatus77]

Interest, whats your source on this?

Both RX480 and GTX 1070 is specced for 150W. Yet they are by no means in the same performance class. In other words, Pascal looks to have substantial perf/watt advantage. And GP106 that competes more directly is 90-100W or so according to leaked specs.

You just delivered a spin, didn't you?

7850 was also rated for 150 w with 1x6pin, just like 480. and it used around 110w.

its rated for 150w doesn't mean it will consume that much power. why do you think 1070 has 1x8 pin even though 1x6 pin would suffice ?

yea because it goes above 150w , rx480 won't. and besides i still haven't said it will be better than pascal in p/w i said it will match it. just wait 20 more days. it's not a spin but educated guess based on board tdp and how 7850 worked. an educated guess something you need to learn

 

[pepone1234]

Interest, whats your source on this?

Both RX480 and GTX 1070 is specced for 150W. Yet they are by no means in the same performance class. In other words, Pascal looks to have substantial perf/watt advantage. And GP106 that competes more directly is 90-100W or so according to leaked specs.

You just delivered a spin, didn't you?

Both ar rated 150W TDP. One uses an 8 pin connector and the other uses a 6 pin connector.

Those two cards obviously are not going to have the same power consumption.

 

[selni]

This clock speed deficit crap had got to stop. Nvidia purposely went for higher clocks with Pascal, it's designed to hit this higher frequency at the cost of some ipc. I have seen nothing to indicate that amd was going for clock speed with Polaris so far. If it is running at 1266 mhz for a gcn core, that to me seem to be reasonable increase over previous gcn cures. It takes more than a node shrink to hit higher frequencies. You need to design for it a well. I don't think AMD went that route.

Sent from my Nexus 6P using Tapatalk

There's been near zero sensible analysis on IPC, just a bunch of people using a term they don't understand and dividing random things that have nothing to do with instructions per clock. No, dividing FPS by (approximate) clockspeed * cores is not IPC with very different GPU layouts and memory subsystems.

However the talk about clockspeeds *is* pretty dumb. Maxwell was already hitting much higher clocks than GCN when they were using the same process, Pascal hitting higher clocks than Polaris isn't a surprise and likely has nothing to do with GF vs TSMC.

 

[dacostafilipe]

... have to compete with Nvidia about, and doing so at worse perf/watt it seems ...

Interest, whats your source on this?

:thumbsdown:

You don't need a source but others do ?

Please stop with this nonsense.

 

[kraatus77]

:thumbsdown:

You don't need a source but others do ?

Please stop with this nonsense.

Never seem him backing up his jibbrish. it's like that bird who doesn't know anything about chess game, but walks over it crapping thinking he won or something.

 

[Silverforce11]

Nothing points to AMD will regain market share with their mainstream/low end lineup.

Also a card like the GTX 1070 will most likely be the single best selling card at all. Just like the GTX 970 was.

GP106 from the rumoured specs looks to be better than Polaris 10 in the metrics. so all in all, the dream of regaining a big part of the lost market share and revenue seems to be just that, a dream.

You have a good history of tech predictions...

Since you predict AMD will lose marketshare, it's almost assured they will have a brilliant Q3 and Q4 this year.

 

[parkerface]

Even when I log in just to activate my ignore list, the garbage still gets quoted for pages at a time.

I can't win. :thumbsdown: