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

[kondziowy]

AMD also have multiple milions PS4 Neo gpus to make for this year so yeah, there is that.

 

[AtenRa]

Now that NV has released two cards which dont even address the same market ,the fact that they wont rush out their release is an obvious sign of failure.

Ehmm, you understand that Polaris is not competing against GP104 and yet your evaluation is that because they dont rush the Polaris release its a sign of failure ??

Contradicting your self in the same sentence ??

 

[MrTeal]

Ehmm, you understand that Polaris is not competing against GP104 and yet your evaluation is that because they dont rush the Polaris release its a sign of failure ??

Contradicting your self in the same sentence ??

I think you should reread his post again for context and tone.

 

[rainy]

Ehmm, you understand that Polaris is not competing against GP104 and yet your evaluation is that because they dont rush the Polaris release its a sign of failure ??

Contradicting your self in the same sentence ??

He was clearly sarcastic, however he should use /s.

 

[AtenRa]

ehm, ok got it

 

[poohbear]

You guys should realize by now AMD doesn't care who is getting anxious. They said they weren't releasing till ready, they said they were aiming for mid year. Doesn't matter if nvidia puts Chuck Norris in their GPUs, AMD is doing what AMD wants.

Yes but don't u see the derth in marketing on AMD'S part and how it's affecting their sales? As an investor in AMD it's really off putting. They have no sense of marketing which is so important for successful product launches.

 

[BlitzWulf]

Yes but don't u see the derth in marketing on AMD'S part and how it's affecting their sales? As an investor in AMD it's really off putting. They have no sense of marketing which is so important for successful product launches.

That's part of the NV launch reality distortion field, after June 10th 85% of GPU purchases till the end of time will be 400-450$ 1070's , Market analysis and common sense be damned . /s

AMD's strategy has not changed and it is still the right one.
They desperately need the influx of cash more than they need the performance crown or to win the the hearts of tech journo's and forum warriors.

If they can bring a compelling product at the right price they might be able to win against their true greatest foe....debt.

 

[ZZZAAA]

AMD also have multiple milions PS4 Neo gpus to make for this year so yeah, there is that.

I don't think PS4K has been even confirmed yet so I doubt they'll release it this year. Read yesterday that Sony is still denying PS4 Neo existence with them expecting PS4 hardware sales to top 60m this year. It's only been out a few years.

Anyway, I just hope I didn't bite early with my Skylake build and AMD releases Vega way before Kaby Lake. That NDA news really let me down...

 

[trane]

I don't think PS4K has been even confirmed yet so I doubt they'll release it this year. Read yesterday that Sony is still denying PS4 Neo existence with them expecting PS4 hardware sales to top 60m this year. It's only been out a few years.

Anyway, I just hope I didn't bite early with my Skylake build and AMD releases Vega way before Kaby Lake. That NDA news really let me down...

Actually, if you read AMD's earnings press release, there's a major semi-custom APU ramping up second half this year. It may or may not be PS4 Neo but kondziowy's point stands.

 

[Sweepr]

Inno3D already talking about GP106:

Polaris will have competition very soon.

 

[trane]

Actually, I think that's not GP106. That's probably a heavily cut down GP104. If it were a GP106, it would a) occupy more than one SKU. b) follow on 3 months from GP104 like with every one xx6 GPU.

What this tells me is that a competitor to Polaris 10 XT is coming. GTX 1060 or 1060 Ti is coming. But under that, AMD is going to go unchallenged for quite some time. Polaris 11 almost certainly has no answer, and Polaris 10 Pro might go without competition for a while too. AMD has the sub-$300 market all to themselves, while Nvidia has the >$300 market all to themselves. And they'll both fight it out at $300.

 

[linkgoron]

I agree with Trane - it looks (to me) like the rumored gp104 660ti.

If true, it looks like the 480 non-x (assuming there is an x) and lower will go unchallenged for a few months as the 950/960 are staying in the market.

 

[Sweepr]

Guess we will see. But rumours say GP104-150 (Geforce GTX 1060 Ti?) is AIB-only.

 

[AtenRa]

Guess we will see. But rumours say GP104-150 (Geforce GTX 1060 Ti?) is AIB-only.

Inno3D is an AIB.

 

[AtenRa]

Here is where AMD targets Polaris 10-11 at, Mainstream and Performance. That is 80-85% of the market.

https://www.jonpeddie.com/publications/add-in-board-report

And in case you missed it, AMD Q1 2016 GPU marketshare is higher than last year Q1.

 

[ShintaiDK]

Its not 80-85% of the revenue. And Polaris seems to only enter the low/middle part of the performance segment.

And now it seems GP106 will launch before Polaris as well. Polaris is a chip they needed months ago, not months from now or whenever they launch it. Every single day is lost revenue.

 

[RussianSensation]

Its not 80-85% of the revenue. And Polaris seems to only enter the low/middle part of the performance segment.

And now it seems GP106 will launch before Polaris as well. Polaris is a chip they needed months ago, not months from now or whenever they launch it. Every single day is lost revenue.

They don't care about getting 85% of revenue only from P10/11 parts. What they need is profits and market share gains. That 1060 coming in between 970/980 is nothing impressive because that's R9 390 level of performance. P10 is rumoured to beat 390X. Besides, if both AMD and NV cards have similar average performance, the AMD card is an automatic winner since Pascal has not fixed DX12 performance. If we look at only DX12 games, the lead of 1070/1080 over Fury X/390X shrinks considerably. That means 1060's trump card will have to be power usage, not DX12 performance. Besides, with how poorly Kepler and Maxwell aged and NV putting GTX700 series on legacy status already, if both AMD and NV's cards perform similarly and cost similarly, P10 will be an automatic winner for driver support and DX12. You are also assuming 1060 will have 8GB of VRAM. I bet it only has 6, again a deficit to Polaris 10. Pascal is just Maxwell+ stop-gap generation before the real new architecture with Volta. Polaris 10 is rumoured to go into Neo and XB Scorpio, virtually guaranteeing superior useful life and driver support over 3-5 years that many mainstream gamers keep their cards for.

The irony here is many of these mainstream/performance users could have purchased an after-market 290/290X for $200-300 as early as November 2014. Considering 970 was only $330 September 2014, 1060 looks like a snooze fest since it's going to launch almost 2 years after similarly priced 290/290X/970. This entire 2016 is looking pretty boring with no real breakthrough other than the price/performance is the 1070.

I find it interesting that you also keep ignoring the possibly of 3 distinct Polaris 10 chips: 2048/2304/2560 shaders as well as AMD playing the price/performance + DX12 long game. It's going to be funny if Battlefield 1 is a DX12 AMD GE title where P10 is almost as fast as a 1070 for $80-100 less.

 

[dazelord]

They don't care about getting 85% of revenue only from P10/11 parts. What they need is profits and market share gains. That 1060 coming in between 970/980 is nothing impressive because that's R9 390 level of performance. P10 is rumoured to beat 390X. Besides, if both AMD and NV cards have similar average performance, the AMD card is an automatic winner since Pascal has not fixed DX12 performance. If we look at only DX12 games, the lead of 1070/1080 over Fury X/390X shrinks considerably. That means 1060's trump card will have to be power usage, not DX12 performance. Besides, with how poorly Kepler and Maxwell aged and NV putting GTX700 series on legacy status already, if both AMD and NV's cards perform similarly and cost similarly, P10 will be an automatic winner for driver support and DX12. You are also assuming 1060 will have 8GB of VRAM. I bet it only has 6, again a deficit to Polaris 10. Pascal is just Maxwell+ stop-gap generation before the real new architecture with Volta. Polaris 10 is rumoured to go into Neo and XB Scorpio, virtually guaranteeing superior useful life and driver support over 3-5 years that many mainstream gamers keep their cards for.

The irony here is many of these mainstream/performance users could have purchased an after-market 290/290X for $200-300 as early as November 2014. Considering 970 was only $330 September 2014, 1060 looks like a snooze fest since it's going to launch almost 2 years after similarly priced 290/290X/970. This entire 2016 is looking pretty boring with no real breakthrough other than the price/performance is the 1070.

I find it interesting that you also keep ignoring the possibly of 3 distinct Polaris 10 chips: 2048/2304/2560 shaders as well as AMD playing the price/performance + DX12 long game. It's going to be funny if Battlefield 1 is a DX12 AMD GE title where P10 is almost as fast as a 1070 for $80-100 less.

Feature-wise the 960 won over 380/380X (HDMI 2.0, HEVC decoding), but this time the 1060 must battle fully equipped Polaris cards. Power consumption for all these new cards must be so low it's irrelevant for most perople.

 

[jpiniero]

Guess we will see. But rumours say GP104-150 (Geforce GTX 1060 Ti?) is AIB-only.

You mean OEM only (and specifically for OEMs pushing VR). nVidia is of course free to change their mind.

 

[Flapdrol1337]

They don't care about getting 85% of revenue only from P10/11 parts. What they need is profits and market share gains. That 1060 coming in between 970/980 is nothing impressive because that's R9 390 level of performance. P10 is rumoured to beat 390X. Besides, if both AMD and NV cards have similar average performance, the AMD card is an automatic winner since Pascal has not fixed DX12 performance. If we look at only DX12 games, the lead of 1070/1080 over Fury X/390X shrinks considerably. That means 1060's trump card will have to be power usage, not DX12 performance. Besides, with how poorly Kepler and Maxwell aged and NV putting GTX700 series on legacy status already, if both AMD and NV's cards perform similarly and cost similarly, P10 will be an automatic winner for driver support and DX12. You are also assuming 1060 will have 8GB of VRAM. I bet it only has 6, again a deficit to Polaris 10. Pascal is just Maxwell+ stop-gap generation before the real new architecture with Volta. Polaris 10 is rumoured to go into Neo and XB Scorpio, virtually guaranteeing superior useful life and driver support over 3-5 years that many mainstream gamers keep their cards for.

The irony here is many of these mainstream/performance users could have purchased an after-market 290/290X for $200-300 as early as November 2014. Considering 970 was only $330 September 2014, 1060 looks like a snooze fest since it's going to launch almost 2 years after similarly priced 290/290X/970. This entire 2016 is looking pretty boring with no real breakthrough other than the price/performance is the 1070.

I find it interesting that you also keep ignoring the possibly of 3 distinct Polaris 10 chips: 2048/2304/2560 shaders as well as AMD playing the price/performance + DX12 long game. It's going to be funny if Battlefield 1 is a DX12 AMD GE title where P10 is almost as fast as a 1070 for $80-100 less.

DX12 performance isn't "fixed"? If I look at computerbase running ashes of the singularity the 1080 is 44% faster than a reference 980Ti.

The fury X beats the 980Ti by 33% in the same test, but it only beats the gigabyte 980Ti gaming by 6%, not that impressive considering the 980Ti is a cut chip, fury x still uses more power and has hbm. So was there really that much to fix in the first place?

For battlefield 1 the P10 should be the best deal, I'm assuming the game is bundled with it so that's already a $60 advantage.

 

[nenforcer]

Apparently even though AMD has always said Mid year,before back to school for Polaris.
Now that NV has released two cards which dont even address the same market ,the fact that they wont rush out their release is an obvious sign of failure.

Never mind that they may have several corporate partners that have made plans expecting launch at a certain date,let's also ignore how terrible of an idea it is to launch a product in to a high volume market without having the production to back it.

I have to agree this is a blunder of epic proportions on AMD's behalf almost as bad as not pushing the 380X forward to launch at the same time as the GTX 970 launched. I mean what were/are they thinking ?

I think you've got your timelines skewed. The GTX 970 4GB launched in September 2014 with a retail price of $329 and a market leading price / power / performance at the time. AMD also released the Radeon R9 285 2GB in September 2014 at $250 which quickly lowered to $200-$225 once the performance wasn't anywhere near the GTX 970. The Radeon 380X launch didn't come until after a year later in November 2015 after the rest of the RX 300 series had already launched in June 2015. Those 2 cards were always seperated by about $100 so not really direct competitors. The GTX 960 2GB released in January 2015 was the answer to the R9 285 2GB and then obviously the 960 4GB competed against the 380/X 4GB. The R9 390 series is what competed with the GTX 970 / 980.

 

[BlitzWulf]

I think you've got your timelines skewed. The GTX 970 4GB launched in September 2014 with a retail price of $329 and a market leading price / power / performance at the time. AMD also released the Radeon R9 285 2GB in September 2014 at $250 which quickly lowered to $200-$225 once the performance wasn't anywhere near the GTX 970. The Radeon 380X launch didn't come until after a year later in November 2015 after the rest of the RX 300 series had already launched in June 2015. Those 2 cards were always seperated by about $100 so not really direct competitors. The GTX 960 2GB released in January 2015 was the answer to the R9 285 2GB and then obviously the 960 4GB competed against the 380/X 4GB. The R9 390 series is what competed with the GTX 970 / 980.

I was being sarcastic, but that's what people seem to be implying.
AMD is supposedly faltering because they are not rushing out a product as an answer to it's competitor releasing two other product's,
neither of which it has never claimed to be trying to directly compete with in the first place.

 

[VulgarDisplay]

AMD needs to forget their timeline and release information soon.

They are leaving the brainless masses in the dark and then trusting them to understand that Polaris is not meant to compete with 1080 and 1070. It doesn't matter how compelling your card is for the price and they simply have not learned this lesson during their market share hemorrhaging. Having the fastest card on the market speaks volumes for the uninformed that typically shop in the price bracket they are targeting. Nvidia's market domination has been fueled by spontaneous trait transference for the last 4 gpu generations. People see the fastest flagship and automatically think that everything nvidia has the fastest card no matter than performance level.

They either need to outright take the performance crown or start getting information out everywhere they possibly can or no one will buy Polaris even if it's awesome. Their current strategy only works on c-type (cognitive) personalities, and only on some of them because they still want the best even if it costs them.

Sent from my XT1575 using Tapatalk

 

[tential]

AMD needs to forget their timeline and release information soon.

They are leaving the brainless masses in the dark and then trusting them to understand that Polaris is not meant to compete with 1080 and 1070. It doesn't matter how compelling your card is for the price and they simply have not learned this lesson during their market share hemorrhaging. Having the fastest card on the market speaks volumes for the uninformed that typically shop in the price bracket they are targeting. Nvidia's market domination has been fueled by spontaneous trait transference for the last 4 gpu generations. People see the fastest flagship and automatically think that everything nvidia has the fastest card no matter than performance level.

They either need to outright take the performance crown or start getting information out everywhere they possibly can or no one will buy Polaris even if it's awesome. Their current strategy only works on c-type (cognitive) personalities, and only on some of them because they still want the best even if it costs them.

Sent from my XT1575 using Tapatalk

Yup vulgar display this is pretty much wrapped up to me. Polaris is already going to have a massive uphill launch thanks to amd.

 

[crisium]

AMD needs to forget their timeline and release information soon.

They are leaving the brainless masses in the dark and then trusting them to understand that Polaris is not meant to compete with 1080 and 1070.

We're even getting "articles" from one particular brainless mass that show this confusion. Some people have a [H]ard time understanding, to be fair, so AMD would only benefit from making things clearer, I agree.