[AMD] Vega coming in 1H 2017

[Arachnotronic]

Looks like they aren't quite committing to 1Q 2017, instead just saying 1H 2017. My guess is we'll see it in Q2 2017.

 

[Phynaz]

I Agree.

 

[thilanliyan]

So no competition for 1070/1080 for 6+ months? That sucks if true.
I hope they release a 490 or something in the meantime.

 

[Face2Face]

No bueno if true... Volta shouldn't be too far behind.

 

[tential]

I'm surprised people won't admit this is a failure launch.
I had high hopes for amd, but leaving the high end open for that long is a joke.

AMD doesn't look like they'll be able to truly compete with Nvidia for a long time.

Especially considering the 480 isn't significantly faster or cheaper than the 1060 I just dont see what amd has done this generation that gets people excited.

These waits for amd gpus are just not worth it. At all.

 

[linkgoron]

You don't get it. That's just Vega 11! Vega 10 isn't for the enthusiast class, it's 1080 class, and will be released ~October!

Anyway, yeah the joke's on us. I can't believe that the AIB 980ti will go uncontested for almost two years, not to mention the 1070+ lines. Extremely miserable launch schedule from AMD.

EDIT:
Looks like AMD have started to release flagships only in odd years.

 

[Headfoot]

Q2 2017 seems likely to me too. They wouldn't revise Q1 2017 to 1H 2017 if they didnt think there was a distinct possibility it will slip later. That's a pretty long gap to leave the high end uncontested. Sucks for enthusiasts. They are likely capturing a good chunk of the marketshare they would have gotten anyways with the low to midrange released already, but I'm sure a solid chunk of the profit lays in the high end even with reduced unit volumes

 

[Arachnotronic]

Q2 2017 seems likely to me too. They wouldn't revise Q1 2017 to 1H 2017 if they didnt think there was a distinct possibility it will slip later. That's a pretty long gap to leave the high end uncontested. Sucks for enthusiasts. They are likely capturing a good chunk of the marketshare they would have gotten anyways with the low to midrange released already, but I'm sure a solid chunk of the profit lays in the high end even with reduced unit volumes

Yeah, and in that same deck, they are very clear about when other products come in Q1 vs 1H 2017, see:

 

[Yakk]

AMD should seriously staet looking at pre-orders for their launches. Miners will be able to order as many as they want, prepaid, and so everyone else will be able to buy cards also on launch day.

 

[Erithan13]

It's disappointing but I honestly don't know why anyone finds this surprising. Take a moment to review how outmatched AMD are against NV at the moment - NV own the majority of the market and are pulling in staggering amounts of money, meanwhile AMD have been in dire straights for what feels like forever and are seriously strained for resources (to say nothing of mismanagement/bad marketing etc). You can criticise AMD for a lot but when you ask why Vega didn't arrive with Pascal you're basically asking AMD to produce a product competitive with NV, released at the same time as NV, presumably available in quantities similar to that of NV, while having only the smallest fraction of resources relative to NV and also being tied down to that most awesome of places GlobalFoundries. Basically, given where AMD are at the moment, I'd be absolutely staggered if they could have produced a competent Vega when it was really needed to go against Pascal. Saying that AMD shouldn't have made the decisions that got them into this place to begin with is one thing, accepting that this is where they are and expecting them to compete on all levels with NV is quite another.

I suppose what makes this really interesting is that it's a good test for those of us who say a lack of competition is a bad thing for the consumer. The 480 vs 1060 is a real battle, one which I hope all of us can agree is a good thing regardless of your opinion on those two cards. Those looking at the high end will have no choice but Nvidia for quite some time, it will be most fascinating to see what sort of pricing that monopoly brings.

 

[sandorski]

It's disappointing but I honestly don't know why anyone finds this surprising. Take a moment to review how outmatched AMD are against NV at the moment - NV own the majority of the market and are pulling in staggering amounts of money, meanwhile AMD have been in dire straights for what feels like forever and are seriously strained for resources (to say nothing of mismanagement/bad marketing etc). You can criticise AMD for a lot but when you ask why Vega didn't arrive with Pascal you're basically asking AMD to produce a product competitive with NV, released at the same time as NV, presumably available in quantities similar to that of NV, while having only the smallest fraction of resources relative to NV and also being tied down to that most awesome of places GlobalFoundries. Basically, given where AMD are at the moment, I'd be absolutely staggered if they could have produced a competent Vega when it was really needed to go against Pascal. Saying that AMD shouldn't have made the decisions that got them into this place to begin with is one thing, accepting that this is where they are and expecting them to compete on all levels with NV is quite another.

I suppose what makes this really interesting is that it's a good test for those of us who say a lack of competition is a bad thing for the consumer. The 480 vs 1060 is a real battle, one which I hope all of us can agree is a good thing regardless of your opinion on those two cards. Those looking at the high end will have no choice but Nvidia for quite some time, it will be most fascinating to see what sort of pricing that monopoly brings.

I suspect the problem has to do with the availability of HBM2, not with AMD unable to produce a competitive GPU.

 

[linkgoron]

I suspect the problem has to do with the availability of HBM2, not with AMD unable to produce a competitive GPU.

If that's true, HBM2 was a horrible engineering/business decision.

Nvidia are managing very well with GDDR5x.

 

[sandorski]

If that's true, HBM2 was a horrible engineering/business decision.

Nvidia are managing very well with GDDR5x.

We'll see how bad a decision that will have been. So far they appear to be making Gains. If that continues the decision won't be that bad.

 

[richaron]

I'd be very surprised if the AMD doom and gloom crew are correct here. In fact the thread title screams how desperate they are to jump to conclusions and manipulate information.

At least use a proper quote and insert "for enthusiasts" into title

I can see a couple of options, obviously the chance of a pro vega product. And also I imagine there could be a segment above "performance mainstream" (or whatever Polaris 10 is) and below "enthusiast" wh baby vega could slip. Though OP's crystal ball has ruled them out...

All that being said, sad if true (which it might be). But since they've already got way too much demand maybe they decided to wait.

 

[Arachnotronic]

I'd be very surprised if the AMD doom and gloom crew are correct here. In fact the thread title screams how desperate they are to jump to conclusions and manipulate information.

Nope, just a correct interpretation of the slide + what management has said in public calls.

At least use a proper quote and insert "for enthusiasts" into title

Vega is the enthusiast family; Polaris is the mainstream family. Putting it in the title would have been redundant.

I can see a couple of options, obviously the chance of a pro vega product. And also I imagine there could be a segment above "performance mainstream" (or whatever Polaris 10 is) and below "enthusiast" wh baby vega could slip. Though OP's crystal ball has ruled them out...

They say Pro polaris is coming in Q4 2016 right in that same slide, FYI.

All that being said, sad if true (which it might be). But since they've already got way too much demand maybe they decided to wait.

That doesn't make sense.

 

[boozzer]

I'm surprised people won't admit this is a failure launch.
I had high hopes for amd, but leaving the high end open for that long is a joke.

AMD doesn't look like they'll be able to truly compete with Nvidia for a long time.

Especially considering the 480 isn't significantly faster or cheaper than the 1060 I just dont see what amd has done this generation that gets people excited.

These waits for amd gpus are just not worth it. At all.

haha, why does 480 have to be significantly faster and cheaper to compete?

hell, I am willing to bet that without the rx 480, the only 1060 we will get as gamers would be the 1060 3gb gamers should be thanking AMD, even if they bought a 1060.

competition is invaluable.

 

[Madpacket]

I think Raja was clear on what market he was going after so this shouldn't be a big surprise. The delay in Vega is disappointing but regaining market share matters most right now which is what they are doing.

 

[Stormflux]

If Vega launches in Q2 2017 and is only comparable to the 1070/80 we have now it will be mostly DOA, save for the die hards, unless they drag the entire current pricing scheme down with them. Being tied to HBM2 that seems unlikely.

With rumors of nVidia moving Volta up to 2017 as well... either nVidia knows something and/or AMD has conceded entirely and is just using Polaris for their WSA and Vega as 14nm/16nm pipe cleaners for their coffins... or Navi. 16nm if the Vega@TSMC rumours hold true.

For such a long wait getting the same but much later will be disappointing and a huge failure. Perf/Watt actually being similar this time due to HBM as GCN4 already loses to Pascal without it in this metric. The only way I can see this working out for AMD is if they are wanting to stagger step with nVidia on the high end performance crown. This would still require AMD to release it early in 2017 as the Volta shadow continues to grow the longer they delay.

Example: Q2 2016 for 1070/1080 => Q1 2017 for Vega10/11 for ~15%+ performance over the 1070/1080(Ti)/TitanX at similar or lesser price points in non-AMD favoured games. Then it will be in nVidia's capable hands to close the window from Vega to Volta. Which I no doubt will happen. Volta being +30% performance over Pascal. A typical like clockwork improvement for nVidia.

Will this stagger scenario happen? I doubt it but I don't see how AMD can make this work out otherwise. And as has been proven in the past, most consumers will just wait for nVidias new cards instead of going for AMD when they hold the crown.

Hopefully Zen works out...

 

[jpiniero]

I think Raja was clear on what market he was going after so this shouldn't be a big surprise. The delay in Vega is disappointing but regaining market share matters most right now which is what they are doing.

I think that's just spin though. If AMD had the resources I'm sure we would have seen Vega by now.

 

[alcoholbob]

Big Vega might end up competing with Nvidia's mid-tier Volta the same year it's launched. It might have to have a significant price drop within a quarter of launch.

 

[Azix]

People, is this really a delay in vega or are we just finally getting a launch window? Wasn't october a rumor?

It's disappointing but I honestly don't know why anyone finds this surprising. Take a moment to review how outmatched AMD are against NV at the moment - NV own the majority of the market and are pulling in staggering amounts of money, meanwhile AMD have been in dire straights for what feels like forever and are seriously strained for resources (to say nothing of mismanagement/bad marketing etc). You can criticise AMD for a lot but when you ask why Vega didn't arrive with Pascal you're basically asking AMD to produce a product competitive with NV, released at the same time as NV, presumably available in quantities similar to that of NV, while having only the smallest fraction of resources relative to NV and also being tied down to that most awesome of places GlobalFoundries. Basically, given where AMD are at the moment, I'd be absolutely staggered if they could have produced a competent Vega when it was really needed to go against Pascal. Saying that AMD shouldn't have made the decisions that got them into this place to begin with is one thing, accepting that this is where they are and expecting them to compete on all levels with NV is quite another.

I suppose what makes this really interesting is that it's a good test for those of us who say a lack of competition is a bad thing for the consumer. The 480 vs 1060 is a real battle, one which I hope all of us can agree is a good thing regardless of your opinion on those two cards. Those looking at the high end will have no choice but Nvidia for quite some time, it will be most fascinating to see what sort of pricing that monopoly brings.

its not like pascal was a monumental task. Making a new architecture, i.e. volta, clearly takes some time, even for nvidia. Shrinking maxwell and tuning it is not the same. Not sure what AMD is doing. They could shrink fiji, tune and make it clock higher and grab some cash that way. Only they know.

Anyway, zen is far more important right now than vega. Far more important. This is what will put AMD back in the big leagues and leave nvidia in varsity like before. It makes sense that they focus on getting that right. CPUs are in more PCs than high end GPUs by a million times.

 

[richaron]

All that being said, sad if true (which it might be). But since they've already got way too much demand maybe they decided to wait.

That doesn't make sense.

Lel, you can't even admit something as simple as AMD can't keep up with demand. Most people would have simply shifted goal posts to GF supply, but that even wouldn't change the truth of my statement.

Vega is the enthusiast family; Polaris is the mainstream family. Putting it in the title would have been redundant.

Nope. Vega will be in pro lines of course. Which is enough reason to call out the "D" in your fud title. You have no information saying these pro cards will also be 1st half 2017 (in fact recent media point much sooner). The possibility of a Vega card in between "performance mainstream" (Polaris 10) and "Enthusiast" segments is also something you're ignoring with your title.

There's a reason the source is needed in the title of a thread. And it should be the same reason that a title should not be a deliberate mis-quote. I'm actually glad you posted, I come here mostly for the news. But the manner you're spreading this information deserves reproach.

 

[jpiniero]

I dunno, looks like it's from AMD's investor slides. Seems pretty legit to me.

Lel, you can't even admit something as simple as AMD can't keep up with demand. Most people would have simply shifted goal posts to GF supply, but that even wouldn't change the truth of my statement.

GloFo is definitely the problem, without question. AMD should be able to get enough wafers to fulfill demand, there's not much going on at GloFo 16FF right now. The 460 and 470 are much more available too, although that could be easily explained by miners are being interested in the 480.

 

[mohit9206]

Man what a bummer.Not that i was interested in Vega to begin with,but still expected it to come out later this year so we could get the real msrps on 1070/1080 instead of the current inflated prices.But now congrats to Nvidia,keep enjoying the complete control on $300+ graphics card market for another year.Thing is when Vega does come out,it won't even be competitive with Pascal let alone Volta and efficiency not even competitive with Maxwell.Its going to be quite an embarassment really.

 

[4K_shmoorK]

Damn thats disappointing.

I'm not sure why some posters here are spinning the slides. Clearly states Vega will not be here, in any form, until 1H 2017. Are there any slides that claim otherwise? Thats an astronomical amount of time. The perfect window for NV to release a Ti right when a competitive Vega card releases. Basically a repeat of Fury X and 980 Ti. Then Volta right around the corner.

Damn I wish it were sooner. Just too long for most enthusiasts IMO when a new arch from NV would be another 6-10 months later. Especially if one already owns a Pascal card.