AMD VEGA confirmed for 2017 H1

[old gamer]

Anyone with sense wants to see AMD release a stonking card and get back to parity with NVIDIA but surely anyone with sense can see a 1080/70 card from AMD will struggle to sell well? As has been said, all NVIDIA have to do is drop the price of those cards and job done.

thing is, maybe 1070/80 performance is the best amd can do which I think is not all doom and gloom, it will be fast enough for my needs and I'm sure it will sell well if priced right.
I don't think the 1070/80 has sold as well as the 970/80 and there must be loads of people still sitting on 290's and 970's waiting on a good bang for buck card and who knows vega just might be that card

 

[five8five]

Vega releasing in May.

http://wccftech.com/amd-vega-gpu-radeon-graphics-may-launch/

Wow that's so far away.

wccftech

 

[KentMan]

The problems AMD face right now in the GFX market are many fold, and they never make it easy for themselves.

1) If they release at say 10% performance over 1080 in DX11 titles, they still have to be cheaper than the 1080 to get people to switch. Also that 10% can be easily matched or beaten by Nvidia if they choose to release a 1080ti, AMD cannot price higher than the 1080, as it will lose against a 1080ti at a similar price point. So AMD need to be cheaper but offer more performance, the ace they have up their sleeves is how their cards generally mature better, while this is an ace for some, others see it as lazy and that the performance should have been there from the start. Brand perception will potentially get a hammering yet again if they fail with this.

2) If they are slower than the 1080, but beat the 1070 they probably need to match the 1070 price, again as above, to get people to move across.

3) Future Tech, AMD love to build this into their products, while this gives them longer life spans, its rubbish for the early adopters, who watch a ton of hardware sit under utilised for aeons until the software catches up, if it ever does.

4) Advertising, while this has gotten better, they need to stop the petty pot shots like the "Poor Volta" thing, its clearly aimed at Nvidia, with the latin dig at them sleeping, even coloured green so as to not be any mistake about what they are poking fun at. If your going to be that blatant, you need to product to back it up, or you are going to lose even more of your ever dwindling user base.

5) Accept your place in the market and actually warm to it, rather than trying to force everyone to accept you as the "Premium Brand" instead of telling everyone thats who you are, and then pricing your underperforming products as such, embrace your seen as the performance to pound spent king and use it, once your brand strength and confidence expands, you can actually start charging premiums people will wear, if the product warrants it. Simply saying your premium but not actually being premium, will lose you even more market share.

I think AMD are at a crossroads right now, they can potentially be the biggest news of 2017 if they get their CPU and GPU Strategies right, i think Lisa Su has done a great job so far and that gives confidence that going forward with Zen and Vega we should hopefully not see some stupid decisions made, but i have to err on caution as this is is AMD and they always drop the ball one way or another.

 

[D.P.]

It probably will be, According to those benchmarks Vega is 10% faster than the 1080, take out AMD 'best case' and add in Vega optimised drivers they will probably be 10% faster than the GTX 1080 in reality and a bit cheaper.

They ain't going to be anything more than that, and thats fine, it will have-to be because much more than that is fantasy.

If that is all that vega is then their market share will liekly slip. No one with a 1070 or 1080 will bother switching and Nvidia can happily lower their prices or ride on their superior marketing and brand recognition.


AMD almost certainly can't make something significantly faster than Pascal for less money but they should have released Vega last year to sit along side the 1080. Evidently they couldn't almost certainly due to diminished R&d and the priority of Zen. But this wont help AMD's GPu sales in the future, it just further cements the technological superiority of Nvidia in many of the public's minds.


Nvidia also may well have some actual refreshes of Pascal, new stepping and tweaks with faster clocks since they have had a year fiddling with Pascal and the 16nm process.

 

[maonayze]

Yeah, if it turns out that this is indeed what we will get from Vega 10 (Big Vega) then it will be a tad underwhelming tbh.

I now think what we will get is a card that competes well with the 1080 and may just overtake it once the drivers are optimised. I'm guessing it will beat it by something between 5%-15% with matured drivers. Which should put it near the 1080Ti unless Nvidia decide to only slightly undercut the TXP performance like they did with Titan X and the 980Ti (I was under the impression that Nvidia were not going to put the 1080Ti so close in performance to the Titan XP this time...but I suppose that could be tweaked between now and then depending on the Final Vega reveal)

I think this is what has been meant by Raja and Lisa Su when they said that Vega would be very competitive. It will indeed be very competitive performance wise to the 1080/Ti but I don't think that will produce the massive shift in mindshare that others seem to want.

Personally I think that AMD could never really just turn everything on it's head at such a speed without a big gamble and putting AMD in jeapody of loosing everything on it. This is Raja's long game, and this needs to be done carefully and steadily. If Ryzen is a success then it will be a surer platform to build on for the future of the company. Only they know the Financial boundaries on which they can compete and they have to stay within them to survive and succeed.

And before Flopper says it......Vega 20 Titan Killer coming in H2 2017

 

[Doom112]

This is exactly why i wish AMD just do not ever market anything. They always make expectation so high that the only chance is that you will take a noise dive.

They always tend to compare their product and speak about their competitor more like they did in Polaris launch and compare RX 480 CF with GTX 1080, which became disaster for them. Since 3 or 4 year Nvidia has never spoken about AMD or never mentioned about AMD in their marketing or presentation and for them it is like AMD do not exist now a days. It is like AMD is always trying to seek attention at any cost.

 

[humbug]

The Mindshare is such that AMD need to be better and cheaper than nVdia before just some of the people who have nVidia on their mind to even consider buying AMD instead.

The reality is it doesn't matter how good and or value for money AMD actually are a small but very loud proportion of Nvidia's mindshare are militant, with that no matter how untrue the rhetoric it stops AMD mindshare growth, which is the whole point.

The situation these days is such that AMD can't win, if even survive long term because these days marketing tactics have gone way beyond just honestly promoting your products, and if you are late to that game you are too late.

@ D.P. i think AMD's days in the consumer space are numbered... with that i think we as PC nerds are a dying breed.

 

[five8five]

Maybe the bigger performance card will come with what they call the Vega 20 later in the year, but it will be very late to the game.

Seems to be a recurring theme -

Fury..........wait until Polaris......wait until Big Vega.......wait until Vega 20.

"This time next year Rodney............"

 

[D.P.]

The problems AMD face right now in the GFX market are many fold, and they never make it easy for themselves.

1) If they release at say 10% performance over 1080 in DX11 titles, they still have to be cheaper than the 1080 to get people to switch. Also that 10% can be easily matched or beaten by Nvidia if they choose to release a 1080ti, AMD cannot price higher than the 1080, as it will lose against a 1080ti at a similar price point. So AMD need to be cheaper but offer more performance, the ace they have up their sleeves is how their cards generally mature better, while this is an ace for some, others see it as lazy and that the performance should have been there from the start. Brand perception will potentially get a hammering yet again if they fail with this.

2) If they are slower than the 1080, but beat the 1070 they probably need to match the 1070 price, again as above, to get people to move across.

3) Future Tech, AMD love to build this into their products, while this gives them longer life spans, its rubbish for the early adopters, who watch a ton of hardware sit under utilised for aeons until the software catches up, if it ever does.

4) Advertising, while this has gotten better, they need to stop the petty pot shots like the "Poor Volta" thing, its clearly aimed at Nvidia, with the latin dig at them sleeping, even coloured green so as to not be any mistake about what they are poking fun at. If your going to be that blatant, you need to product to back it up, or you are going to lose even more of your ever dwindling user base.

5) Accept your place in the market and actually warm to it, rather than trying to force everyone to accept you as the "Premium Brand" instead of telling everyone thats who you are, and then pricing your underperforming products as such, embrace your seen as the performance to pound spent king and use it, once your brand strength and confidence expands, you can actually start charging premiums people will wear, if the product warrants it. Simply saying your premium but not actually being premium, will lose you even more market share.

I think AMD are at a crossroads right now, they can potentially be the biggest news of 2017 if they get their CPU and GPU Strategies right, i think Lisa Su has done a great job so far and that gives confidence that going forward with Zen and Vega we should hopefully not see some stupid decisions made, but i have to err on caution as this is is AMD and they always drop the ball one way or another.

No one with a 1080 will switch to a vega GPU if it is only 10% faster, they will just loose a lot of money for no noticeable improvement in gaming. If AMD do release cheaper and 10% then Nvidia will likely just lower prices a little, if sales drop. this makes it even less liekly people will switch.


Vega just seems to be the wrong GPU at the wrong time, or all the ru ours are wrong.


n regards to AMD embracing the bargain basement company line, the reason AMD have tried to re-brand themselves as premium is because of profits and future R7D. Selling GPUs on the cheap may keep the market share form imploding but it is disastrous for the future. AMD have to make decent profits on GPUS sold in the mainstream and above otherwise they stay on the downward spiral of diminishing R&D and tight cashflow limitations.

 

[maonayze]

This is exactly why i wish AMD just do not ever market anything. They always make expectation so high that the only chance is that you will take a noise dive.

They always tend to compare their product and speak about their competitor more like they did in Polaris launch and compare RX 480 CF with GTX 1080, which became disaster for them. Since 3 or 4 year Nvidia has never spoken about AMD or never mentioned about AMD in their marketing or presentation and for them it is like AMD do not exist now a days. It is like AMD is always trying to seek attention at any cost.

The thing is Doom, two 480s do indeed compete within a whisker of a 1080 and beat a 1070. This is why I was thinking that if they concentrated on getting the CF drivers and patches right then that would be another way of competing with the bigger Nvidia cards (Obviously buying two cards needed to be competitively priced to the big card as well, as in slightly cheaper).

I still hope that they will now start to optimise for CF because isn't Vega 20 just two Vega 10 chips on one card? i also heard that there may be the possibility of getting them to be recognized as one card rather than two...now if they could pull that off it would change a few things (Remember Rebellions are built on hope ).

 

[D.P.]

This is exactly why i wish AMD just do not ever market anything. They always make expectation so high that the only chance is that you will take a noise dive.

They always tend to compare their product and speak about their competitor more like they did in Polaris launch and compare RX 480 CF with GTX 1080, which became disaster for them. Since 3 or 4 year Nvidia has never spoken about AMD or never mentioned about AMD in their marketing or presentation and for them it is like AMD do not exist now a days. It is like AMD is always trying to seek attention at any cost.

I noticed this too. It is a known phenomena in marketing and business. Successful companies create good products that sell themselves and can be sold based on their own attributes. Companies that are struggling tend to try and make comparisons "but we are X% faster than our competitor, have y% more special juice in our product". You don't see the likes of Apple comparing their phones to the latest Samsung etc, their phones sell themselves. Apple's marketing is about telling the consumer why they want to bu a new Iphone, not that is has 1Gb more memory or 15% faster GPU clocks than an S7. Do you don;t Apple saying their phones don't burst in to flames and are not restricted to fly in the US. They don;t need to,.

 

[Rroff]

This is exactly why i wish AMD just do not ever market anything. They always make expectation so high that the only chance is that you will take a noise dive.

They always tend to compare their product and speak about their competitor more like they did in Polaris launch and compare RX 480 CF with GTX 1080, which became disaster for them. Since 3 or 4 year Nvidia has never spoken about AMD or never mentioned about AMD in their marketing or presentation and for them it is like AMD do not exist now a days. It is like AMD is always trying to seek attention at any cost.

AMD seem to flip between too much silence and too much hype :S

 

[humbug]

AMD seem to flip between too much silence and too much hype :S

What would you call their handling of Vega so far?

 

[Doom112]

The Mindshare is such that AMD need to be better and cheaper than nVdia before just some of the people who have nVidia on their mind to even consider buying AMD instead.

The reality is it doesn't matter how good and or value for money AMD actually are a small but very loud proportion of Nvidia's mindshare are militant, with that no matter how untrue the rhetoric it stops AMD mindshare growth, which is the whole point.

The situation these days is such that AMD can't win, if even survive long term because these days marketing tactics have gone way beyond just honestly promoting your products, and if you are late to that game you are too late.

@ D.P. i think AMD's days in the consumer space are numbered... with that i think we as PC nerds are a dying breed.

AMD marketing is only to make 16 year old kid happy by giving them fantasies. For example, AMD poking Volta, where they are only competing with a GTX 1080 after a whole year. AMD market their product like they will totally destroy Nvidia ,however, in reality is totally opposite.

When AMD balance reality with their marketing than they will be a much better company.

 

[Doom112]

What would you call their handling of Vega so far?

Disaster. Do you ever seen Intel and Nvidia ever showing a demo of their unreleased product.

 

[Rroff]

What would you call their handling of Vega so far?

Mixed - they could have done with at a low key level getting some of the info about architectural changes out weeks ago or so and at this point they should atleast be showing the press a bit more "flesh" even under NDA than what they are doing.

Disaster. Do you ever seen Intel and Nvidia ever showing a demo of their unreleased product.

Nothing wrong with that as such its more about how you go presenting an unreleased product - which AMD isn't doing very well with IMO.

 

[KillBoY_UK]

No one with a 1080 will switch to a vega GPU if it is only 10% faster, they will just loose a lot of money for no noticeable improvement in gaming. If AMD do release cheaper and 10% then Nvidia will likely just lower prices a little, if sales drop. this makes it even less liekly people will switch.

I would gladly buy a Vega to replace my 980ti's if it was cheaper and as fast or faster than a 1080ti. AMD drivers have come on leaps and bounds the past 12months and are night and day since I was burnted by their drivers with my old 5850xfire setup.

 

[humbug]

AMD marketing is only to make 16 year old kid happy by giving them fantasies. For example, AMD poking Volta, where they are only competing with a GTX 1080 after a whole year. AMD market their product like they will totally destroy Nvidia ,however, in reality is totally opposite.

When AMD balance reality with their marketing than they will be a much better company.

Like what? i read a lot of criticisms about AMD yet no one seems to have any answers to accompany those criticism.

 

[D.P.]

The thing is Doom, two 480s do indeed compete within a whisker of a 1080 and beat a 1070. This is why I was thinking that if they concentrated on getting the CF drivers and patches right then that would be another way of competing with the bigger Nvidia cards (Obviously buying two cards needed to be competitively priced to the big card as well, as in slightly cheaper).

I still hope that they will now start to optimise for CF because isn't Vega 20 just two Vega 10 chips on one card? i also heard that there may be the possibility of getting them to be recognized as one card rather than two...now if they could pull that off it would change a few things (Remember Rebellions are built on hope ).

You can say the same about the 1060, that has always been the allure of SLI/XF, but reality is very different. In today games multi-gpu just isn;t viable, too many engines are just not very compatible so requires significant works form developers who are interested in a tiny tiny market segment.


I wonder if in the future multiGPu will take off when we can have large interposes with multiple GPU chips having shared access to a common memory pool of HBM3 etc., with APIs making the parallelism entirely invisible to the developer in the same way the 2000 cores of a modern GPU are seamlessly working in parallel Combining multiple smaller chips is much more economic, as will sharing memory.

 

[humbug]

You can't SLI the GTX 1060 ^^^^

Mixed - they could have done with at a low key level getting some of the info about architectural changes out weeks ago or so and at this point they should atleast be showing the press a bit more "flesh" even under NDA than what they are doing.

So reveal things like this before traditional events, so what do they then do at those events? Like CES or GDC.