AMD Roadmap Hints At Three GPU Launches By 2018

[kcarbotte]

AMD revealed an updated roadmap for its future GPUs. The company has been criticized for being slow out of the gate with its upcoming Polaris launch, but if the roadmap is any indication, we’re about to see an acceleration in the company’s pace.

AMD Roadmap Hints At Three GPU Launches By 2018 : Read more

 

[InvalidError]

I believe it is easy enough to figure out what "scalability" means: HBM pushed the memory interface width to the limits of what is practical to manufacture, HBM3 or whatever comes around after HBM2 will push data rates to the limits of what is electrically practical too. If you cannot push the bandwidth between a processor and its memory much further, the only direction left is parallel chips: high bandwidth interconnects between chips with an architecture optimized for distributed load sharing.

 

[junkeymonkey]

'' AMD Roadmap Hints At Three GPU Launches By 2018 ''

so does this mean AMD's motto went from ''next year'' to now in ''2 years '' ?? lol......

I was reading a article of amd and I laughed cause every third sentence it had next year bla bla bla next year ,ect ....... lol

how bout something today or at least this year ??

AMD, Enabling today.Inspiring tomorrow ,next year

 

[turkey3_scratch]

Junkeymonkey, I'd say AMD is not the only company getting slower. Intel keeps holding off Broadwell-E, they are now on a 2.5 year tick tock cycle. Nvidia's Pascal might release after Polaris cards, and what was initially thought of as Pascal cards in 2015, we're now looking realistically at late 2016. Every PC company seems to be getting slower, not just AMD.

It also looks like now that HBM2 won't be featured in Polaris, bummer. I think Pascal is strictly using GDDR5X, maybe 1 HBM card.

No way in heck do I expect Vega GPUs to be released early 2017. Expect late 2017, or early 2018.

 

[ern88]

I really want something this year that would be the same price of what I paid for my Sapphire HD7950 vapor x. And will give me about 50% plus performance of that card.

 

[voodoochicken]

I'm just bumming with on-board graphics of my z170 board I just got until SOMETHING gets released.

 

[junkeymonkey]

''Intel keeps holding off Broadwell-E, they are now on a 2.5 year tick tock cycle. Nvidia's Pascal might release after Polaris cards,'

well don't you got the great skylake and slap that Maxwell behind it [combined with win-10] ?? its not like AMD is pushing them. they [intel/nvidia] will release things once they see its time to got more fresh money from you .. dang man they got to make there money and get all they can from you on what they offer you now first

 

[turkey3_scratch]

[quotemsg=17665239,0,252980]I really want something this year that would be the same price of what I paid for my Sapphire HD7950 vapor x. And will give me about 50% plus performance of that card. [/quotemsg]

With a die shrink by a multiple of 2 we should be expecting hopefully 150% more performance IMO.

 

[heinlein]

"The company said that Polaris is expected to deliver up to 2.5x performance per watt, but the first products will be targeted at notebook computers."

Unless a VR capable Polaris is available when I get my pre-ordered Oculus I will be buying a Nvidia card. I know others feel the same. Don't blow this opportunity AMD,

 

[turkey3_scratch]

I expect Polaris to come before Pascal.

 

[kcarbotte]

[quotemsg=17665145,0,125865]I believe it is easy enough to figure out what "scalability" means: HBM pushed the memory interface width to the limits of what is practical to manufacture, HBM3 or whatever comes around after HBM2 will push data rates to the limits of what is electrically practical too. If you cannot push the bandwidth between a processor and its memory much further, the only direction left is parallel chips: high bandwidth interconnects between chips with an architecture optimized for distributed load sharing.[/quotemsg]

And you might be on to something, but I'm not about to put that in a news post. We're not in the business of rampant speculation. We take what's given to us and interpret what we have. estimating the performance number of a graph without labels is a far cry from making broad assumptions.

 

[junkeymonkey]

I think like before it will get delayed just right up to just in time for black Friday , say ,mid October ??... all the hype built up and all that x-mas money burning a hole in your pocket

 

[hotroderx]

Junkeymonkey, I'd say AMD is not the only company getting slower. Intel keeps holding off Broadwell-E, they are now on a 2.5 year tick tock cycle. Nvidia's Pascal might release after Polaris cards, and what was initially thought of as Pascal cards in 2015, we're now looking realistically at late 2016. Every PC company seems to be getting slower, not just AMD.

It also looks like now that HBM2 won't be featured in Polaris, bummer. I think Pascal is strictly using GDDR5X, maybe 1 HBM card.

No way in heck do I expect Vega GPUs to be released early 2017. Expect late 2017, or early 2018.

A company stand point this makes since spread the cost of research out over a longer time period, while milking every dime you can from your current products!. Sadly right now Intel and Nvidia don't have extremely fierce competition.


They know this and also know that shutting AMD down completely would be bad for business all around. Since most likely the federal Trade commission would step in and do something to both companies saying they didn't foster a healthy competitive environment are some other bs.

This is just the way of things until AMD can become healthier as a company. I dont think its the technology has stalled its more why spend the money if you dont have to. Cause at the end of the day lets be honest Intel and Nividia do this for the stock holders!. I also think the only reason we hadn't seen a steady climb in pricing is because Intel knows with there current releases (same with nvidia) that if they raise pricing people will just be satisfied with what they have vs keeping pricing the same and suddenly people can justify the 2 year upgrade for a marginal boost in performance.

 

[kcarbotte]

[quotemsg=17665294,0,68688]"The company said that Polaris is expected to deliver up to 2.5x performance per watt, but the first products will be targeted at notebook computers."

Unless a VR capable Polaris is available when I get my pre-ordered Oculus I will be buying a Nvidia card. I know others feel the same. Don't blow this opportunity AMD,[/quotemsg]

AMD said there are desktop GPUs coming, but it remains to be seen if those will outperform the Fury X / 980Ti or if they will be around the same performance or lower, but with significantly lower price points.

 

[ern88]

So will we see any high end GPU'S from AMD or Nvidia this year.?

 

[kcarbotte]

[quotemsg=17665424,0,252980]So will we see any high end GPU'S from AMD or Nvidia this year.?[/quotemsg]

Only people working inside those companies know that right now.

 

[hannibal]

New manufacturing process, so it is safer to make low end GPUs first. But lets see. The HBM2 is coming quite late to full production, so it is expensive and rare before Hynix and Samsung force each other to compete. So that means 2017 will be high-end year.
Also GDDR5+ is coming late this year, so even that is going to go 2017, and if it is much more expensive than normal gddr5, we are not seeing it in middle range cards soon.
Seems to be that 980ti and Fury cards are for VR in this year. And it seems that in VR AMD has god position at this moment.

 

[photonboy]

justeymonkey,
I don't understand what your POINT is. They are launching Polaris this year. There are always new generations on the roadmap for several years.

Again, you asked why not "at least this year", well 2016 is THIS year and that's when Polaris is on the chart. So...

 

[kcarbotte]

[quotemsg=17665780,0,60597]New manufacturing process, so it is safer to make low end GPUs first. But lets see. The HBM2 is coming quite late to full production, so it is expensive and rare before Hynix and Samsung force each other to compete. So that means 2017 will be high-end year.
Also GDDR5+ is coming late this year, so even that is going to go 2017, and if it is much more expensive than normal gddr5, we are not seeing it in middle range cards soon.
Seems to be that 980ti and Fury cards are for VR in this year. And it seems that in VR AMD has god position at this moment.[/quotemsg]

GDDR5X should be similar in price, if not cheaper than current GDDR5. That was part of the motivation for it.
I talked to Micron in the fall about the memory and it was aluded that this memory will be for lower tier cards to keep the cost down, while the high end moves to HBM.

 

[Knicks2012]

My warranty for my GPU ends in 2018, will be cool to see the new GPUs from AMD and Nvidia when Im looking to jump to 4k.

 

[InvalidError]

[quotemsg=17665273,0,1712875]With a die shrink by a multiple of 2 we should be expecting hopefully 150% more performance IMO.[/quotemsg]
Area-wise, that would be a die shrink by a factor of 4X: going from 28nm to 14nm is half the width and half the length.

In principle, they should be able to pack 3-4X as much performance in the same die area but if they did that, they would only get one product cycle out of 14-16nm before being stuck near the process' physical limits. As much as I would like the first 14nm step to provide 150% more performance for a given die size or price point, I doubt we are going to see more than 100% and even that might be too optimistic.

In the past, new GPU generations only brought 40-66% more performance to a given price point.

 

[none12345]

"how may times have we had to hear all that and nothing to show ?? wheres zen that's been talked up seems like 5 years now ?? come on"

Hrm, i dont think zen has been mentioned for anything close to 5 years, i want to guess about 2 years, but im not sure if its even been that long. The wiki page for zen was created in may of 2015, so thats not even a year ago.

Maybe it just seems like 5+ years because the cpu industry has been so boring for the last 5+ years. Not just amd, but intel as well, nothing exciting for quite some time now.

Same with the gpu industry for the last 3 years or so. Sure nvidia had a nice reduction in power usage with their last set of gpus, but performance wise, being stuck on 28nm for both amd/nvidia has sucked big time. Again the graphics industry has been pretty darn boring the last couple years as well for that reason.

Process node issues as of late have just been so depressing. Its not like the good old days where chips were twice as fast every year or every other year at most. I miss those days!

 

[junkeymonkey]

guys say the price of cards today is high but think back when a 8800gtx was over 500 bucks as the mac daddy card now look at what you get today for 550 bucks -- heck a low mid range card at 170 bucks is more powerful then the 8800 - I think you got a lot of power today for the same mony you spent way back then [price point]

 

[vaughn2k]

HBM1 and HBM2 is using a PCB interposer together with the GPU. I think what it meant in the Next Gen memory is that, AMD will be using TSV technology to connect the HBM into the GPU itself, further reducing latency... Its a guess, but its probable..

 

[MrKB]

I really want something this year that would be the same price of what I paid for my Sapphire HD7950 vapor x. And will give me about 50% plus performance of that card.

I just bought a Sapphire R9 390 8GB and I love it. I think you'd get about a 30-40% boost depending on what you used your card for. If you want that 50% I would consider the R9 Nano, it's a little more expensive but there are some benefits to that card as well.