AMD Confirms, Polaris GPUs Will Be Released Before The Back To School Season (WCCF)

[flopper]

That is the most optimistic interpretation. Realistic?

Absolutly.
Polaris is demoed already so AMD is ready, Nvidia pascal is absent.
I find people funny, 10nm or 7nm? that level of tech is slowing down and each node will last longer due to the same reason as always, even for intel its a huge cost to shrink down and build fabrics for the tech.
I wonder what people say when 10 or 7nm might be it and a whole new tech is needed to overcome the technical difficulties and the cpu simply stop evolving and Moore died.

 

[Kenmitch]

Additionally, they expect to have ZEN processors available in limited quantities by the end of 2016 with production ramping up in early 2017.

In the world of the CPU limited quantities means what?

1,000
10,000
50,000
100,000
???

Time will tell I guess. Looking forward to at least see what these alleged chips can or can't do.

 

[raghu78]

That is the most optimistic interpretation. Realistic?

Given the fact that AMD received production samples back in early Nov 2015 and the fact that they demoed silicon running at CES 2016 in 1st week of Jan 2016 I would say the first Polaris chips taped out sometime in late July/early Aug 2015. It takes 10-12 weeks from tapeout to first silicon.

http://www.kitguru.net/components/a...cts-using-second-gen-14nm-process-technology/

http://globalfoundries.com/newsroom...logy-success-for-next-generation-amd-products

Normally it takes 9-12 months from tapeout to chips on the market. So late June/early July 2016 is a very realistic timeline for launch of the smaller Polaris chips.

 

[Phynaz]

Said by Lisa Su
we went to FinFET relatively early

Huh?

 

[MeldarthX]

Given the fact that AMD received production samples back in early Nov 2015 and the fact that they demoed silicon running at CES 2016 in 1st week of Jan 2016 I would say the first Polaris chips taped out sometime in late July/early Aug 2015. It takes 10-12 weeks from tapeout to first silicon.

http://www.kitguru.net/components/a...cts-using-second-gen-14nm-process-technology/

http://globalfoundries.com/newsroom...logy-success-for-next-generation-amd-products

Normally it takes 9-12 months from tapeout to chips on the market. So late June/early July 2016 is a very realistic timeline for launch of the smaller Polaris chips.

For graphics it can be as early as 6 months; with reports that AMD is about 2 quarters ahead on node shrink; Nvidia's never been good at node shrinks.

We could easy small Polaris as early as march/april and big around June. AMD's said their target is to have full stack out for back to school.

While would be awesome to see something March/April - very well full stack could be launched in June ......

 

[Kenmitch]

Huh?

When was the ability to go to finfet 1st available for AMD to use?

I doubt Intel offered to produce chips for AMD.

 

[raghu78]

For graphics it can be as early as 6 months; with reports that AMD is about 2 quarters ahead on node shrink; Nvidia's never been good at node shrinks.

We could easy small Polaris as early as march/april and big around June. AMD's said their target is to have full stack out for back to school.

While would be awesome to see something March/April - very well full stack could be launched in June ......

On a brand new 14LPP process node and a new Polaris architecture I would say 6 months will never happen. We are already 6 months from tapeout. If AMD needs another stepping to solve some bugs and improve efficiency and yields I would not be surprised. I would say late June and early July 2016 is very realistic and right in time for back to school.

I do not expect big Polaris with HBM2 any time before Sep/Oct 2016. I expect the big Polaris to be a 4096 sp (GCN 2.0), 8GB HBM2, 2048 bit memory bus, 512 GB/s flagship card with a die size of 300-350 sq mm.

http://www.kitguru.net/components/g...mory-chips-opens-way-for-32gb-graphics-cards/

HBM2 is expected to start production in Q2 2016 and it takes 6 months for the first chips on market from start of production. Interposer based 2.5D manufacturing is more complex and takes longer turn around times.

 

[MeldarthX]

On a brand new 14LPP process node and a new Polaris architecture I would say 6 months will never happen. We are already 6 months from tapeout. If AMD needs another stepping to solve some bugs and improve efficiency and yields I would not be surprised. I would say late June and early July 2016 is very realistic and right in time for back to school.

I think its also realistic for big Polaris; but for little; you don't show of prototypes at a major event if you're not going to launch in next couple months...... Specially if they need a respin - which looks like they don't. I'm more than confident to say we should say small Polaris around April; big June/July .......as AMD's said they want to make sure they are in skus; laptops for back to school. Means they have to moving from prototypes to full production now......

 

[MeldarthX]

On a brand new 14LPP process node and a new Polaris architecture I would say 6 months will never happen. We are already 6 months from tapeout. If AMD needs another stepping to solve some bugs and improve efficiency and yields I would not be surprised. I would say late June and early July 2016 is very realistic and right in time for back to school.

I do not expect big Polaris with HBM2 any time before Sep/Oct 2016. I expect the big Polaris to be a 4096 sp (GCN 2.0), 8GB HBM2, 2048 bit memory bus, 512 GB/s flagship card with a die size of 300-350 sq mm.

http://www.kitguru.net/components/g...mory-chips-opens-way-for-32gb-graphics-cards/

HBM2 is expected to start production in Q2 2016 and it takes 6 months for the first chips on market from start of production. Interposer based 2.5D manufacturing is more complex and takes longer turn around times.

Also AMD's always handled node shrinks very well.....I very much doubt they'd blow this

 

[mysticjbyrd]

I do not expect big Polaris with HBM2 any time before Sep/Oct 2016. I expect the big Polaris to be a 4096 sp (GCN 2.0), 8GB HBM2, 2048 bit memory bus, 512 GB/s flagship card with a die size of 300-350 sq mm.

http://www.kitguru.net/components/g...mory-chips-opens-way-for-32gb-graphics-cards/

HBM2 is expected to start production in Q2 2016 and it takes 6 months for the first chips on market from start of production. Interposer based 2.5D manufacturing is more complex and takes longer turn around times.

The chart on your source seems to imply production will begin very early into the second Quarter, maybe April.

Where did you get the 6 month time frame from?

 

[MeldarthX]

well as someone posted out in its own thread; I'll put this here also......Big Polaris was also spotted at CES.......thus putting more into my theory we'll see them sooner than later

http://www.tweaktown.com/news/49581...-next-gen-polaris-spotted-ces-2016/index.html

 

[LTC8K6]

That is the most optimistic interpretation. Realistic?

If I can't believe a direct quote from Su, then what can I believe?

 

[guskline]

Interested in what the Big Polaris brings to the table.:thumbsup:

 

[moonbogg]

Interesting that they expect 14nm to be a long node. We know a single node can provide and meet expectations for a long time as we saw with 28nm. I guess 14nm should stretch at least that far. I think this next gen is going to be very good.

 

[Azix]

From the PC World article:

"AMD is establishing a number of products on 14-nm FinFET in 2016, Su said. “But it will be a long node. It will last three, four, five years.""

Seeing as this was in a section entitled "Pulling even with Intel," that's not a good sign. Intel is scheduled to hit 10-nm next year, with 7-nm in 3 or 4 years, and an outside chance of 5-nm in 5 years. 3 or 4 years at 14-nm may be ok on the GPU side, but Intel will be burying them on the CPU side even worse than they are now. If AMD wants to still be solvent in 5 years, they need to compete better in the mobile space where everything is heading and where they are a complete non-factor for both GPU's and CPU's now. Being 2 or 3 nodes behind is not an effective strategy.

AMD doesn't have a fab, so its not like they can do anything about that. I doubt intel will make much progress for a while though. Things are going to be harder and harder

 

[swilli89]

I would be over joyed if AMD wound up at the end of 2016 with great positioning in the graphics front with the Polaris family of GPUs and a ZEN CPU that surprises everyone.

Here's hoping one of our hobby's most valuable players starts its journey back to health in 2016:thumbsup:

 

[Vesku]

During the chat, Koduri [Chief Architect of Radeon Technologies Group] said:

We have two versions of these FinFET GPUs. Both are extremely power efficient. This is Polaris 10 and that’s Polaris 11. In terms of what we’ve done at the high level, it’s our most revolutionary jump in performance so far. We’ve redesigned many blocks in our cores. We’ve redesigned the main processor, a new geometry processor, a completely new fourth-generation Graphics Core Next with a very high increase in performance. We have new multimedia cores, a new display engine.

In summary, it’s fourth generation Graphics Core Next. HDMI 2.0. It supports all the new 4K displays and TVs coming out with just plug and play. It supports display core 4.3, the latest specification. It’s very exciting 4K support. We can do HAVC encode and decode at 4K on this chip. It’ll be great for game streaming at high resolution, which gamers absolutely love. It takes no cycles away from games. You can record gameplay and still have an awesome frame rate. It’ll be available in mid-2016.

http://techfrag.com/2016/01/17/amds...ris-10-the-most-power-efficient-gpus-to-date/