AMD's Future Chips & SoC's: News, Info & Rumours.

Page 4 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.


I remember being mocked about even thinking about that possibility. Heh.

@gamerk: the Development Triangle is the same thing: Quality - Time - Cost; only 2 are economically viable at any given time during conception and implementation of a solution.

Cheers!
 
Intel has to do something in response to the Nvidia moves on driverless vehicles, and maintain it's profit margin. I'm sure AMD is a much cheaper provider while quality looks great in comparison. Bottom line, it's the better deal for Intel! And AMD users will ride high on this one! And I'm making money hand over fist today thank you AMD!
 
AMD Financial Analyst Day Presentation - Epyc - Mobile - Threadripper - Vega
by Hilbert Hagedoorn on: 05/16/2017 09:48 PM

"AMD EPYC

AMD Just announced the brand name for Data-center CPU based on Ryzen, the brand name will be called EPYC.
EPYC obviously is what previously known as "Naples". A series server processors based on"(Ry)Zen" x86 processing engine with up to 32 cores. Naples has 32 cores and is capable of 64 simultaneous threads, eight memory channels, supporting up to 2TB RAM per CPU and 128 PCIe 3.0 lanes. Memory can run 2400/2677 MHz per channel. The memory controller is also capable of using bigger than 16GB DIMMs and in total you could fit 4 TB of DDR4 memory. The Naples processor (well SoC) connects to whatever you need it through over a 128 high-speed I/O-lanes mostly Gen 3. This means you could connect 4 GPU's, 12 NVMe-SSD's and some fast 10 GigE Ethernet ports to go along with it. When you couple two processors in SMT, the IO is shared though and 64 lanes will be used for the interconnect in-between the two Naples processors. Each Naples processor has four Zen based 8-core dies interconnected.
Ryzen Mobile

Ryzen Mobile Solutions will Launch at the End of 2017 (Q4). Commercial market will see 1H 2018 availability.
Ryzen Mobile will get on-die 'Vega' cores, aka Vega based architecture.
Mark Papermaster mentions Zen 2 to be based on 7nm as well as Zen 3 (2020)
Vega is released in June says Papermaster, based on 14nm and 14nm+ node fabrication. Navi, the next-next-gen GPU architecture will be based in 7nm.
Ryzen 3 will be released in Q3 of this year
Ryzen PRO

Ryzen Pro is announced, these are Ryzen based processors for the commercial market. AMD will release them for the desktop and mobile market.
Threadripper

Jim Anderson announces 16-core / 32 thread desktop parts, threadripper. It will be available Summer 2017 and more information will become available during Computex.
AMD Radeon RX VEGA - Radeon Rising

Earlier on in the presentations Mark Papermaster mentioned Radeon Rx Vega to become available in June. Meanwhile Chief architecht of the Radeon Technology Group Raja Koduri is talking a lot about the benefits of the new Vega architecture and heterogeneous computing (cpu+gpu) and its various possible implementations. AMD things that Vega is going to be big in the data-center, he shows examples of DeepBench inbetween the NVidia P100 and Vega. Nvidia scored 133 Ms, the Vega setup 88. In this score lower is better.
AMD launches Radeon Vega Frontier Edition - the card is a mission intelligent enterprise graphics card for professional usage (data-crunching) and product designers. The card comes with 16 GB of HBM2 graphics memory and will perform in the 13 TFLOP (fp32) performance bracket. The 16GB Radeon Vega Frontier Edition will become available late June 2017."

http://www.guru3d.com/news-story/10pm-cest-amd-financial-analyst-presentation-live-feed.html

EDIT: adding YouTube Link https://www.youtube.com/watch?v=590h3XIUfHg
 


Aka the rumors of RR delayed to 2018 were correct.
 


"AMD’s first chips based on the company’s new Zen architecture are all designed for desktop computers. But if you’ve been wondering when the new chip technology would make its way to notebooks and convertible tablets, AMD has an answer: in the second half of 2017."
https://liliputing.com/2017/05/amd-ryzen-mobile-chips-raven-ridge-coming-later-year.html

"We don’t know the pricing yet, and the only availability information that we have at this point indicates that Ryzen Mobile APU-equipped notebooks should make their way to the market sometime during the second half of 2017, whereas commercial machines will not see the new chipsets until the first half of 2018."
https://www.digitaltrends.com/computing/amd-announces-ryzen-mobile-apus/

"Today on their Financial Analyst Day 2017, AMD's Mark Papermaster confirmed the existence of Ryzen mobile products. Namely, Papermaster said "(...) with Ryzen mobile coming out at the end of the year." These will launch on the second half of 2017, in two-in-1 systems, ultraportable, and gaming products. These will have integrated Zen cores and, for the first time-ever announcement, Vega graphics cores integrated into the processor. This looks to be a very, very interesting APU solution from AMD. These Zen and Vega foundations should deliver 50% more CPU performance and 40% more GPU performance, help AMD achieve an up to 50% increase in power efficiency on their mobile platformcompared to their 7th Gen APU solutions."
https://www.techpowerup.com/233384/amd-ryzen-mobile-solutions-to-launch-at-the-end-of-2017

"At the same time, AMD confirmed that Ryzen mobile accelerated processing units (APUs) codenamed Raven Ridge and offering four cores and eight threads with embedded Vega-architecture graphics would be heading to devices in the second half of the year. Finally, Ryzen Pro - the company's business-centric processor family - is scheduled to land in the second half of the year on the desktop and in the first half of 2018 for mobile devices."
https://www.bit-tech.net/news/hardware/2017/05/17/amd-epyc-threadripper-vega/1

All the news shows second half of 2017. There is a pro version slotted for Q1 2018.
0db39e7703fb.jpg
 
I just had a very couple of dumb questions in my head...

Why is it that important for a Company to miss a projection made 1+ year ago? Do we need to get ourselves prepared to go and deliver food to those poor people standing in line when the *ballpark* estimate companies gives turns out to be on the *next ballpark*?

Is there some value in discussing that or on getting it right or wrong?

I know people here owns AMD stock, but other than that, is there any value for the regular reader?

Cheers!
 
Asus ROG teases the world's first AMD Ryzen laptop
"Something has awakened."
Brad Chacos By Brad Chacos Senior Editor, PCWorld | MAY 18, 2017 12:24 PM PT

"Asus will be revealing a Ryzen gaming laptop at Computex in two weeks, in other words (and as first spotted by Overclock3D)."

"AMD’s Ryzen-based “Raven Ridge” APUs will launch in notebooks in the third quarter, but they’re paired with Radeon Vega graphics cores. "

"AMD says Ryzen laptops will provide a 50-percent boost in CPU performance, a 40-percent boost in GPU performance, and use 50 percent less power than the APUs it offers today."

http://www.pcworld.com/article/3197566/computers/asus-rog-teases-the-worlds-first-amd-ryzen-laptop.html

"Later this year AMD will launch Raven Ridge, a new APU using Zen CPU cores, but it'll also rock Vega NCUs as well. The engineering sample in question was a 4C/8T part at 3GHz base clock and 3.3GHz turbo clock, with 704 Vega NCUs at 800MHz. The Raven Ridge engineering sample featured 2MB of L2 cache and 4MB of L3 cache, while the new chip sported 11 compute units - so if we blend in 64 stream processors per CU, we should see 704 stream processors in total on the Vega NCU. The chip was pushing 572.68 Mpix/s, too."

Read more: http://www.tweaktown.com/news/57464/amd-raven-ridge-apu-4c-8t-cpu-vega-ncu-tech/index.html

I can't wait to see benchmarks!

 
AMD News Galore: Threadripper, EPYC, Ryzen Pro Processors, Integrated Vega Graphics For APUs, 7nm Ryzen Roadmap
by Paul Alcorn May 16, 2017 at 4:30 PM

"Finally, as encouraging as many of the announcements are, AMD has more to come. The company divulged that it will debut its Zen 3 products on Global Foundries' second-generation 7nm+ process in 2020. The first generation 7nm process will obviously debut in the interim, but the company did not provide a firm timeline for it."

http://www.tomshardware.com/news/amd-threadripper-epyc-pro-vega,34423.html
 


Interestingly the PCWorld author doesn't even know what is inside this laptop

If that winds up being the case, this ROG Ryzen laptop might use a more straightforward—and thus far unannounced—mobile variant of AMD’s standard Ryzen CPUs instead. It’s a long shot—AMD’s only previewed APUs thus far. Or maybe consumer Vega will launch right around the same time as the Frontier Edition. Or maybe this laptop will take a while to actually launch. Time will tell!
Click to expand...

We know the mobile Zen APUs are scheduled for the holiday season, and there is some rumor of that this laptop could use an underclocked 8-core Zen CPU (45W TDP) paired with a mobile GPU.
 


Very nice, I still hold by first statement on the need for such monsters, but I guess if you can afford it and you want to future proof. New unreal engine can do theorhetically hundreds of threads if you got the hardware, not sure about the other top two like new ID or upcoming source 2.

 


Even current games use a good 70-80 threads; threading is NOT hard. The problem is most of them don't do any real amount of work. Now, having more cores still helps as you aren't interrupting the heavy threads as much (latency).
 

Can you plz list game engines that use more then 8 threads atm? I can only think of next gen engines using this many.

And from what I heard and my little coding experience for devs its a lot harder to code in multithread, so I dont know about it not being hard. The latency issues also are part of the coding dificulty from what I gather if not mistaken
 


He's taking it form the light threading point of view, where you share the same memory space and don't go out of it. I'm willing to say most of those threads are just libraries spawning their own things, but not the developer's threads.

When you spawn heavy threads, the OS treats them differently. You can see nowadays that regular programs spawn sub-processes that are a full new program still linked to the main one. Best example is Chrome. That will still spawn light threads on its own, but still have a bunch of heavy threads using CPU.

Cheers!
 

Ahhhh!!!
Yeah I meant as in say 8 threads use 8 full cores
 


In regards to Chrome and other browsers: Spawning new processes was a way to get around the old 2GB Win32 memory barrier, but has absolutely no place in modern computing. That's the primary reason browsers are such memory hogs; you're duplicating the Win64 APIs every time you open a new tab, by virtue of treating it as a new process.

In regards to threads: Feel free to add the "threads" category to Task Manager. And while a lot of those threads are API threads, just as many are spawned by the developer. Most threads don't do a whole lot of work; they take care of some background task that is reasonably independent of the main program.
 


I'm not disagreeing, my dear gamerk. I can see where you're coming from and it's totally valid.

It's just that the implication (like he clarified already) is about threading the parts of the programs that do meaningful work. That depends on design and implementation, so it's really a "case by case". We can't really put every game in the same bag, but there are things we know could be done better (thread wise).

I'll just keep the can of worms a bit closed for now, but I think we can agree moving to more threading (process heavy threads, that is) looks to be getting easier. Even Java introduced changes in its API with Java 8 to make threading easier.

Cheers!
 


But I'm arguing it was never hard in the first place; all you need is to invoke CreateThread (lets disregard any non-Windows OS for a second) and boom, thread. Not that complicated. What's hard is distributing workload in such a way where those threads do any meaningful amount of work, and ensuring you don't end up blocking for one reason or another.

The increase in threading you're seeing has more to do with the fact the rendering pipeline has been broken up in DX12/Vulkan, breaking up that massive GPU rendering thread into smaller chunks. Aside from that, nothings changed.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom