AMD Achieves High-End Embedded Performance Leadership with New R-Series

[btarunr]

AMD today announced new AMD Embedded R-Series SOC processors that establish performance leadership across a targeted range of embedded application market requirements for digital signage, retail signage, medical imaging, electronic gaming, media storage and communications and networking. Designed for demanding embedded needs, the processors incorporate the newest AMD 64-bit x86 CPU core ("Excavator"), plus third-generation Graphics Core Next GPU architecture, and state-of-the-art power management for reduced energy consumption. Combined, these AMD innovations and technologies provide industry-leading graphics performance and key embedded features for next-generation designs.

The single-chip system-on-chip (SOC) architecture enables simplified, small form factor board and system designs from AMD customers and a number of third party development platform providers, while providing astounding graphics and multimedia performance, including capability for hardware-accelerated decode of 4K video playback. With a robust suite of peripheral support and interface options, high-end AMD Radeon graphics, designed for the industry's first Heterogeneous Systems Architecture (HSA) 1.0 certification, and support for the latest DDR4 memory, the new AMD R-Series SOC addresses the needs of a wide range of markets and customers.

"AMD continues its push into x86 embedded platforms and that's paying off with an increasing number of customers and applications," said Jim McGregor, principal analyst, TIRIAS Research. "There is a need for immersive graphics, high-quality visualization, and parallel computing in an increasing number of embedded applications. Across these fronts, the AMD Embedded R-Series SOC is a very compelling solution."

"With so much momentum around immersive experiences, especially for visual and parallel computing, the embedded industry needs a high-performance, low-power and efficient architecture with superior graphics and compute capabilities," said Scott Aylor, corporate vice president and general manager, AMD Embedded Solutions. "Our new AMD Embedded R-Series SOC is a strong match for these needs in a variety of industries including digital signage, retail signage, medical imaging, electronic gaming machines, media storage, and communications and networking. The platform offers a strong value proposition for this next generation of high-performance, low-power embedded designs."

Industry-Leading Graphics
With the latest generation AMD Radeon graphics as well as the latest multimedia technology integrated on-chip, the AMD Embedded R-Series SOC provides enhanced GPU performance and support for High Efficiency Video Coding (HEVC)1 for full 4K decode and DirectX 12. The new AMD Embedded R-Series SOCs offer 22 percent improved GPU performance when compared to the 2nd Generation AMD Embedded R-Series APU2 and a 58 percent advantage against the Intel Broadwell Core i7 when running graphics-intensive benchmarks3. Specifications for the integrated AMD Radeon graphics include:

• Up to eight compute units4 and two rendering blocks
• GPU clock speeds up to 800MHz resulting in 819 GFLOPS
• DirectX 12 support

Fully HSA Enabled
Customers in several industries such as machine learning, medical imaging and digital signage often need to execute compute intensive, parallel processing algorithms. HSA is a standardized platform design that unlocks the performance and power efficiency of the GPU as a parallel compute engine. It allows developers to more easily and efficiently apply the hardware resources in today's SoCs, enabling applications to run faster and at lower power across a range of computing platforms. The AMD Embedded R-Series platform incorporates a full HSA implementation which balances the performance between the CPU and GPU. Leveraging the heterogeneous Unified Memory Architecture (hUMA) allows for reduced latencies and maximizes memory access to both the CPU and GPU to increase performance.

Designed for Embedded
The AMD Embedded R-Series SOC was architected with embedded customers in mind and includes features such as industrial temperature support, dual-channel DDR3 or DDR4 support with ECC (Error Correction Code), Secure Boot, and a broad range of processor options to meet an array of embedded needs. Additionally, configurable thermal design power (cTDP) allows designers to adjust the TDPs from 12W to 35W in 1W increments for greater flexibility. The AMD Embedded R-Series SOC also has a 35 percent reduced footprint when compared to the 2nd Generation AMD Embedded R-Series APU, making it an excellent choice for small form factor applications.
Key features and specifications include:

• First embedded processor with dual-channel 64-bit DDR4 or DDR3 with Error-Correction Code (ECC), with speeds up to DDR4-2400 and DDR3-2133, and support for 1.2V DDR4 and 1.5V/1.35V DDR3
• Dedicated AMD Secure Processor supports secure boot with AMD Hardware Validated Boot (HVB); initiates trusted boot environment before starting x86 cores
• High-performance Integrated FCH featuring PCIe Gen3 USB3.0, SATA3, SD, GPIO, SPI, I2S, I2C, UART

The AMD Embedded R-Series SOC provides industry-leading ten-year longevity of supply. The processors support Microsoft Windows 7, Windows Embedded 7 and 8 Standard, Windows 8.1, Windows 10, and AMD's all-open Linux driver including Mentor Embedded Linux from Mentor Graphics and their Sourcery CodeBench IDE development tools.

View at TechPowerUp Main Site

 

[cdawall]

Ok sweet release an AM1 version.

 

[Patriot]

Ok sweet release an AM1 version.

Yeaaaah such a big sell for upgradable platform... and no more chips made for it.

 

[Musaab]

When will they stop this money bleeding business. they should focus on one or two products and finish them then market them. A medium company like AMD and the companies in the same size can't deal with many projects at the same time. Nvidia for example built all their business around single product and make change to drivers (GeForce, Quadro and Tesla) or adding others products to their own (Tegra and Grid). AMD is almost the same size as Nvidia and need to do the same. Nvidia is now dealing with single design which is Maxwell while AMD is dealing with CGN1, CGN1,1 and CGN 1,2 in their production line, they need to stop this fragmentation, and focus before it is too late.

 

[Mysteoa]

When will they stop this money bleeding business. they should focus on one or two products and finish them then market them. A medium company like AMD and the companies in the same size can't deal with many projects at the same time. Nvidia for example built all their business around single product and make change to drivers (GeForce, Quadro and Tesla) or adding others products to their own (Tegra and Grid). AMD is almost the same size as Nvidia and need to do the same. Nvidia is now dealing with single design which is Maxwell while AMD is dealing with CGN1, CGN1,1 and CGN 1,2 in their production line, they need to stop this fragmentation, and focus before it is too late.

First there is not such ting as GCN 1.1 and 1.2. This naming was made from the community. Second they are working on Arctic island which will be on 16/14 nm and surely will have bigger changes in the GCN then the previous minor changes.

 

[cdawall]

Yeaaaah such a big sell for upgradable platform... and no more chips made for it.

Exactly it has what 3 processors?

 

[Musaab]

First there is not such ting as GCN 1.1 and 1.2. This naming was made from the community. Second they are working on Arctic island which will be on 16/14 nm and surely will have bigger changes in the GCN then the previous minor changes.

If want to name them Hawaii, Fiji, ....,ect it's still the same problem they are dealing with three different generations on this production line at the same time.
Adding Arctic Island will complicates the matter more. They should work on one generation and make all the chips from that generation by changing cores number to cuts the cost. The same can be said about CPUs.

 

[uuuaaaaaa]

If want to name them Hawaii, Fiji, ....,ect it's still the same problem they are dealing with three different generations on this production line at the same time.
Adding Arctic Island will complicates the matter more. They should work on one generation and make all the chips from that generation by changing cores number to cuts the cost. The same can be said about CPUs.

http://wccftech.com/amd-arctic-islands-16nm-launch-coming-quarters/

 

[Musaab]

http://wccftech.com/amd-arctic-islands-16nm-launch-coming-quarters/

I hope that link is true and not just changing the manufacturing process from 28nm to 14/16nm. And this rise a question, are they planning to use the HBM in all their products or they will preserve it for high-end and take NVIDIA's approaches of the DDR5X. Because HBM's cost is still high and they still can't make enough.

 

[uuuaaaaaa]

I hope that link is true and not just changing the manufacturing process from 28nm to 14/16nm. And this rise a question, are they planning to use the HBM in all their products or they will preserve it for high-end and take NVIDIA's approaches of the DDR5X. Because HBM's cost is still high and they still can't make enough.

I think that they will have HBM in their high tiered products only and use GDDR5 for everything that is not meant for 4k and above.

 

[Musaab]

I think that they will have HBM in their high tiered products only and use GDDR5 for everything that is not meant for 4k and above.

And above??????? No way. Not in the next three years.

 

[Nkd]

If want to name them Hawaii, Fiji, ....,ect it's still the same problem they are dealing with three different generations on this production line at the same time.
Adding Arctic Island will complicates the matter more. They should work on one generation and make all the chips from that generation by changing cores number to cuts the cost. The same can be said about CPUs.

Do you really have any idea what you are talking about. Those are code names and nvidia has codnames for products. Arctic island is a new gpu design. They can call it whatever they want. Nvidia does the same. I have no idea what you are saying though. They are already doing what you are saying, 390 a cut down version of 390x and fury a cut down version fur x. What exactly are you trying to get to here?

 

[Musaab]

Do you really have any idea what you are talking about. Those are code names and nvidia has codnames for products. Arctic island is a new gpu design. They can call it whatever they want. Nvidia does the same. I have no idea what you are saying though. They are already doing what you are saying, 390 a cut down version of 390x and fury a cut down version fur x. What exactly are you trying to get to here?

WE CAME IN PEACE

You are wrong Fury/X/Nano are the same chip let us call it GCN 1.2 from 2015, 390/X are the same chip that was found in 290/X and let us call it GCN 1.1 from year 2013, 380/X are the same chip as the 280/X which was originally found in the 7970/7950 we can call it GCN 1.0 from year 2011 they only disable some cores through wiring, Bios or driver.
for Nvidia's side the 980Ti/TITAN contains chip called GM210, 980/970 GM204, 960/950 GM206. all these chips are build on the Maxwell cores v2 of 2014, They share the core design but with totally different DIE size and cores number.
So AMD has to deal with three different generations of cores design, While Nvidai Use only one core Design for all their Graphics Products. more over AMD's share toke that huge dive because most people don't want to get last generation rebranded cards, they want something new despite the performance per dollar advantage AMD has.

 

[uuuaaaaaa]

You are wrong Fury/X/Nano are the same chip let us call it GCN 1.2 from 2015, 390/X are the same chip that was found in 290/X and let us call it GCN 1.1 from year 2013, 380/X are the same chip as the 280/X which was originally found in the 7970/7950 we can call it GCN 1.0 from year 2011 they only disable some cores through wiring, Bios or driver.
for Nvidia's side the 980Ti/TITAN contains chip called GM210, 980/970 GM204, 960/950 GM206. all these chips are build on the Maxwell cores v2 of 2014, They share the core design but with totally different DIE size and cores number.

IIRC R9 380 is based on a cut down Tonga (R9 285) and the allegedly R9 380x (unreleased) is full Tonga. Tonga is GCN 1.2.

 

[Musaab]

IIRC R9 380 is based on a cut down Tonga (R9 285) and the allegedly R9 380x (unreleased) is full Tonga. Tonga is GCN 1.2.

You are right the 380/X is/will be GCN 1.2 Tonga but the 270x is the three generation old card. Thanks uuuaaaaa.

 

[IamEzio]

WE CAME IN PEACE


You are wrong Fury/X/Nano are the same chip let us call it GCN 1.2 from 2015, 390/X are the same chip that was found in 290/X and let us call it GCN 1.1 from year 2013, 380/X are the same chip as the 280/X which was originally found in the 7970/7950 we can call it GCN 1.0 from year 2011 they only disable some cores through wiring, Bios or driver.
for Nvidia's side the 980Ti/TITAN contains chip called GM210, 980/970 GM204, 960/950 GM206. all these chips are build on the Maxwell cores v2 of 2014, They share the core design but with totally different DIE size and cores number.
So AMD has to deal with three different generations of cores design, While Nvidai Use only one core Design for all their Graphics Products. more over AMD's share toke that huge dive because most people don't want to get last generation rebranded cards, they want something new despite the performance per dollar advantage AMD has.

Maybe all the GTX9XX series are Maxwell. but they have alot more cards than that currently selling, Desktop "GT" series of GPU's that half of them are not even based of kepler . and in the best case they have a Kepler "version" with the same name, and when you buy a cheap laptop with "better than integrated by 5 percent " discrete gpu like the "Geforce 920" you actually get a 730M or something like that just 3 years later , nvidia has so much confusion in the lower end market about models with the same name and based on different chips that its amazing nobody saying a word , so what if this is low end stuff. i am sure this is half of their market , because lets make it clear that alot more people buy cheap laptops with cheap GPU's than high end gaming machines . also for AMD all of the RX3XX and RX2XX are based on GCN , even if it is missing some features in when it comes to the GCN 1.0 cards still sold its still mostly the same architecture and has the same support and optimizations .

 

[uuuaaaaaa]

Maybe all the GTX9XX series are Maxwell. but they have alot more cards than that currently selling, Desktop "GT" series of GPU's that half of them are not even based of kepler . and in the best case they have a Kepler "version" with the same name, and when you buy a cheap laptop with "better than integrated by 5 percent " discrete gpu like the "Geforce 920" you actually get a 730M or something like that just 3 years later , nvidia has so much confusion in the lower end market about models with the same name and based on different chips that its amazing nobody saying a word , so what if this is low end stuff. i am sure this is half of their market , because lets make it clear that alot more people buy cheap laptops with cheap GPU's than high end gaming machines . also for AMD all of the RX3XX and RX2XX are based on GCN , even if it is missing some features in when it comes to the GCN 1.0 cards still sold its still mostly the same architecture and has the same support and optimizations .

There are two GTX860m, one uses kepler and the other maxwell...
http://www.notebookcheck.net/Review-Nvidia-GeForce-GTX-860M-Maxwell-vs-Kepler.114908.0.html

 

[Musaab]

Maybe all the GTX9XX series are Maxwell. but they have alot more cards than that currently selling, Desktop "GT" series of GPU's that half of them are not even based of kepler . and in the best case they have a Kepler "version" with the same name, and when you buy a cheap laptop with "better than integrated by 5 percent " discrete gpu like the "Geforce 920" you actually get a 730M or something like that just 3 years later , nvidia has so much confusion in the lower end market about models with the same name and based on different chips that its amazing nobody saying a word , so what if this is low end stuff. i am sure this is half of their market , because lets make it clear that alot more people buy cheap laptops with cheap GPU's than high end gaming machines . also for AMD all of the RX3XX and RX2XX are based on GCN , even if it is missing some features in when it comes to the GCN 1.0 cards still sold its still mostly the same architecture and has the same support and optimizations .

I do agree with you. Nvidia has a terrible rebranding strategy specially in laptops 7xx-8xx, they do rebranding in business sector a lot (GT 645), but it was profitable moves take "Geforce 920" it's manufactured on an old node for dirt cheap, the foundry get some profit from old nods that has deserted by the big players, Nvidia got cheap GPU for almost the cost of the silicon, the laptop manufacturer can install discreet GPU for maybe 15$ and stick Green label and last the customers get discreet GPU for 20-25$ above the price of the same laptop without discreet GPU and that is not bad. some time the companies get a big over stock of the sub 100$ chips after the end of a generation (GTX550 Fermi GF116) for example became GT645 OEM, what they did was re branding them and sell them to OEMs who usually install them in business PCs where there is no use of discreet GPU. But AMD's problem is that they need to work on old chips to upgrade them to get some extra performance, so they have to design the new chip and at the same time work to upgrade the old chip for rebanding and that is a lot of work, Nvidia's chips are modular like LEGO you can make it big, you can make it small at the end they will work and the fabs are happy because they in fact make the same design only change the size and customer know that the small cards are as good and modern as the big once some times better "GTX 750/Ti" . that is why Nvidia keep selling a lot of their over priced cards.

 

[Trovaricon]

I am not sure if it should be considered funny or not, but now announced Sapphire evaluation platform DB-FP4(-D4) board (for R series in announcement) is considered by me to be the best layout / display connectivity SFF 'big' APU board available.
It is slim&short mATX form factor (170x226mm) with 4 DIMMs! With a little imagination you can actually fit this inside the '7 liter' cases like Silverstone Milo 05/06 or InWin! BM series.

- M2 support
- Spacing between PCI-e slots
- 3!!! Display ports (AFAIK no retail board has 3 digital outputs - how sad is this?)

Now the question is: wth are board makers producing for the market? Except of the M2 the rest could have been in FM2+ boards years ago.

The only problem with that Sapphire board is, that I expect it to cost 'only' 10 times more that I am willing to pay for it (Sapphire FP3 DBs cost $2.6k)

 

[anubis44]

When will they stop this money bleeding business. they should focus on one or two products and finish them then market them. A medium company like AMD and the companies in the same size can't deal with many projects at the same time. Nvidia for example built all their business around single product and make change to drivers (GeForce, Quadro and Tesla) or adding others products to their own (Tegra and Grid). AMD is almost the same size as Nvidia and need to do the same. Nvidia is now dealing with single design which is Maxwell while AMD is dealing with CGN1, CGN1,1 and CGN 1,2 in their production line, they need to stop this fragmentation, and focus before it is too late.

This isn't a 'money bleeding business', AMD is monetizing existing technology. They've already developed this stuff, and now they're shrinking it down for a different market. What would be a waste of money would be if they DIDN'T sell this existing tech to as many different markets and customers as possible.