Advanced Micro Devices this week launched the AMD embedded R-series accelerated processing unit (APU) platform designed for mid- to high-end graphics-intensive applications such as digital signage, casino gaming, point-of-sale systems and kiosks, as well as parallel-processing-intensive applications spanning medical imaging and security/surveillance.
AMD did not unveil exact specification of its R-series APUs for embedded application, but indicated that they have up to five Piledriver-class x86 general-purpose cores, memory controller as well as Radeon HD 7000-series class graphics adapter. Thanks to such features as integrated high-definition video encoder, high-performance graphics engine and other, the new chips have a lot of chances to become rather popular in case they support appropriate operating systems. The new chips have thermal design power in the range between 17W and 35W, hence, their specs are likely to be similar to mobile A-series "Trinity" APUs released last week.
“With the AMD Embedded R-Series, we are taking our APU technology to the next level. By leveraging its seamlessly integrated heterogeneous system architecture, developers can tap into a high-performance and efficient parallel processing engine to accelerate their graphics- and compute-intensive applications, all while using industry-standard libraries such as OpenCL and DirectCompute," said Buddy Broeker, director, AMD embedded solutions.
Developers working with the AMD embedded R-series APU can implement remote management, client virtualization and security capabilities to help reduce deployment costs and increase security and reliability of their AMD R-series based platform through AMD DAS 1.0 featuring DASH 1.1, AMD virtualization and trusted platform module (TPM) 1.2 support.
A variety of AMD customers have already announced mini-ITX mainboards, computer-on-module devices, system solutions for high-performance digital signage applications, system solutions for high-performance casino gaming applications and other products based on R-series APUs, the company said.
Tags: AMD, APU, Trinity, Piledriver, Radeon, Fusion
Comments currently: 3
Discussion started: 05/23/12 05:25:14 AM
Latest comment: 05/27/12 06:14:32 AM
Expand all threads
| Collapse all threads
This model is particularly interesting R-452L 4 cores 1.6/2.4 GHz 19W dtp!
Really a ultra part! Only problem is that a stupid Amds chipset drives 6.5-7.8W.
If morons integrate chipset real 7000 gpu and move to 28 nm lithography they cud become concurrent! It's insanity that a chipset draws more power on a ultra part then a piledriver module!
PS : They also must improve multitasking between modules and a best way is parallel processing model! We really don't need more tasks or if we need it its better that those are virtual ones like Intel implemented.
If they really try hard in of they cud menage to have production part's before the end of the year even better they cud integrate 8 gen. gpu. So by the beginning of the year we cud have 4x module piledriver part with even better gpu with dtp of 20-25W.
05/23/12 05:25:14 AM]
- collapse thread
AMD's mobile Chipset do NOT use 6.5-7.8W That's their Desktop chips, their Mobile chips has a TDP of 4.7W, while being produced on 65nm, compared to Intels Mobile chipset with the same features, produced on 22nm has a TDP of 4.1W, seems process doesn't have much effect on a chipsets TDP.
What you talk about is nearly impossible considering how new the GCN architecture is, and it's the reason we'll only see it next time around with Kaveri.
The way Bulldozer Modules work, well there's nothing wrong with, a Quadcore is perfectly good at multitasking the way it is, and AMD uses their own method that is superior in scaling compared to Intels method. (See CMT vs SMT).
I suspect it will only be a few months before we start seeing demoes of Kaveri, like with Trinity it didn't take long before we actually saw AMD Demoeing it. The only thing I wish AMD would try and do better would be to deliver Desktop users with some higher end parts, but hopefully we'll see something nice with Vishera just around the corner. By the end of the year we'll hear talks about HD8000 and GTX700 chips.
05/24/12 10:16:58 AM]
I actually think that there will be the rev 2 of Trinity and it will be based on rev 2 of piledrivers about 12-15% faster then rev 1. This shud be based on 28 nm technology and a logical step should be that they integrated real 7000 gpu already based and printed on 28 nm lithography. This recent gpu is 2x per wat efficient. Intel can't follow this in next ½ year period. In the other hand Amd is working on the chipset integration in next gen APUs.Chipset DTPs i mentioned are for R series.
Mi estimations based on 4 module APUs dtp are based on those 3 observations ( gpu + chipset integration + 28 nm production ). Those are a minimal dtp consumptions that they cud menage for UVL products.
When i mentioned multitasking i whose more refering to ability of an piledriver module working as single I/O trade with possible second spare trade if need be.
05/27/12 06:14:32 AM]
Add your Comment
Enter your username and e-mail address. Password will be sent to you.