An Accelerated Processing Unit (APU) is generally any processing system that includes

additional processing capability designed to accelerate one or more types of computations

outside of a CPU. This may include a Graphics Processing Unit, a Field Programmable Gate
Array, or similar specialized processing system. Variations on the usage of this term include a
marketing based variation in which APU describes a processing device which integrates a
CPU and a GPU on the same die thus improving data transfer rates between these
components while reducing power consumption. APUs can also include video processing and
other application-specific accelerators. Examples: AMD Fusion, Intel's Sandy Bridge and
NVIDIA's Project Denver.

The term Accelerated Processing Unit or APU was first used in a public context with
respect to accelerated computing on the Scalability.org blog, and prior to that in various
presentations and business plans written by Joe Landman of Scalable Informatics. Other uses
include Xilinx using the term for an Auxiliary Processor Unit.

Accelerated processing unit captures industry flair

AMD has introduced the first accelerated processing unit, or APU, based on Fusion
technology. The Embedded G-Series processor architecture has been adopted by several
companies to raise the bar on graphics and performance for embedded systems. At this year's
Embedded World in Nuremberg, Germany (1 to 3 March) several companies will be showing
how G-Series technology can be used.
The AMD Embedded G-Series (Figure 1) is based on the company's Fusion technology. The
single piece of silicon incorporates the low-power, x86 CPU based on the Bobcat core, with
DirectX 11-capable GPU and parallel processing engine. It is described as a new class of
accelerated processor and the company claims that it combines more compute capabilities on
a single die than any processor in the history of computing.
The company identified that the embedded market is calling for increased power efficiency,
reduced footprint, high CPU performance, full feature sets with an emphasis on graphics
performance to produce the APU.
Differentiation can be the critical factor in a product's success and longevity in the market.
Chris Cloran, corporate VP and GM, Computing Solutions Group, AMD said "By combining
the processing of the CPU with the GPU on a single energy-efficient chip, customers can take
advantage of better price/performance, superior 1080p HD playback, and small form factors
for innovative designs."

The power of the architecture will enable graphics-intensive design, such as in digital
signage, internet-ready set top boxes, mobile and desktop thin clients, casino gaming
machines, point-of-sale kiosks, and small form factor PCs, as well as SBCs.

The open development ecosystem for the Embedded G-Series platform includes multiple
BIOS options, support for various Microsoft Windows, Linux, and real-time operating
systems, the integrated OpenCL programming environment, and source-level debug tools.

Specifications

The series is made up of one or two cores with 1Mbyte L2 cache, 64bit floating point unit.
The processors operate at up to 1.66GHz and have a choice of SIMD engines, such as the
DirectX 11-capable graphics as well as 3D and graphics processing. There is also a third
generation unified video decoder. For power management there is C6 and power gating.
There is a DDR3 800-1066 memory with support for 64bit channel and two DIMMs. All of
this is packaged in a BGA with a physical footprint of 890mm², including the Fusion I/O
controller hub. AMD is exhibiting at Embedded World, Hall 9: 557.

One manufacturer to use the technology is Congatec, in its COM Express line. Visitors to
Embedded World will be able to see the conga-BAF (Figure 2) in Nuremberg (Hall 12: 122).

The COM Express standard has been expanded with the Fusion technology to combine
processors and graphics cores in a compact package. It boasts a strong performance/W ratio
and flexible task allocation on the CPU and GPU. The module can stream high-definition
multimedia content and play the next-generation 3D games.

The company integrates five processors from the series, from the T44R 1.2GHz single-core
(64kbyte L1 cache, 512kbyte L2 cache) with 9W TDP to the T56N 1.6 GHz dual core
(64kbyte L1 cache, 512kbyte x2 L2 cache) with 18W TDP. The module uses the Hudson E1
Fusion controller hub and provides a two-chip solution with up to 8Gbyte dual-channel
DDR3 memory.
The integrated graphics core with the Universal Video Decoder 3.0 for seamless processing
of Blu-ray content via HDCP (1080p), MPEG-2 HD and DivX (MPEG-4) video interfaces
supports DirectX 11 and OpenGL 4.0 for fast 2D and 3D image display as well as OpenCL
1.1. The choice of graphics interfaces includes VGA, single channel LVDS as well as
DisplayPort and DVI/HDMI for direct control of two independent displays.
Six PCI Express x1 lanes, eight USB 2.0 ports, four SATA, one EIDE and one Gigabit
Ethernet interface allow flexible system expansion with high data bandwidth. The board
controller features, ACPI 3.0 power management and high-definition audio complete the
package.

Computing design

CompuLab has selected the Embedded G-Series as the basis for its next-generation fit-PC3
with a design in a durable, fanless casing for what it says is its most powerful, miniature,
fanless IPC to date.

The IPC measures 6.3x6.3x0.98inch and exploits the APU's low power x86 CPU and
discrete-class GPU to provide performance, visual experience and features that are claimed to
be comparable to much larger systems while consuming a fraction of the power.

Irad Stavi, director of business development, CompuLab said. "We expect that the dual-core
CPU, visual performance, low power and the x86 eco-system provided by the platform
together with the small size, durability and versatility of the fit-PC3 design will prove an
attractive combination for many industry segments."

The IPC can host an extension card that can be tailored to fit custom I/O requirements for the
set of interfaces provided by the Embedded G-Series platform.

Three versions, one company

Kontron has introduced three small form factor embedded platforms based on Embedded G-
Series. It has developed COMs, embedded motherboards and SBCs for multimedia delivery
applications based on the APUs' energy efficient processor cores and advanced GPU with
support for DirectX 11.

The COM Express will debut at Embedded World (Hall 12: 404), followed by a compact
embedded motherboard and an ultra-small SBC in Q2 2011. Additionally, with the support
for OpenCL and Microsoft DirectCompute, parallel processing executed by the graphics core
will speed up vector processing applications such as situational awareness and video
surveillance required by the industrial automation, military and medical markets where radar,
sonar and image recognition will benefit from where the advantages of the built-in GPU for
cost savings compared to using a number of multi-core processors.

Figure 1: Embedded G-Series APUs raise the bar on graphics, power and processing
performance.
Figure 2: Congatec capitalises on the graphics performance in the conga-BAF.
Figure 3: The fit-PC3 squeezes visual performance into a slim modem.

AMD has started promoting its new Fusion concept, which, among other things, encompasses central and graphics
processor on the same piece of silicon. This approach is significantly different from the existing integration of the
competitive company, and we came in contact with Samuel Naffziger, AMD Senior Fellow, who explained how they
thought this whole system and which are the important innovations comparing to the current approach.

Four years after buying graphics chipmaker ATI, AMD shows the goods at
Computex: called Accelerated Processing Unit (APU), these new chips which merge
CPU and GPU are supposed to save power and cost (according to AMD) although it
doesn’t seem to us like they would generate any groundbreaking savings on both
counts. A few months ago, Intel has shown that it too had embedded a graphics chip
to a CPU.

That said, AMD’s Fusion is superior to Intel’s CPU/GPU hybrid graphics, and
APUs are able to also perform general purpose computing to help the main
processor, if apps support it – not all CPU/GPU hybrid are created equal. It would be
a *real* game changer if developers were starting to rely on having another set of
processing units inside every AMD processors (I suspect that AMD will pay select
devs to do it…). For that, the APUs need to support DirectCompute and/or OpenCL,
two standards that Intel’s graphics solution doesn’t support now.

Before you get too excited, don’t forget that APUs are “just” integrated graphics,
even if AMD tries really hard to make them cool. They embed much fewer sub-
processors than their “add-on” counterparts, and therefore the raw power is much
less. We’ll have to wait for the details and independent numbers before deciding if
this is going to be useful in the short-term. (photo courtesy of Trusted Reviews)

Earlier this week AMD had some big news and the term “APU” was tossed around a lot. This
surely had a lot of people wondering what the heck an APU is and how it differs from a
central processing unit (CPU).

The video above shows how an APU works by explaining the flow of data through the
components involved. It also compares how a CPU works to how an APU does things. It’s
helpful enough, but if you already know some of the basic terms (CPU, GPU, etc.) then
Wikipedia’s definition might be better:

An APU integrates a CPU and a GPU on the same die thus improving data transfer rates
between these components while reducing power consumption. APUs can also include video
processing and other application-specific accelerators. Examples: Intel’s Sandy
Bridge, AMD Fusion and NVIDIA’s Project “Denver”

Make more sense now? An APU is the combination of a CPU (generally a multi-core one), a
graphics processing unit, and then some way to get them to play together nicely.

Why do this in the first place? Because it turns out that GPUs are good at things besides
graphics, so the CPU can offload computing tasks to them. To optimize this cooperation
bottlenecks between the CPU and the GPU had to be removed, the result being the APU!

AMD outlines plans for future processors

by Cyril Kowaliski — 2:45 PM on December 14, 2006
During its four-hour Financial Analyst Day presentation today, AMD revealed new elements of its
processor roadmap spanning the next couple of years, as well as its plans to scale beyond the
current multi-core model. Intel talked about processors with "tens to hundreds of cores" at IDF
earlier this year, but AMD believes the core race is just a repeat of the megahertz race and that
adding more cores isn't the best way to go about scaling processor performance in the future.
Instead, AMD is cooking up what it calls "Accelerated Processing Units":
 Source: AMD.

Accelerated Processing Units, or APUs, will be multi-core chips that include any mix of processor
cores and other dedicated processors. Fusion, AMD's integrated CPU and graphics processor, is
AMD's first step in that direction. However, the company eventually intends add more specialized
cores that can handle tasks other than general-purpose computing and graphics. AMD didn't give
any specific examples, but one could easily imagine future Fusion-like chips with cores for physics
processing, audio/video encoding, and heck, maybe even AI acceleration.

Whatever it ends up doing, AMD says it will be able to tailor APUs with different combinations of
core types to specific markets. Fusion, for example, will be aimed chiefly at the mobile market
when it arrives in 2009. Enthusiasts needn't worry, though: AMD says it doesn't plan to integrate
high-end GPUs and CPUs into massive silicon fireballs, because both production costs and power
envelopes for such chips would be too high.

AMD also revealed some of its more immediate plans for the processor market by showing off new
desktop and mobile roadmaps:
Source: AMD. The desktop roadmap is more or less self-explanatory, placing the introduction of
quad-core desktop CPUs in mid-2007 and the launch of derived dual-core models in the latter part
of the year. All new chips will have support for HyperTransport 3.0 and DDR2 memory, and
accompanying motherboards will boast PCI Express 2.0 support. AMD apparently doesn't intend to
switch sockets or move to DDR3 memory until the middle of 2008.
On the mobile front, things are a little more interesting. Early next year, AMD will launch a new
mobile chip code-named "Hawk" that will boast lower power utilization than current Turion 64 X2
and Mobile Sempron chips. The chip will be paired with platforms that will have support for hybrid
hard drives as well as a somewhat novel concept dubbed hybrid graphics.

According to AMD, notebooks with hybrid graphics will include both discrete and integrated
graphics processors. When such notebooks are unplugged, their integrated graphics will kick in and
disable the discrete GPU. As soon as the notebook is plugged back into a power source, the discrete
GPU will be switched on again, apparently without the need to reboot. AMD says this technology
will enable notebooks to provide the "best of both worlds" in terms of performance and battery life.

In late 2007, AMD will introduce a mobile chip dubbed Griffin that will support split power planes
and HyperTransport 3.0, meaning it'll almost certainly be based on AMD's quad-core architecture.
Griffin will be accompanied by platforms supporting PCI Express 2.0, DirectX 10-class integrated
graphics, DisplayPort video output, and a Universal Video Decoder—essentially a dedicated video
processor.

That about covers what AMD unveiled about its upcoming processors today. The presentation did
include a fair amount of discussion about the company's financial performance and consumer
electronics plans, though; we'll fill you in on those topics a little later.

To say that we've been waiting for AMD's Fusion CPU / GPU combo for a long time would be an
understatement. In fact, while AMD was busy talking about it, Intel swept in with its own Arrandale
and Clarkdale chips that pack graphical and computing processing into the same chip. Lest we were
discouraged, then, AMD is making a return to form with news that its first Fusion APU (Accelerated
Processing Unit) is about to start sampling to manufacturers, with a now definite 2011 launch
window. Codenamed Llano, this will be a quad-core beastie with intended operating speeds of more
than 3Ghz and graphics parts borrowed from ATI's successful line of Evergreen GPUs. That means
DirectX 11, a feature Intel is unlikely to match, whereas AMD will have everything Intel currently
does and more, with a 32nm production process, on-die integration (rather than just the same chip
packaging), and power gating allowing for dynamic per-core overclocking a la Turbo Boost. It's been
lonely without you AMD, now just fulfill this promise and all that absenteeism will be forgiven.
AMD announced the launch of its new generation of processors, the Fusion platform. The
new family of chipsets, using the accelerated processing unit (APU), are designed to handle
multi-core processors as well as high-end graphics using a single die.

Leading computer manufacturers such as Sony, Dell, HP, Asus, Lenovo, MSI and Toshiba
would be launching APU-based systems soon.

AMD (Advanced Micro Devices) claims that the new design offers ordinary people the
computing capability that was earlier associated with supercomputers. The new processor’s
power-saving feature would be a boon for users who were increasingly seeking richer graphic
content on the Internet.

AMD’s two centres in Hyderabad and Bangalore employ about 1,100 persons. The team had
played a critical role in the development of the processor. One of the key features of the new
chipset is its “AllDay Power” capability, which enables usage of computers for more than 10
hours on a single charge. The Hyderabad team played a major role in achieving this.

Ravi Swaminathan, Managing Director and Regional Vice-President, Sales and Marketing,
AMD India, said the new chipset design promise to herald a “fundamental transformation, in
the computer landscape.” “The new family of processors represent the biggest shift in PC
technology in 40 years,” Mr. Swaminathan said.

Claiming that AMD was the only chip-maker to have established capabilities in designing
and manufacturing central processing units as well as graphics processing units, Mr.
Swaminathan said, “The APU will set the new industry standard.” He said the capabilities of
the new range of processors make them ideal for use in a range of products — from mobile
phones to desktop computers. While the E-series and the C-series of APUs would be
available in India soon, the higher-end A-series APU would be available by May, Mr.
Swaminathan said. [via]

Fusion platform, the new generation processors from Computer processor manufacturer AMD was
launched and announced on Tuesday. (01-02-2011).

The Fusion platform are specially designed to handle multi-core processors which uses accelerated
processing unit (APU) which is considered to be the new family of chipsets by AMD. The Fusion
platform can also be used as high-end graphics using a single die. Mr. Ravi Swaminathan, Managing
Director and Regional Vice-President, Sales and Marketing, AMD India, said “The new family of
processors represent the biggest shift in PC technology in 40 years,”
Mr. Ravi Swaminathan also said that the Fusion platform will set a new industry standards in the
market and he also claimed that AMD is the only company to manufacturer chipsets which capabilities
in designing and manufacturing central processing units and graphics processing units. He also
added that the new family chipsets will be available in different range of products which will include
mobile phones to desktop computers.

The E-series accelerated processing unit (APU) and the C-series accelerated processing unit (APU)
will be available in India very soon and the high end version A-series APU would be available by May,
Mr. Swaminathan said. He also added that the leading laptop and computer manufacturers like Sony,
Dell, HP, Asus, Lenovo, MSI and Toshiba will also launch APU-based systems very soon.

The AMD's Indian technology development team, based in Hyderabad and Bangalore, had played “a
critical role” in the development of the processor added Michael Goddard, Corporate Vice-President,
Chief Engineer, Low-Power Clients, Products Group. The highlighting feature of new family chipset is
the “AllDay Power”.