Skip navigation

AMD Corporate

4 Posts authored by: mark.papermaster Employee

Exactly three years ago we announced a six-year goal to dramatically improve the energy efficiency of our processors. AMD’s 25x20 initiative is our vision to improve the energy efficiency of our accelerated processors (APUs) for mobile technology devices by 25x from a 2014 baseline. We’ve done tremendous work to date. We’ve developed new processor architectures, power efficient technologies, and power management techniques all toward the goal of accelerating energy efficiency of not only our APUs, but also of our compute processors (CPUs) and graphics processors (GPUs). Let’s look at the results so far.

 

 

25x20 Chart.png

 

 

At the three-year halfway point, we’ve launched two new products (6th and 7th Generation AMD  A-Series APUs) for mobile products, with the next generation Ryzen™ Mobile to follow in the second half of 2017. Since 2014, we have achieved a nearly 4x improvement in energy efficiency, placing AMD ahead of pace to achieve our 25x20 goal1.

 

Achieving these substantial improvements in energy efficiency over several years, and associated reductions in the carbon footprint of our products, required broad and deep innovation. We’ve designed and implemented many new capabilities into our processors over this time. For example, Heterogeneous Unified Memory Access (hUMA) allows the CPU and GPU to directly access the same memory as peers instead of requiring all access to go through the CPU. This hUMA has led to as much as 17 times boost in performance for certain features when using the GPU’s parallel processing in concert with the CPU2. But, because it’s a shared power/thermal infrastructure, the power demands are equivalent to using the CPU alone.

 

Another significant AMD innovation is adaptive voltage and frequency scaling. This scaling involves the implementation of unique, patented silicon speed capability sensors and voltage sensors in addition to traditional temperature and power sensors. Silicon speed capability and platform voltage control can vary significantly part-to-part and platform-to-platform. The speed and voltage sensors enable each individual processor to adapt to its silicon characteristics, platform behavior, and operating environment. By adapting in real-time to these parameters, the processor can dynamically optimize its operation for maximum efficiency.

 

These innovations and the many other technologies we’ve created and implemented over the last three years are substantive and foundational engineering work. Our efforts have been widely recognized too, having received multiple awards worldwide from prestigious organizations.

 

25x20 Awards 2016 to 2017_for Mark P blog.jpg

 

When we think about these recognitions in terms of customer experience, we’re honored by these awards from the technology and energy efficiency communities. The excellent engineering, progress towards the goal, and public accolades show that we are on the right path. AMD remains committed to the 25x20 goal, and will continue to leverage our portfolio of intellectual property around architectural innovation, power-efficient technology, and power management techniques.

 

However, the best is yet to come. The 25x20 results so far have come from our 28nm products. We anticipate a strong improvement in performance per watt by moving to a 14nm FinFET process. We will launch our Ryzen Mobile APU built on this process in the second half of 2017. What’s more, we’re not resting on the design front. We are adding more power management with a suite of power reduction techniques called Pure Power. In Ryzen Mobile, we are estimating an increase in CPU performance of up to 50 percent, and graphics over 40 percent, with a cut in power use of up to 50 percent.

 

At the halfway point, we are on track to achieve 25x20. I’m very happy with what we’ve done to advance energy efficiency and achieve this position, and very excited about the innovation still to come. Creating differentiated low-power products remains a key element of our business strategy. We will continue to report on our progress toward achieving the 25x20 energy efficiency goal, as we strive to bring customers the benefit of industry leadership.

 

 

Mark Papermaster is CTO and SVP for Technology & Engineering at AMD. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied. 

 


[1] Typical-use Energy Efficiency is defined by taking the ratio of compute capability as measured by common performance measures such as Cinebench and 3DMark 11, divided by typical energy use as defined by ETEC (Typical Energy Consumption for notebook computers) as specified in Energy Star Program Requirements Rev 6.1 10/2014. “Kaveri” represents the baseline of relative compute capability (1.0) and relative energy efficiency (1.0). 6th Generation APU "Carrizo” relative compute capability (1.23) divided by relative energy efficiency (0.35) equals 3.5X improvement over “Kaveri”. 7th Generation APU “Bristol Ridge” relative compute capability (1.36) divided by relative energy efficiency (0.34) equals 3.95X improvement over “Kaveri”. Testing conducted by AMD Performance labs on optimized AMD reference systems. PC manufacturers may vary configuration yielding different results. An equal blend of Cinebench R15 CPU performance score and 3DMark11 GPU performance score is used to represent APU performance. In order of Cinebench / 3DMark11 scores, Kaveri scored 232/2142, Carrizo scored 277/2709, and Bristol Ridge scored 279/3234.

 

[2] Footnote 17 in http://www.amd.com/Documents/energy-efficiency-whitepaper.pdf

Mark Papermaster, CTO, and SVP for Technology & Engineering

 

This week, two new consortia, with AMD support, announced open standard development efforts targeting high-performance interconnect technologies inside future servers and datacenters. A third consortium, CCIX, which was announced in May and counts AMD as a founding member, released its first specification and welcomed new members.

 

AMD participation in these three standards bodies should come as no surprise, as we have a long history of supporting open standards in technology development. Our leadership in helping form the HSA Foundation to make programming heterogeneous computing devices easier is one example of this commitment. Additionally, the recent establishment of GPUOpen provides developers with free, open development tools and software for PC games, computer-generated images and GPU compute applications.

 

The latest announcements reinforce the shared goal of consortia members to bring open standards into the datacenter. While there are unique aspects to each at the technical level, there are also many similarities, and many of the participating companies, like AMD, are members of all three.At the engineering level, the standards can generally be distinguished this way:

 

Gen-Z is focused on an optimized solution for rack-scale connections to memory and accelerators as an open alternative to the proprietary Omni-Path.

 

OpenCAPI targets connections to IBM’s POWER architecture with memory and accelerators. AMD is focused on getting our Radeon Technologies Group products into this ecosystem by joining this effort.

 

CCIX is primarily a coherent processor-to-accelerator interconnect that leverages the existing PCIe® physical interconnect ecosystem.

 

Members of all three organizations are committed to an open approach to speed innovation and collaboration versus existing proprietary and closed standards. Future innovation will be critical to taking computing into the next decade and beyond as big data analytics, machine learning and other applications with extremely high computational needs become more prevalent.

 

I encourage you learn more about each organization from their respective announcements. Through these efforts, we are setting a solid foundation for customers and partners to support our “Zen”-based server platforms and AMD Radeon™ Pro Graphics for compute, as well as their follow-ons, long into the future!

 

Mark Papermaster is CTO, and SVP for Technology & Engineering at AMD. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied. 

It’s easy to look around and see the amazing societal benefits from the explosion of computing over the last twenty years. For example, advancements in medical computing have decreased the time taken for DNA sequencing, allowing doctors to more quickly identify which drug would work best against certain types of cancers. Along with this explosion, the energy consumption and environmental footprint from computing has increased. This is a challenge that affects all of us, and the brains of the computer – the microprocessor – is an important part of the total equation.

 

In the past, processor designers could rely heavily on the mechanics of Moore’s Law to provide regular leaps in computing power and energy efficiency. But the steady improvements from better manufacturing processes and size shrinks predicted by Moore’s Law have slowed in recent years, largely because transistors are now so small they are running into fundamental limits of physics that drive costly processing. Moving forward, much of the new gains will likely come from innovative chip designs.

 

At AMD, energy efficiency has long been a key product design focus for our microprocessors, including our APUs, CPUs and GPUs. In June 2014 we announced our goal for a 25 times improvement in energy efficiency of our mobile APUs by 2020. That was, and is, a stretch goal considering that from 2009 to 2014 we achieved a 10 times improvement. We thought 10x was pretty good, but in 2014 we set our sights even higher.

 

At the two-year mark, I’m pleased to report we are on track toward achieving our 25x20 goal. And in the process our latest chips have dramatically improved in computing performance as well as energy efficiency¹.

 

These gains are no small achievement. In fact, our progress on energy efficiency was honored this week by Environmental Leader with a Project of The Year award. The judges viewed our 25x20 vision as “ambitious and audacious” and noted that the project has already demonstrated strong progress. Among the results, so far, is a 50 percent decrease in the carbon footprint of systems built around our 6th generation A-Series processors.

 

AMD continues to achieve these leaps in energy efficiency by focusing on numerous design enhancements, improved transistor density, and system level efficiency optimizations that result in power and performance improvements. These are detailed in our white paper, describing AMD’s commitment to energy efficiency. For example, our 6th generation A-Series processor introduced in 2015 was 2.4 times more energy efficient than its 2014 predecessor. And the recent introduction of our new 7th generation A-Series processor in 2016 – code named “Bristol Ridge” – adds an additional 14 percent improvement. Taken together, at the two-year mark, we are solidly above the trend line needed to achieve our goal of 25 times energy efficiency improvement by 2020².

 

In addition to our mobile APUs, we’re also making important strides in the energy efficiency of our GPU technology. With similar engineering advancements in architecture and power management, plus a boost from next generation process technology, AMD’s upcoming “Polaris” GPUs are on target to at least double the performance per watt over previous generations³.

 

We have more ground to cover to reach our 25x20 goal but we remain confident we are taking the right steps to meet this ambitious target. Look for additional reports on our progress and the energy efficient innovations coming up that will help get us to the goal line.

 

Mark Papermaster is CTO and SVP for Technology & Engineering at AMD. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.  

 

 

 


1) Typical-use Energy Efficiency as defined by taking the ratio of compute capability as measured by common performance measures such as SpecIntRate, PassMark and PCMark, divided by typical energy use as defined by ETEC (Typical Energy Consumption for notebook computers) as specified in Energy Star Program Requirements Rev 6.0 10/2013. “Kaveri” relative compute capability (4.5) of baseline divided by relative energy efficiency (0.45) of baseline = 10X. "Carrizo” relative compute capability (6.0) of baseline divided by relative energy efficiency (0.22) of baseline = 27.4X (which is 2.7x that of “Kaveri”). “Bristol Ridge” relative compute capability (6.9) of baseline divided by relative energy efficiency (0.18) of baseline = 38.5X (which is 3.8x that of “Kaveri”).

2) Typical-use Energy Efficiency as defined by taking the ration of compute capability as measured by common performance measures such as SpecIntRate, PassMark and PCMark, divided by typical energy use as defined by ETEC (Typical Energy Consumption for notebook computers) as specified in Energy Star Program Requirements Rev 6.0 10/2013. “Kaveri” relative compute capability (4.5) of baseline divided by relative energy efficiency (0.45) of baseline = 10X. “Carrizo” relative compute capability (6.0) of baseline divided by relative energy efficiency (0.22) of baseline = 27.4X (which is 2.5x that of “Kaveri”), “Bristol Ridge” relative compute capability (6.9) of baseline divided by relative energy efficiency (0.18) of baseline = 38.3X (Which is 3.4x that of “Kaveri”).

3) Testing conducted by AMD internal labs as of Dec 15, 2015 with AMD’s previous “Hawaii” and “Bonaire” architecture based platforms and preliminary “Polaris” architecture based engineering sample. Systems tested with Intel i7-4770K with 8GB DDR3-1600 RAM, Driver 15.30 beta, Windows 10 64bit running a “Perlin Noise” benchmark measured in fps.  AMD’s "Hawaii" based platform averaged 377 fps at 1000 MHz while consuming 195.4 W, resulting in 1.9 frames per watt.  AMD’s "Bonaire" platform averaged 131.4 fps at 1015 MHz while consuming 71.6 W, resulting in 1.8 frames per watt. Preliminary engineering data showed AMD’s Polaris architecture based engineering sample as resulting in more than 2x the performance per watt as compared to “Hawaii” and “Bonaire” based platforms in this testing. POL-2

Sustainable is Attainable (3).jpgDeveloping energy efficient processors has long been a design focus at AMD. In 2014 we started talking about our goal to improve the energy efficiency of our mobile processors – those found in laptops, for example – by 25 times by 2020. That’s a lot. Using a car analogy, this rate of improvement would be like turning a 100-horsepower car that gets 30 miles per gallon into a 500 horsepower car that gets 150 miles per gallon in only six years.

We’re well on our way towards achieving this goal with the launch this summer of the6th Generation AMD A-Series APU, also known as “Carrizo.” This latest processor achieves a 2.4x improvement in energy efficiency over our prior generation, placing AMD ahead of the pace needed to realize our 2020 goal. With the product designs currently in development, I’m very confident we will achieve this goal.

 

Developing energy efficient processors for our customers and delivering longer battery life for end-users also aligns perfectly with our efforts to reduce the carbon footprint of our company and our products. It’s also the right thing to do. Climate change is arguably one of the most significant issues facing humankind today. One of the corollaries to improving energy efficiency of our products should be a reduction in their “carbon footprint.” To determine this footprint, the amount of “carbon emissions” from each phase of the product lifecycle, from manufacturing to transportation to consumer use has to be measured. These carbon emissions, also known as “Greenhouse Gases” (GHGs), when emitted to the atmosphere accumulate and trap the sun’s energy, leading to climate change.

 

We recently conducted a carbon footprint study to determine whether the improvement in energy efficiency with the 6th Generation AMD A-Series APU also resulted in a reduction in greenhouse gases. This involved an apples-to-apples comparison of this processor compared to our prior generation product using a standardized reference platform. The study followed the guidelines set forth by the widely accepted Greenhouse Gas Protocol established by the World Resources Institute and World Business Council for Sustainable Development. The study then underwent third party critical review to check for adherence to those standards.

 

The results are now available. The findings show that the carbon footprint of the 6th Generation AMD A-Series APU is approximately 46 percent less than the prior generation processor. This assumes a three year service life and typical usage scenario. The use phase of the processor is the largest contributor to the overall carbon footprint. These results indicate that an end user upgrading to a notebook computer using the latest processor can expect a 50 percent reduction of GHG emissions over the service life of that product compared to one using the prior generation.

 

This sounds good, and it is. And this is another data point to validate that our design efforts are on track. But just what does this mean in terms of the impact? For each individual processor this is fairly small. With larger numbers it becomes more interesting. Consider the use case of a large enterprise. If this firm upgraded 100,000 machines from the prior generation to ones using the newest processors – with other components remaining the same – they could potentially achieve reductions of 4.9million kWh and 3,350 metric tons of GHGs over a three year service life. The power saved would be enough to power 461 US homes for a year.

 

That’s significant. If similar improvements in energy efficiency played out across the larger universe of Information and Communications Technologies (ICT) that includes computing and communication devices encompassing not only computers but also televisions, cell phones, network hardware, satellite systems and more, the impact would be dramatic.

Complete results and the details of our Carbon Footprint study, entitled Comparative Carbon Footprint Assessment of the Manufacturing and Use Phases of Two Generations of AMD Accelerated Processing Units, can be viewed here.

 

Mark Papermaster is Senior Vice President and Chief Technology Officer at AMD. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.