AMD Strategy Transformation Brings Agile Delivery of Industry-Leading IP to the Market [AMD press release, Feb 2, 2012]
At its annual Financial Analyst Day, AMD (NYSE: AMD) detailed a new “ambidextrous” strategythat builds on the company’s long history of x86 and graphics innovation while embracing other technologies and intellectual property to deliver differentiated products.
AMD is adopting an SoC-centric roadmap designed to speed time-to-market, drive sustained execution, and enable the development of more tailored customer solutions. SoC design methodology is advantageous because it is a modular approach to processor design, leveraging best practice tools and microprocessor design flows with the ability to easily re-use IP and design blocksacross a range of products.
“AMD’s strategy capitalizes on the convergence of technologies and devices that will define the next era of the industry,” said Rory Read, president and CEO, AMD. “The trends around consumerization, the Cloud and convergence will only grow stronger in the coming years. AMD has a unique opportunity to take advantage of this key industry inflection point. We remain focused on continuing the work we began last year to re-position AMD. Our new strategy will help AMD embrace the shifts occurring in the industry, marrying market needs with innovative technologies and become a consistent growth engine.”
Roadmap Updates Focus on Customer Needs
Additionally, AMD today announced updates to its product roadmaps for AMD Central Processing Unit (CPU) and Accelerated Processing Unit (APU) products it plans to introduce in 2012 and 2013. The roadmap modifications address key customer priorities across form factors including ultrathin notebooks, tablets, all-in-ones, desktops and servers with a clear focus on low power, emerging markets and the Cloud.
AMD’s updated product roadmap features second generationmainstream (“Trinity”) and low-power (“Brazos 2.0”) APUs for notebooks and desktops; “Hondo,” an APU specifically designed for tablets; new CPU cores in 2012 and 2013 with “Piledriver” and its successor “Steamroller,” as well as “Jaguar,” which is the successor to AMD’s popular “Bobcat” core. In 2012, AMD plans to introduce four new AMD Opteron™ processors. For a more in-depth look at AMD’s updated product roadmap, please visit http://blogs.amd.com.
Next-generation Architecture Standardizes and Facilitates Software Development
AMD also provided further details on its Heterogeneous System Architecture (HSA), which enables software developers to easily program APUs by combining scalar processing on the CPU with parallel processing on the Graphics Processing Unit (GPU), all while providing high bandwidth access to memory at low power. AMD is proactively working to make HSA an open industry standard for the developer community. The company plans to hold its 2nd annual AMD Fusion Developer Summitin June, 2012.
New Company Structure Strengthens Execution
In conjunction with announcing its restructuring plan in November 2011, AMD has strengthened its leadership team with the additions of Mark Papermaster as senior vice president and chief technology officer, Rajan Naik as senior vice president and chief strategy officer, and Lisa Su as senior vice president and general manager, Global Business Units. These executives will help ensure that sustainable, dependable execution becomes a hallmark of AMD.
AMD started talking about ‘Trinity’ and ‘Hondo’ last summer. See in Acer repositioning for the post Wintel era starting with AMD Fusion APUs [June 17, 2011]
What AMD could definitely be proud of for 2011 is A “Brazos” Story: The Little Chip That Could (And Then Just Kept On Going) [AMD Fusion blog, Feb 1, 2012]:
In late 2010, AMD shipped its first-ever Accelerated Processing Units (APUs), internally codenamed “Brazos,”which combined the tremendous processing power of graphics and x86 on a single chip.
We had high expectations for the low-voltage “Brazos” APU: great computing, HD, long battery life and DirectX 11 capable graphics, all on a single chip. Yet still we were blown away by the initial industry reception. It was only a year ago we left CES with seven highly-sought after innovation and technology awardsfor the little product we ultimately named the C- and E-Series APUs, including:
- 2010 PC Magazine Technical Excellence Award
- CES 2011 Design & Engineering Innovations Award
- CHIP China 2010 Highlight Awards
- Computer World China Innovation Award
- Notebooks.com Best Innovation CES 2011
- Popular Mechanics Editors’ Choice Award
- Shopping Guide China Most Advanced Digital Product Award
After CES we should have re-nicknamed “Brazos” the “Little Chip That Could.” And all throughout 2011, “Brazos” kept on chugging. We added the “Best in Show” Award at Embedded Systems Conference and the “2011 Best Choice of Computex TAIPEI Award” to the list of accolades. In the second quarter we sold more than five million C- and E-Series APUs. What a tremendous start to a new way of processing for AMD and the industry.
But “Brazos” kept on impressing, showing up in a variety of form factors – notebooks, netbooks, small desktops and all-in-ones– from top global OEM partners.
So it was no surprise or mistake that we ended 2011 with more than 30 million APUs shipped. It all started with little “Brazos,” which has now earned its place in history as AMD’s fastest ramping platform ever.
John Taylor, Director of Worldwide Product Marketing at AMD
AMD Codename Decoder – November 9, 2010 [AMD Business blog]
An APU is an accelerated processing unit, a new generation of processors that combine either low-power or high-performance x86 CPU cores with the latest GPU technology (such as DirectX® 11) on a single die.
Planned for introduction: Q1 2011
Market: Multiple devices, including notebooks ultrathins, HD netbooks and small form factor desktops.
What is it? A sub-one watt capable x86 CPU core that first comes to market in the “Ontario” and “Zacate” Accelerated Processing Units (APU) for mainstream, ultrathin, value, and netbook form factors as well as small form factor desktop solutions. “Bobcat” is designed to be an extremely small, highly flexible, out-of-order execution x86 core that easily can be scaled up and combined with other IP in SoC configurations.
Planned for introduction: Q1 2011
Markets: Value Mainstream Notebooks, HD Netbooks and Small Form Factor Desktops
What is it? “Brazos” is AMD’s 2011 low-power platform, available with two APUs; “Zacate” – currently planned to be marketed as the E Series – is an 18-watt TDP APU for ultrathin, mainstream and value notebooks as well as desktops and all-in-ones. “Ontario” – currently planned to be marketed as the C Series – is a 9-watt
APU for netbooks and small form factor desktops and devices. Both “Brazos” platform APUs include a DirectX® 11-capable GPU.
Planned for introduction: Q1 2011
Market: Server and Client
What is it? A multi-threaded high-performance x86 CPU core contained in the “Zambezi” processor for client PCs and “Interlagos” and “Valencia” processors for servers. Included in the “Scorpius” enthusiast desktop PC platform and “Maranello,” “Adelaide,” and “San Marino” server platforms, “Bulldozer” is designed to be a completely new, high performance architecture that employs a new approach to multithreaded compute performance for achieving advanced efficiency and throughput. “Bulldozer” is designed to give AMD an exceptional CPU option for linking with GPUs in highly scalable, single-chip APU configurations. “Bulldozer” offers AMD another exceptional CPU option for combining with GPUs in highly scalable, single chip APU configurations, beginning in 2012 APU designs.
Planned for introduction: Client (1H 2011); Server (2H 2011)
Market: Notebooks and Desktops
What is it? Part of the “Sabine” platform, “Llano” is a 32nm APU including up to four x86 cores and a DirectX® 11-capable GPU, primarily intended for performance and mainstream notebooks and mainstream desktops. “Llano” is engineered to deliver impressive visual computing experiences, outstanding performance with low power and long battery life.
Planned for introduction: Mid-2011
Market: Primarily ultrathin notebooks and HD netbooks
What is it? A 9W APU featuring dual or single “Bobcat” x86 cores currently planned to be marketed as the C Series, and primarily intended to serve the low power and highly portable PC markets for netbooks and small form factor desktops and devices.
Planned for introduction: Q1 2011
What is it? “Zacate” is AMD’s 18W APU designed for the mainstream notebook and desktop market. Zacate will feature low-power “Bobcat” CPU cores and support DirectX 11 technology.
Planned for introduction: Q1 2011
More information about 2011 AMD APU past on this blog:
– Acer repositioning for the post Wintel era starting with AMD Fusion APUs [June 17, 2011]
– Supply chain battles for much improved levels of price/performance competitiveness [Aug 16, 2011]
– Acer & Asus: Compensating lower PC sales by tablet PC push [March 29 – Aug 2, 2011]
– CES 2011 presence with Microsoft moving to SoC & screen level slot management that is not understood by analysts/observers at all [Jan 7, 2011]
– Changing purchasing attitudes for consumer computing are leading to a new ICT paradigm [Jan 5, 2011]
AMD started talking about ‘Trinity’ last summer. See in Acer repositioning for the post Wintel era starting with AMD Fusion APUs [June 17, 2011]
Advanced Micro Devices’ CEO Discusses Q4 2011 Results – Earnings Call Transcript [Seeking Alpha, Jan 24, 2012]
We are seeing particularly strong customer interest in our expanded low-power APUs for 2012. The low-power versions of our next-generation chip, Trinity APU, delivers mainstream performance while using half the power of our traditional notebook processor. This processor fits into an ultrathin notebook design, as thin as 17 millimeters, providing industry-leading visual performance and battery life at very attractive price points. Trinity remains on track to launch for midyear.
We achieved record quarter client revenue driven by an increase in supply of Llano APUs. And in Q4 of 2011, APUs accounted for nearly 100% of mobile microprocessors shipped and more than 60% of the total client microprocessors shipped. Microprocessor ASP increased sequentially due to an increase in mobile microprocessor ASP and an increase in server units shipped.
There is no doubt that the customer acceptance of our APU architecture is quite strong. We’ve now shipped over 30 million of these APUs to date. And we’re seeing a strong uptake in terms of that architecture, what it means to the customer. They are looking for a better experience, and I think that’s a key reason why we’ve seen the momentum in our business and the ability to deliver on that. Our focus on execution around the APUs and around Llano is definitely paying off. And I think as we move forward, we should be able to continue to build on that momentum.
We’ve actually increased our Llano 32-nanometer product delivery by 80% from the third quarter, and now Llano makes up almost 60% of the mobile microprocessing revenue. … We’re going to continue to build on the strong relationships that we’ve been developing with GLOBALFOUNDRIES as we move forward.
The movement to thin and light is nothing new. Customers want mobility. And the idea of ultrathin is something that we’re very focused on. And if you think about it with our APU strategy that I mentioned, with the next-generation product, Trinity APU, we already are well ahead of the pace last year when we set a record-setting year for design wins with the Trinity product in 2012. With that product, we can deliver ultrathin in the range of 17 millimeters. And what’s really important and I think we have to all focus on is ultrathin and mobility, the ability for computing to reach customers across the planet. … And I’ll add that the improvements that we’ve made in Trinity in both our CPU and the GPU are really delivering outstanding results in performance per watt. So as well for the ultrathins being able to hit the 17-millimeter low-profile, we’re also getting a doubling of the performance per watt. So it’s an exciting application of our APU technology.
… as you think of the industry trends around consumerization, cloud and convergence, there’s no doubt, as we’ve seen these kinds of inflection points in the industry, there’s always a significant downward pressure in terms of the price points. So if you’re dragging huge asset base along with you and there comes pressure into the market around those price points, that could put pressure into their [Intel’s] — into a business model. … We think the emerging market and the entry — and the high-growth markets around entry and mainstream will be the hottest segment, and I think that’s playing to our hand. We’re going to emphasize this strategy. We want to embrace this inflection point that’s emerging. We want to accelerate it, because shift happens when there’s these inflecting points.
Of course, we see the investment of our competitor, but the fabless ecosystem is not sitting still. And if you look at the investments that are done on their — TSMC, at a GLOBALFOUNDRIES and a GLOBALFOUNDRIES and alliances level, then the numbers are very comparable. GLOBALFOUNDRIES and their partnership models invest about $9 billion this year. TSMC seeds around $6 billion, if I recall the number correctly. So this is, in terms of scale and absolute numbers, are very comparable to what Intel is putting on the table.
… I feel pretty good about where we are in terms of the transition around 32 nm. … And I want to emphasize, we’ve made real progress, but we’re not finished with that. And we need to continue to work every day with those tiger teams we’ve put in place. We’re tracking the test vehicles through the lines to make sure that we’re getting that consistent improvement, because that will reduce our consumption of wafers and give us far more flexibility in our supply chain. So while we have improved by 80% from the third quarter, we’re not all the way there yet … there’s more yield improvements possible on that 32-nanometer line. … And those same techniques and practices that the teams — the tiger teams applied on 32-nanometer, that momentum continues in the 28-nanometer. And so that poises us well going into the coming 2012.
… I think it’s fair to say from the improvements we have seen and the — and our foundry partners that we are not going to be supply-constrained in the first quarter. … I think the progress we have seen on Trinity has impressed us. And of course, all the learnings that have been done on 32-nanometer with the Llano product will be transferred to Trinity. So the start-off pace with Trinity is going to be significantly better from a yield perspective compared to where we were at Llano launch. So that makes us quite optimistic looking forward.
Here are also a couple of illustrations highlighting that 2011 APU success with the details of new APU strategy additions from Lisa Su‘s (Senior Vice President and General Manager, Global Business Units) presentation for the 2012 Financial Analyst Day held on February 2, 2012 (see her full presentation in PDF):
APUs BRING LEADERSHIP GRAPHICS/COMPUTE IP TO MAINSTREAM [#10]
“Llano” APU offers nearly 3X the performance in the same power envelope over conventional CPUs (2)
Fully leverages the growing ecosystem of GPU-accelerated apps
Source: AMD Performance labs
(1) Testing performed by AMD Performance Labs. Calculated compute performance or Theoretical Maximum GFLOPS score for 2013 Kaveri (4C, 8CU) 100w APU, use standard formula of (CPU Cores x freq x 8 FLOPS) + (GPU Cores x freq x 2 FLOPS). The calculated GFLOPS for the 2013 Kaveri (4C, 8CU) 100w APU was 1050. GFLOPs scores for 2011 A-Series “Llano” was 580 and the 2013  A-Series “Trinity” was 819. Scores rounded to the nearest whole number.
(2) Testing performed by AMD Performance Labs. Calculated compute performance or Theoretical Maximum GFLOPS score (use standard formula of CPU Cores x freq x 8 FLOPS) for conventional CPU alone in 2011 was 210 GFLOPs while the calculated GFLOPs for the 1st Gen APU using standard formula (CPU Cores x freq x 8 FLOPS) + (GPU Cores x freq x 2 FLOPS) was 580 or 2.8 times greater compute performance.
Related new codenames (from the AMD provided At-a-Glance Codename Decoder [Feb 2, 2012]):
“Trinity” APU (Traditional Notebooks, Ultrathin Notebooks and Desktops)
- “Trinity” is AMD’s second generation APU and improves the power and performance of AMD’s A-Series APU lineup for mainstream and high-performance notebooks and desktops. “Trinity” will feature next-generation “Piledriver” CPU cores and new, DirectX® 11-capable, second generation AMD Radeon™ HD 7000 series graphics.
- New for 2012, AMD will offer a BGA or pin-less format, low power “Trinity” APU specifically designed for ultrathin notebooks.
- Planned for introduction: Mid-2012
“Piledriver” Core Micro Architecture
- “Piledriver” is the next evolution of AMD’s revolutionary “Bulldozer” core architecture.
- The “Trinity” line-up of APUs will be the first introduction of “Piledriver.”
“Kaveri” APU (Notebooks and Desktops)
- “Kaveri” is AMD’s third generation APU for mainstream desktop and notebooks.
- These APUs will include “Steamroller” cores, and new HSA-enabling features for easier programming of accelerated processing capabilities.
- Planned for introduction: 2013
“Steamroller” Core Micro Architecture
- “Steamroller” is the evolution of AMD’s “Piledriver” core architecture.
AMD OPTERON™ FUTURE TECHNOLOGY [#26]
Additional new codename (from the AMD provided At-a-Glance Codename Decoder):
“Excavator” Core Micro Architecture
- “Excavator” is the evolution of AMD’s “Steamroller” core architecture.
APU ADOPTION: RECORD DESIGN WINS, STRONG END-USER DEMAND [#11]
Shipped > 30m APUs to date
11 of the world’s top 12 OEMs shipping AMD APU-based platforms
“Brazos” APUs shipped more units in its first year than any previous mobile platform in AMD history
“Llano” APUs ramped to represent nearly 60% of mobile processor revenue by Q4 2011
Additional new codenames (from the AMD provided At-a-Glance Codename Decoder):
“Southern Islands” Discrete Graphics
- Internal codename for the entire family of desktop graphics ASICs based on Graphics Core Next architecture and utilizing 28nm process technology.
- “Southern Islands” products include “Tahiti” (AMD Radeon™ HD 7900 series), “Pitcairn,” “Cape Verde” and “New Zealand.”
“Brazos 2.0” APU (Essential Desktop and Notebook, Netbook, All-In-One and Small Desktop)
- The “Brazos 2.0” family of APUs will follow “Brazos”, AMD’s fastest ramping platform ever.
- In addition to increased CPU and GPU frequencies, “Brazos 2.0” will offer additional features and functionality as compared to “Brazos”.
- Planned for introduction: H1 2012
“Hondo” APU (Tablet)
- “Hondo” is AMD’s sub-5W APU designed for tablets. “Hondo” will feature low-power “Bobcat” CPU cores and support DirectX® 11 technology in a BGA or pin-less format.
- Planned for introduction: H2 2012
AMD started talking about ‘Hondo’ (as well as ‘Trinity’) last summer. See in Acer repositioning for the post Wintel era starting with AMD Fusion APUs [June 17, 2011]
(3) Projections and testing developed by AMD Performance Labs. Projected score for 2012 AMD Mainstream Notebook Platform “Comal” on the “Pumori” reference design for PC Mark Vantage Productivity benchmark is projected to increase by up to 25% over actual scores from the 2011 AMD Mainstream Notebook Platform “Sabine”. Projections were based on AMD A8/A6/A4 35w APUs for both platforms.
(4) Projections and testing developed by AMD Performance Labs. Projected score for the 2012 AMD Mainstream Notebook Platform “Comal” the “Pumori” reference design for 3D Mark Vantage Performance benchmark is projected to increase by up to 50% over actual scores from the 2011 AMD Mainstream Notebook Platform “Sabine”. Projections were based on AMD A8/A6/A4 35w APUs for both platforms.
(5) Testing performed by AMD Performance Labs. Battery life calculations using the “Pumori” reference design based on average power draw based on multiple benchmarks and usage scenarios. For Windows Idle calculations indicate 732 minutes (12:12 hours) as a resting metric; 421 minutes (7:01 hours) of DVD playback on Hollywood movie, 236 minutes (3:56 hours) of Blu-ray playback on Hollywood movie, and 205 minutes (3:25 hours) using 3D Mark ‘06 as an active metric.
Projections for the 2012 AMD Mainstream Platform Codename “Comal” assume a configuration of “Pumori” reference board, Trinity A8 35W 4C – highest performance GPU, AMD A70M FCH, 2 x 2G DDR3 1600, 1366 x 768 eDP Panel / LED Backlight, HDD (SATA) – 250GB 5400rpm, 62Whr Battery Pack and Windows 7 Home Premium.
Additional new codenames (from the AMD provided At-a-Glance Codename Decoder):
“Sea Islands” Graphics Architecture
- New GPU Architecture and HSA Features
- Planned for introduction: 2013
“Kabini” APU (Essential Desktop and Notebook, Netbook, All-In-One and Small Desktop)
- The “Kabini” APU is AMD’s second generation low-power APU and follow-on to “Brazos 2.0.”
- In addition to new “Jaguar” cores, these APUs will be enhanced with new Heterogeneous Systems Architecture (HSA), enabling features for easier programming of accelerated processing capabilities.
- Planned for introduction: 2013
“Temash” APU (Tablet and Fanless Client)
- The “Temash” APU is AMD’s second generation tablet APU and follow-on to “Hondo.”
- In addition to new “Jaguar” cores, these APUs will be enhanced with new Heterogeneous Systems Architecture-enabling features for easier programming of accelerated processing capabilities.
- Planned for introduction: 2013
“Jaguar” Core Micro Architecture
- “Jaguar” is the evolution of AMD’s “Bobcat” core architecture for low-power APUs.
MOBILE MARKET PROJECTIONS [#29] AMD Direction:
Leadership graphics, web applications and video processing leveraging APUs
Agile, flexible SoC designs
Ambidextrous solutions across ISAs and ecosystems
Fanless, sealed designs
These APU related strategic moves have been summarized by the same John Taylor as Strengthening our Client Roadmap [AMD Fusion blog, Feb 2, 2012]:
Roadmaps signify our plans to customers and business partners, outlining the new products and technologies that we are bringing online. In an ideal world plans would never change. But in reality, change is a certainty in the tech industry – new form factors immerge, technologies and applications shift and consumer tastes remake technology plans.
Like any technology company, AMD desires to anticipate change in the industry. So we course-correct as we work with customers to ensure that we create products that address the optimal blend of timing, features and performance, cost and form factors.
Today at our Financial Analyst Day in Sunnyvale, AMD senior staff detailed how AMD will focus its investments in R&D and marketing going forward, including roadmaps for 2012-2013. As Phil Hughes summarized, the announced roadmaps are designed to extend platform longevity, accelerate time to market and enhance performance and features. These roadmaps strengthen AMD’s ability to make the most of shifting market dynamics, all the while giving stand-out experience across device categories through our graphics and video IP. This blog provides some insight into our 2012 and 2013 roadmaps – the words in quotes are the codenames for the particular AMD processor offerings discussed today.
2012 Client Roadmap
AMD’s “Brazos 2.0” Accelerated Processor Unit (APU) family will be used for essential desktop and notebook, netbook, tablet, all-in-one and small desktop form factors. This allows us to address a fast-growing segment of the PC market where we have proven success with the original “Brazos” line-up – the C-Series, E-Series and Z-SeriesAPUs. We will add plenty of new features to the “Brazos 2.0” APU family, including increased CPU and GPU performance, longer battery life, a bevy of integrated I/O options and improvements to AMD Steady Video technology. “Brazos 2.0” is scheduled to hit the market in the first half of 2012.
As we demoed at CES, AMD’s “Trinity” APU for desktop and notebook remains on track for introduction in mid-2012, with plans to pack up to four “Piledriver” CPU cores and next-generation DirectX® 11-capable graphics technology, together delivering up to 50% more compute performance than our “Llano” offerings, including superior entertainment potential, longer battery-life and an even more incredibly brilliant HD visual experience.
New for 2012, AMD will introduce a low voltage “Trinity” APU that will be ideal for the next-generation of ultrathin notebook. This “Trinity” APU matches the experience enabled by the AMD 2011 APU in up to half the TDP. As we said, “Trinity” is on track for introduction in mid-2012.
In 2012 we will also introduce the ultra-low voltage “Hondo” APU for tablets. These low-power (power maxes out at 5W TDP) APUs will have “Bobcat” CPU cores and support DirectX 11 technology in a BGA or pin-less, thin processor package. Look for these in the second half of 2012 – more details to come later.
On the desktop platform side of things, the “Vishera” CPU will replace the “Komodo” CPU for desktop. This change enables accelerated time to market for improved performance and next-generation CPU features while maintaining the existing AM3+ motherboards. The “Vishera” CPU ushers in many exciting updates, includes 8 “Piledriver” cores, and when compared with the previous generation, provides higher frequencies, improved instruction per clock performance, advanced instruction sets (thus increasing application performance), additional DDR3 memory support and next-generation AMD Turbo Core Technology. We plan to launch “Vishera” in the second half of 2012.
2013 Client Roadmap
2013 brings major evolution to the client roadmaps as the vision presented by Rory, Mark and Lisa today begin to manifest – including moving our low power APUs to a system on a chip (SoC) design with the AMD Fusion Controller Hub integrated right into a single chip design.
In the performance APU category our third-generation APU, “Kaveri,”will employ “Steamroller” (the evolution of AMD’s “Piledriver” core architecture) x86 cores for enhanced instructions per clock and power advantages. Applications that take advantage of GPU accelerate will give users an amazing experience thanks to our Graphics Core Next and new Heterogeneous Systems Architecture (HSA) enabling features for easier programming of accelerated processing capabilities.
In the low power category, the “Kabini” SoC APU takes over for “Brazos 2.0.” This second generation low power APU integrates “Jaguar” x86 cores for augmented performance and energy efficiency. These APUs will also benefit from select HSA features and functionality.
We keep on innovating for the ultra-low power space in 2013. Our second generation, ultra-low-power “Temash” SoC APU will follow “Hondo” for tablet and other fanless form factors. This APU will also leverage the “Jaguar” low-power x86 cores and HSA features.
We at AMD strongly believe these roadmap updates help us time new product introductions with customer design phases to hit key sales cycles across a range of form factors and experiences. We are moving with the market and on the path to deliver exceptional productivity and user experience in a wide array of form factors.
John Taylor, Director of Worldwide Product Marketing at AMD
He also provided the following answers to questions regarding how AMD spells out Windows 8 tablet strategy [CNET, Feb 2, 2012]:
Q: Before, we go to Windows 8, what is your smartphone strategy, if any?
Taylor: The smartphone market is eight, nine, ten, maybe a dozen players. [They have] lower ASPs (average selling price), lower [profit] margins, different competitive dynamic. So, there is no shift on the smartphone strategy.
And Window 8?
Taylor: But you will see much more focus on tablets, the convertible or hybrid devices that fit between tablets and notebooks, very thin [designs].
What chips exactly will get you there?
Taylor: For tablets, it will decidedly be the Hondo chip. We’re acknowledging that we still have a couple of watts to shave off to really be a more ideal tablet platform (to achieve optimal power efficiency). But we think that Temash gets us much, much closer to that in 2013.
And Windows 8 convertibles?
A 17-watt [power consumption] is the lowest that we’ll offer. That’s called Trinity. It will be unmatched in that [17-watt design] space. Discrete graphics-like performance. All types of dedicated video processing capabilities, better battery life than the competition. And all of these ways that we’re driving the new generation of accelerated applications. If you think about the Web apps that are being built for Win 8, using HTML5 and the graphics enginethat drives that higher level experience.
I will add to that the following two illustrations from the AMD Product and Technology Roadmaps[AMD FAD, Feb 2, 2012]:
“Vishera” CPU (Desktop)
- The “Vishera” desktop CPU incorporates up to eight “Piledriver” cores, advanced instruction sets and other performance enhancing additions
- This next-generation CPU will maintain the AM3+ infrastructure.
- Planned for introduction: H2 2012
In addition to the above described expansion of the original APU strategy for the clients there is a kind of naming change with AMD Fusion System Architecture is now Heterogeneous Systems Architecture [AMD Fusion blog, Jan 18, 2012]
Since its introduction to the public in June 2011 at the AMD Fusion11 Developer Summit, the AMD Fusion System Architecture (FSA) has received widespread support and interest from our business partners and technology industry leaders. FSA was the blueprint for AMD’s overarching design for utilizing CPU and GPU processor cores as a unified processing engine, which we are making into an open platform standard. This architecture enables many benefits, including high application performance and low power consumption.
Our software partners are already taking advantage of the power and performance advantage of APU and GPU acceleration, with more than 200 accelerated applications shipped to date. The combination of industry standards like OpenCL and C++ AMP, alongside FSA, is ushering in the era of heterogeneous computing.
Together with these software partners, we have built a heterogeneous compute ecosystem that is built on industry standards. As such, we believe it’s only fitting that the name of this evolving architecture and platform be representative of the entire, technical community that is leading the way in this very important area of technology and programing development.
FSA will now be known as Heterogeneous Systems Architecture or HSA. The HSA platform will continue to be rooted in industry standards and will include some of the best innovations that the technology community has to offer.
Manju Hegde and I will be hosting a breakout session on HSA at AMD’s Financial Analyst Day on February 2nd 2012, which will be webcast live here. More information on the latest advances in HSA design will be released at a future date.
Also, if you haven’t already made plans to attend the AMD Fusion12 Developer Summit in June 2012 in Bellevue, Washington, I encourage you to save the date. Leaders from the technology and programming development communities will converge at the summit to discuss Heterogeneous Computing and the next-generation user experiences that are enabled by this platform.
Phil Rogers, corporate fellow at AMD.
From the Analyst Day breakout session presentation I will include the following illustrations here as the food for thoughts and further interests:
For Windows 8 related HSA, “C++ AMP” (indicated on the last illustration) is worth to expand on via Introducing C++ Accelerated Massive Parallelism (C++ AMP) [MSDN Blogs, June 15, 2011]
A few months ago, Herb Sutter told about a keynote he was to delivered today in the AMD Fusion Developer Summit (happening these days). He said by then:
“Parallelism is not just in full bloom, but increasingly in full variety. We know that getting full computational performance out of most machines—nearly all desktops and laptops, most game consoles, and the newest smartphones—already means harnessing local parallel hardware, mainly in the form of multicore CPU processing. (…) More and more, however, getting that full performance can also mean using gradually ever-more-heterogeneous processing, from local GPGPU and Accelerated Processing Unit (APU) flavors to “often-on” remote parallel computing power in the form of elastic compute clouds. (…)”
In that sense, S. Somasegar, Senior Vice President of the Developer Division made this morning the following announcement:
“I’m excited to announce that we are introducing a new technology that helps C++ developers use the GPU for parallel programming. Today at the AMD Fusion Developer Summit, we announced C++ Accelerated Massive Parallelism (C++ AMP). (…) By building on the Windows DirectX platform, our implementation of C++ AMP allows you to target hardware from all the major hardware vendors. (…)”
C++ AMP, as Soma tells in his post, is actually an open specification. Microsoft will deliver an implementation based on its Windows DirectX platform (DirectCompute, as Daniel Moth specifies in a later posta few minutes ago).
Daniel added that C++ AMP will lower the barrier to entry for heterogeneous hardware programmability, bringing performance to the mainstream. Developers will get an STL-like library as part of the existing concurrency namespace (whose Parallel Patterns Library –PPL and its Concurrency Runtime –ConcRT are also being enhanced in the next version of Visual C++ –check references at the end of this post for further details) in a way that developers won’t need to learn a different syntax, nor using a different compiler.
Update (6/16/2011): “Heterogeneous Parallelism at Microsoft”, the keynote where Herb Sutter and Daniel Moth introduced this technology with code and graphic demos is available for on-demand watching.
Update (6/17/2011): Daniel Moth’s session “Blazing-fast Code Using GPUs and More, with C++ AMP” is available as well! Beside, Dana Groff tells what’s new in Visual Studio 11 for PPL and ConcRT.
Pedal to the metal, let’s go native at full speed!
- S. Somasegar’s announcement: http://blogs.msdn.com/b/somasegar/archive/2011/06/15/targeting-heterogeneity-with-c-amp-and-ppl.aspx
- Daniel Moth’s blog post: http://www.danielmoth.com/Blog/C-Accelerated-Massive-Parallelism.aspx
- Herb Sutter’s keynote at the AMD Fusion Developer Summit: http://channel9.msdn.com/Events/AMD-Fusion-Developer-Summit/AMD-Fusion-Developer-Summit-11/KEYNOTE
- Daniel Moth: Blazing-fast Code Using GPUs and More, with C++ AMP (session presented at AMD Fusion Developer Summit): http://channel9.msdn.com/Events/AMD-Fusion-Developer-Summit/AMD-Fusion-Developer-Summit-11/DanielMothAMP
- Announcing the PPL, Agents and ConcRT efforts for Visual Studio 11, by Dana Groff: http://blogs.msdn.com/b/nativeconcurrency/archive/2011/06/16/announcing-the-ppl-agents-and-concrt-efforts-for-v-next.aspx
- AMD Fusion Developer Summit Webcasts: http://developer.amd.com/afds/pages/webcast.aspx
With that in mind the upcoming 2012 AMD Fusion Developer Summit will definitely bring quite important updates as promised by the last breakout session illustration:
More on that: Adobe and Cloudera among Keynotes at AMD Fusion12 Developers Summit [AMD Fusion blog, Feb 3, 2012]
Finally, regarding the ‘ambidextrous’ strategy mentioned in the first sentence of the press release:
- ‘ambidextrous’ generally means ‘very skillful and versatile’ coming from ‘able to use the right and the left hand with equal skill’
- it is described in the press release as:
… adopting an SoC-centric roadmap designed to speed time-to-market, drive sustained execution, and enable the development of more tailored customer solutions. SoC design methodology is advantageous because it is a modular approach to processor design, leveraging best practice tools and microprocessor design flows with the ability to easily re-use IP and design blocks across a range of products. …
- and detailed in Mark Papermaster‘s (Senior Vice President and Chief Technology Officer) presentation for the 2012 Financial Analyst Day held on February 2, 2012 (see his full presentation in PDF) via the following illustrations:
Which caused probably the biggest interest and questions among participating analysts what made even The Wall Street Journal to report as AMD Will Incorporate Others’ Technology in Its Chips [Feb 3, 2011]:
The company on Thursday said it may pursue what it calls an “ambidextrous” strategy that would allow it to offer chips that include circuitry developed by other companies as well as its own. One obvious option would be low-power microprocessor technology from ARM HoldingsPLC that now dominates chip markets for cellphones and tablet computers.
AMD Chief Executive Rory Read, at a meeting with analysts here and in a subsequent interview, stopped short of saying that AMD would definitely add ARM-based technology to its chips in the future. But he noted that the company is laying the technical groundwork for modular chips that could accept blocks of circuitry developed by ARM as well as other companies.
“We have a relationship with ARM, and we will continue to build on it,” Mr. Read said in an interview. “We will continue to evolve that relationship as the market continues to evolve.”
Such possibilities are a sign of how the exploding market for mobile devices is causing many companies to alter their strategies. The x86 design used by AMD and Intel is the foundation of virtually all personal and most server computers.
But the two companies have struggled to make headway in the mobile-device market, in large part because of the lower power consumption of ARM-based designs. Meanwhile, ARM licensees—which include Qualcomm Inc., Texas Instruments Inc. and Nvidia Corp.—are adding to the pressures by edging toward the PC market, as MicrosoftCorp. finishes development of a new operating system that supports ARM and x86 chips.
AMD’s management team, in a meeting with analysts here, took pains to dispute the notion that AMD may become marginalized as ARM-powered competitors enter the PC market. Rather, they argued, AMD’s strength in graphics and microprocessors—and a strategy of customizing chips for large customers—will expand AMD’s opportunities.
Indeed, Mr. Read argued, it is Intel’s outsize influence of the tech industry that will tend to decline. “We will see the breakdown of proprietary control points,” Mr. Read said.
Though Mr. Read didn’t commit to embracing ARM’s designs, others who heard his presentation said the direction is clear. “AMD was very deliberate today about their goal to integrate more third-party intellectual property,” said Patrick Moorhead, a former AMD vice president and now principal analyst at Moor insights & Strategy. “Nothing they communicated excluded the potential for ARM.”
AMD’s remarks also underscore an industry shift—driven largely by the mobile market—away from separate chips and toward multi-function products that the industry calls SoCs, for systems on a chip, which save space and power in mobile devices and other hardware.
Intel and AMD have begun offering SoCs for laptop computers. But AMD discussed extensive plans to create more such products at a faster rate, using a flexible design scheme that can accommodate technology submitted by other companies.
Mr. Read, who previously served as a senior executive at PC maker Lenovo GroupLtd., has recruited others that also worked at IBM and have experience with other chip technologies than x86.
One is Mark Papermaster, AMD’s senior vice president and chief technology officer, who worked at Apple Inc. and Cisco Systems Inc. after leaving IBM in 2008. Another is Lisa Su, a senior vice president and general manager of AMD’s global business units, who most recently worked at Freescale Semiconductor HoldingsLtd., an ARM user.
Ms. Su gave an updated road map for a series of future chips, including products that AMD expects to be used in tablets that are powered by Microsoft’s forthcoming Windows 8 operating system. But Mr. Read said AMD would likely stay away from trying to sell chips for smartphones soon, characterizing the market as too crowded with competitors.