Marvell ARMADA beats Qualcomm Snapdragon, NVIDIA Tegra and Samsung/Apple Hummingbird in the SoC market [again]
Follow-ups: – Marvell SoCs to win both Microsoft and Nokia for Windows Phone and Windows 8 platforms (after the Kinect success) [Feb 1, 2012]
– First real chances for Marvell on the tablet and smartphone fronts [Aug 21 – Sept 25, 2011]
While in the last 2 years Qualcomm Snapdragon, then NVIDIA Tegra and Samsung/Apple Hummingbird (Samsung’s S5PC110 and Apple’s A4 via Intrinsity’s acquisition) got by far the biggest public attention in the System-on-a-Chip (SoC) space it has been Marvell’s ARMADA the real winner all along.
Follow-up: Marvell to capitalize on BRIC market with the Moby tablet [Feb 3, 2011]
Follow-up: Kinoma is now the marvellous software owned by Marvell [Feb 15, 2011]
Follow-up: ASUS, China Mobile and Marvell join hands in the OPhone ecosystem effort for “Blue Ocean” dominance [March 8, 2011]
Follow-up: High expectations on Marvell’s opportunities with China Mobile [May 28, 2011]
Update [Jan 17, 2011]: Report: iPad 2 to use fast graphics chip [Jan 17, 2011] (emphasis is mine)
The iPad 2 will sport powerful, new graphics hardware, along with a higher-resolution display, according to a report.
That graphics chip would be Imagination’s SGX543, according to Apple Insider.
If this rumor is on the money, it is, indeed, a potent graphics technology. Imagination describes the POWERVR SGX543MP as allowing “up to 16 cores…in a high-performance, multiprocessor graphics solution without performance or silicon area compromises.” This graphics tech would be used in conjunction with a dual-core ARM processor, as CNET previously reported.
And Apple’s next-gen iPhone 5 would also feature this chip design–the so-called Apple A5 processor.
“This makes sense,” said Linley Gwennap, principal analyst at the Linley Group, a chip consulting firm. “The A5 processor must have at least dual Cortex-A9 CPUs (central processing units) to be competitive with [Nvidia’s] Tegra 2 and other current smartphone CPUs,” Gwennap said in response to an e-mail query. The Cortex-A9 is a design being used by most major ARM chip suppliers, such as Texas Instruments, Samsung, and Nvidia.
Gwennap continued. “The single-core SGX543 does not have enough graphics performance to keep up with Tegra 2, but a dual-core SGX543 should be within the same range. Even a dual-core SGX543 would fall well behind the graphics performance of Marvell’s new Armada 628, which should be in phones in 2H11,” he said. (Imagination also has the SGX545.)
Addition [later]: Marvell’s tri-core ARM chip has near-PS3-level graphics [Sept 2010], since Sony Playstation 3 GPU has 250 million triangles per second performance vs. ARMADA 628’s 200 MT/s.
Update [Nov 4]: Marvell ARMADA with sun readable and unbreakable Pixel Qi screen, and target [mass] manufacturing cost of $75 on this blog [Nov 4]
Update [Nov 2]: Sehat Sutardja: An Engineering Marvell by IEEE Spectrum [Nov 2, in print Oct 27 but with the title of Marvell Inside] is describing the extremely deep electronic engineering mentality lead with its CEO as the secret recipe for success from the very beginning:
Sehat already had plans for the first product: a better read channel for disk drives. It sounds incredibly specialized and it is, but it’s also one of the drive’s key components. The read channel takes the analog signal coming from the magnetic head as it scans the disk, converts the noisy signal to digital, and puts that information out onto the bus that will take it to the computer. Existing read channels used a bipolar transistor on a complementary-metal-oxide semiconductor substrate (BiCMOS), but Sehat planned to use only CMOS. That way the channels could be manufactured by a chip foundry like the Taiwan Semiconductor Manufacturing Co., so Marvell wouldn’t have to build its own fab. Using CMOS also meant that the device would consume less power. This would, however, present an engineering challenge: Existing CMOS read-channel designs were much slower than BiCMOS.
… they convinced Seagate Technology to take a chance. Ken Burns, an executive at Seagate, told them that the company’s next-generation drive would need a read channel at 240 Mb/s—could Marvell deliver? … They told Burns yes. In less than three months the Marvell team hit the 240-Mb/s mark, and Seagate became Marvell’s first customer. … Today, in terms of units sold, Marvell has about 60 percent of the market for hard drive systems-on-a-chip.
“This little start-up, with one product line, put Texas Instruments out of the read-channel business,” Ohr [an analyst with Gartner] says.
Now we have a new ARMADA SoC processor product from Marvell which is making that lead even more evident:
Marvell Raises Technology Bar Again with World’s First 1.5 GHz Tri-Core Processor Delivering Dual Stream 1080p 3D Video for Smartphones and Tablets [Sept 23] – Game Changer: Ultra-low power, ultra-high performance ARMADA three-core processor first to feature 3D graphics performance with quad unified shaders for 200 million triangles per second delivered on mobile devices
Marvell ARMADA 628 has indeed an impressive set of features on a single die. To quote from the press release (emphasis is mine):
Key Features
- World’s first “tri-core” application processor
– Up to 1.5 GHz for the two main cores and 624 MHz for the third low power core
– “Heterogeneous multiprocessing” with “hardware-based Cache Coherence”
– 1 MB System Level 2 Cache
– Platform leading multimedia capabilities, including support for both WMMX2 and NEON acceleration; and a highly optimized pipelined VFPv3 floating point engine
– Member of the ARMADA family of processors for easy software porting- 1080p dual stream 3D video applications (30 FPS, multi-format)
- Ultimate 3D graphics performance with quad unified shaders for 200 million triangles per second (MT/s)
- High performance, integrated image signal processor (ISP)
- Ability to project images on multiple simultaneous displays
– 2 LCDs
– 1 HDMI
– 1 advanced EPD [like the ones used in E-Ink based e-readers, e.g. Amazon Kindle] controller- Peripherals support: USB 3.0 Superspeed Client, MIPI CSI, MIPI DSI, HDMI with integrated PHY, UniPro, Slimbus, SPMI
Addition [later]: Marvell’s tri-core ARM chip has near-PS3-level graphics [Sept 2010], since Sony Playstation 3 GPU has 250 million triangles per second performance vs. ARMADA 628’s 200 MT/s.
We should emphasize two additional facts from the press release:
The new ARMADA 628 tri-core processor incorporates a number of advanced processing and power management features. The tri-core design integrates two high performance symmetric multiprocessing cores and a third core optimized for ultra low-power. The third core is designed to support routine user tasks and acts as a system management processor to monitor and dynamically scale power and performance. The tri-core architecture provides superior performance and lower power over dual-core designs while maintaining industry compatibility and leadership – ensuring a richer, faster and smoother experience than any other ARM-based processor available today.
… In addition to the tri-core CPU, there are six additional processing engines to support stunning 3D graphics, 1080p video encode/decode, ultra high fidelity audio, advanced cryptography, and digital photo data processing – for a total of nine dedicated core functions.
Additional information from Linley Gwennap‘s Marvell Debuts Tricore Architecture [Sept 23] infonugget: The chip is manufactured in 40 nm geometry [note, this is the first such mobile processor from Marvell publicly announced], all three processors use Marvell’s high-end Sheeva PJ4 design, the third CPU uses however a different circuit layout, based on the same synthesizable core, but optimized for lower speed and lower power, the new chip is fully compatible with ARM’s Neon instruction set (unlike the previous versions of the PJ4), the 3D graphics performance [200 MT/s] is more than four times of the current high-end ARMADA 610 [45 MT/s], and the introduction of an optimized third CPU is an innovative feature which should deliver better battery life than a dual-CPU chip that simply shuts down one CPU. According to Linley Gwennap:
To maximize battery life, the processor uses the smaller CPU most of the time, but it can shift to the powerful dual CPUs when maximum performance is required. The new Marvell processor is currently sampling and due to enter production next March.
With the current availability of the samples to the customers it is “just” the time needed to create the high-volume, consumer market products for the mass availability. Because the ARMADA 628 is a member of a whole family of processors designed for easy software porting it is not a great engineering challenge to develop the new devices based on this new and “marvelous” SoC. Hopefully the device vendors are also well ahead in their business development activities for ultra-low power but at the same time ultra-high performance handheld products in the smartphone and tablet space. If not yet then they will have enough market experience for early next year when all the supply (in addition to the hopefully mass-produced by that time ARMADA 628 SoC) and distribution arrangements could be organised by them.
How the competition is standing up against this leadership offering? Badly, since there is nothing comparable to ARMADA 628 in a 1-2 years timeframe ahaead of us:
Qualcomm: 1.5GHz dual-core Snapdragon devices to arrive at end of 2011, 1.2GHz in Q1 [Sept 8]
Qualcomm 1.5GHz Snapdragon due end of 2011, not Q1 [Sept 8]
NVIDIA CEO: Tegra 3 almost done, Tegra 4 on the way, expect a new Tegra annually [Sept 21]
NVIDIA Finds Its Rhythm: Tegra 3 Imminent with Annual Updates [Sept 21]
which titles may sound quite competitive but in fact these are just mumblings by their desparate CEO as evidenced by these two videos from an unrelated press conference (obviously the CEO was just using the opportunity to “fend-off” the upcoming annoucement known to him already):
– NVIDIA CEO Jen-Hsen Huang talks about Tegra [Sept 21], with a very low quality voice recorded by someone in the audience
– the other one, with a better voice record, embedded in the above mentioned NVIDIA Finds Its Rhythm: Tegra 3 Imminent with Annual Updates article which is making it obvious that the only real specific what was said was:
Tegra 2 is just our first entry (into the mobile market), Tegra 3 is almost done, Tegra 4 is being built. Just think in the context of the NVIDIA rhythm. Every single year, there will be a new Tegra.
Nothing more!
Finally regarding comparable high SoC processor plans from Samsung/Intrinsity[Apple] combo there is no new information at all. The Hummingbird chip (Samsung’s S5PC110 and Apple’s A4) is even no match to the #2 (after ARMADA 628) Qualcomm 1.2 GHz dual-core Snapdragon, also in the sampling phase right now.
When considering the competition one should also consider that Marvell has even high-impact end-customer initiatives as parts of its long-term strategy:
– Moby for education/learning and healthcare/medical (see also their $75 tablet project with Pixel Qi and OLPC, mentioned before, as the natural evolution for Moby)
Follow-up: Marvell to capitalize on BRIC market with the Moby tablet [Feb 3, 2011]
– Plug Computer with only 2 watts of power dissipation (complete for the below spec) for typical applications such as a high performance home server, multimedia server, web proxy etc. The latest CES 2011 release [Jan 5, 2011] is stating even more:
Marvell's Plug Computer 3.0 announced on CES 2011 with a 2 GHz ARMADA™ 300 processor, Embedded Wi-Fi, Bluetooth enabled, 512MB flash, 512MB DDR2, 1.8 inch HD, 1 USB 2.0, 1 Mini USB, 1 10/100/1000 (Gigabit) Ethernet, and SDIO slot.
From streaming the latest winter vacation video to a connected TV in the bedroom, to enjoying favorite music on a Wi-Fi enabled receiver in the living room, Plug Computer 3.0 makes enjoying media on intelligent devices throughout the home easier and on-the-go more enjoyable than before.
The release of Plug Computer 3.0 opens up additional applications and usage in key new markets: Smart Grid, Home Automation, Medical Monitoring, Multimedia Content Sharing, Security and Access Control, Industrial Automation, Agricultural, Mesh and Grid Computing.
More information
* Quite important note for the information given above and below: Marvell has developed and continuing to develop two types of synthesizable cores: PJ1 and PJ4. Unfortunately there is no public information on that except the appearance in various processor products. These cores allow different kinds of core optimisations (an example is the 3d CPU vs dual high-peformance ones in ARMADA 628) and various ARM Instruction Set Architecture (ISA) implementations (as seen below in our product catalogue references). This is a highly competitive differentiator for Marvell.
Update: the PJ1 was also called 88SV331x, and the codename for it “MOHAWK”
- on Marvell by Marvell: A Next Generation Semiconductor Company – History of Marvell CPUs – ARMADA* [overview of the current offerings, not updated yet with ARMADA 628] – Awards
- Marvell Unveils High-Speed CPUs for the Enterprise [Apr 28, 2008]: the “Discovery Innovation Series” MV7800 type embedded processors introduced in single core versions with frequency of 0.6 – 1.2 GHz and dual core versions with frequency of 0.8 – 1.0 GHz. A month later (see the next item) it is disclosed that the high-speed CPUs are the Sheeva PJ1* variants.
- Marvell Sheeva Embedded CPU Technology Set to Drive Next Generation of Consumer, Mobile and Enterprise Applications [June 2, 2008], Marvell Introduces SoCs to Boost Digital Home Gateway and PC Performance: [June 2, 2008]: Sheeva PJ1*-based “Kirkwood series” 88F6000 SoCs announced as samples available. “Marvell’s new SoCs will initially come in three versions, the 88F6180, 88F6192 and the 88F6281, with the 88F6281 capable at running of speeds up to 2 GHz, while consuming an ultra-low sub 2 Watts of power.”
- The Untold Story of Marvellʼs Processor Development, by Linley Gwennap, Principal Analyst [Aug 2008]
- Marvell offers their ideas on design for low power [Oct 10, 2008]
- Marvell Unleashes Powerful New Applications Processor for Mass-Market Digital Consumer Devices [Jan 8, 2009]: PXA168 as the high member of the current, upto 1.2 GHz Sheeva PJ1*-based ARMADA 100 series announced as “sampling now”.
- ARM SoC launched with Linux support [Jan 8, 2009] – this article gives a good overview of the Sheeva PJ1* products from the first “Discovery Innovation Series” MV7800 type embedded processors introduced in May 2008, then the “Kirkwood series” 88F6000 SoCs, also for the embedded market in June 2008, and then to the PXA168 application processor of January’09, later renamed ARMADA 168.
- Marvell Empowers Mass Market TD-SCDMA OPhones with PXA920 Chipset [Sept 8, 2009]: the introduction of Marvell’s one of the most strategic SoC products, the integrated (on a single die) 2.75G (EDGE), 3G (TD-SCDMA) and 3.5G (TD-HSDPA and TD-HSUPA) communication processor with an upto 0.832 GHz Sheeva PJ1* core, the PXA920 for the Chinese market. This got also another name later, Pantheon. It is quite important to realize the strategic intent behind this product, quoting the real customer behind:
“We believe that the PXA920 solution will help us realize China Mobile’s vision of sub-1000 RMB [~$150, unsubsidized] TD OPhones in the near future,” said Bill Huang general manager of China Mobile Research Institute. “Marvell has worked with us from the start of the PXA920 program and we are excited by the rapid progress we have made towards realizing this milestone. China Mobile will work closely with Marvell and handset eco-system partners to deploy the PXA920 and we look forward to the rapid launch of next generation TD-SCDMA OPhones based on the Marvell PXA920.”
Update: That opportunity was realized only 2 years later. See:First real chances for Marvell on the tablet and smartphone fronts [Aug 21, 2011]
- New Marvell HD Media Processor System-on-a-Chip Brings Award-Winning Qdeo Processing to Next-Generation Consumer Electronics Devices [Sept 16] : the introduction of 88DE3010 High Definition (HD) Media Processor System-on-a-Chip (SoC) with two Sheeva PJ1* cores at up to 1.2 GHz, a month later also getting a new name, ARMADA 1000. This was also the last Sheeva PJ1* variety SoC introduced so far.
- Marvell Launches ARMADA – New Application Processors for Next Generation Mobile Computing and Intelligent Connected Devices [Oct 19, 2009]: rebranding of Sheeva based products and introducing the first Sheeva PJ4*-based product in the family, the ARMADA 510 having a frequency of 1.2 GHz and with “sampling to customers this quarter”. With this Marvell introduced ARM v6 / ARM v7 capability and also was one of the industry’s first running ARM v7 instruction set. This was also an important milestone for Marvell’s new Sheeva-based product strategy (emphasis is mine):
The ARMADA family of application processors sets a new standard by offering CPU cores with PC-class performance, support for Adobe Flash technology and Blu-ray functionality. Additional features include complete broadband connectivity for the next wave of innovative electronics, spanning price points from low cost consumer gadgets all the way to premium performance devices. Customer adoption has been rapid: to date, Marvell ARMADA application processors have won more than 50 design wins.
- On ARMADA by Anand Lal Shimpi: Marvell’s ARMADA: Custom Designed ARM SoCs Break 1GHz [Oct 19, 2009] — please note that the only objection raised here, “The biggest issue I can see with Marvell’s PJ4 is that it doesn’t support ARM’s NEON SIMDfp instruction set” is not existing any more with ARMADA 628.
- Marvell Introduces ARMADA 300 — Setting New Embedded Computing Performance Standards [Dec 16, 2009]: the latest members of the Sheeva PJ1*-based embedded “Kirkwood” series processors, the 88F6282 and 88F6283 are introduced as the ARMADA 300/310SoCs. The high-performance ARMADA 300 with frequency of 1.6 – 2.0 GHz, low power ARMADA 310 with frequency of 0.8 GHz and 1.0 GHz. Quoting the press release:
The ARMADA 300 series offers more than 2.0 GHz processing horsepower in a sub 2W power envelope [it is even less for the ARMADA 310: “under normal load, this device operates on average in less than one watt power consumption envelope” said Sahat Sutardja, Marvell’s CEO at the March 4, 2010 Earnings Call] for the entire System on Chip – a level of performance never before seen with this class of processor – and the flexibility, interfaces and price points that fit the needs of today’s digital economy.
- Marvell Changes the Game Again – New ARMADA 610 Application Processor First to Bring 1080p Full-HD Encode and Decode and 3D Multimedia Performance to Mobile Connected Devices [Jan 5, 2010]: the introduction of the second Sheeva PJ4*-based member of the ARMADA family. The ARMADA 610 has been designed for mobility, features a 1 GHz CPU, has an integrated 3D engine which renders 45M triangles-per-second (via a complete floating point pipeline and unified vertex and fragment/pixel shading) for an immersive gameplay experience with the ability to drive the latest in 3D enabled user interfaces, a lot of integrated peripheral controller on the same die etc. It is targeted at “eReader, Mobile Internet Devices (MIDs), tablets, smartbook/netbooks, connected portable media players, smartphones, and new personal information appliances.” Subsequently it was “designed to use extremely low power while maintaining high processing performance [2.42 DMIPS/MHz] and attractive price points for manufacturers.” It was “currently sampling to early customers.”
- Marvell Announces Another Breakthrough in Chip Technology: World’s First Quadruple Core Processor for ARM Instruction Set [Jan 6, 2010] (emphasis is mine):
Based on the same CPU architecture as the Marvell® ARMADA™ 500 and 600 processor series, Marvell’s quadcore implementation can deliver gigahertz-plus processing per core and is designed for customer-specific products such as mass consumer market and high volume gaming applications.
“Introducing our quadcore technology to the world represents a pivotal moment in CPU development for the consumer electronics industry,” said Ms. Weili Dai, Marvell’s Co-founder and Vice President and General Manager of Marvell Semiconductor, Inc.’s Consumer and Computing Business Unit. “Today’s media-rich consumer applications are already pushing the limits. By making quadcore capabilities available to our customers we will enable the newest generation of cutting edge devices that consumers will always demand — more horsepower, higher performance, better battery life, and more attractive price points than ever before for mass consumer market adoption.”
Marvell has a long history of delivering multicore technology to customers for use in a broad variety of applications ranging from pachinko machines, printers, gaming, networking, gateways, all the way up to computing-intensive enterprise applications. This announcement of the first implementation of quadcore technology for the ARM ecosystem further demonstrates Marvell’s ability to deliver high performance, flexible technology that meets the silicon requirements of numerous tier-one customers, regardless of the end application.
“With the introduction of the ARMADA line of processors, we see the wide variety of devices that Marvell’s application processors can power,” said Rob Enderle, Principal Analyst for the Enderle Group. “Before ARMADA, the ARM ecosystem was thought to be limited by performance barriers. Now, with this announcement of its quadcore technology, Marvell is showing the world the ARM ecosystem’s true potential while cementing its position among the leaders in advanced CPU development for mass market consumer applications.”
Marvell’s quadcore solution is designed to meet customer specific requirements for mass consumer market opportunities. For more information about Marvell quadcore technologies please visit www.marvell.com or contact a sales representative.
- Marvell Unveils 1.6GHz Quad-Core ARMADA XP Platform for Enterprise-Class Cloud Computing Applications [Nov 8] (emphasis is mine):
… the Company will demonstrate its quad-core processing, enterprise-class cloud computing platform, Marvell® ARMADA™ XP (Extreme Performance). The ARMADA XP is the fastest ARM processor available on the market today for enterprise class applications. The new platform integrates four Marvell designed ARM compliant 1.6GHz CPU cores along with a host of I/O peripherals to offer one of the highest levels of integration in the industry. By employing advanced design methodology and process technology, the ARMADA XP is optimized to consume strikingly low power at 1.6GHz, delivering the best performance per watt [16,600 DMIPS performance at less than 10 watts see below in the “Key Features” section] to empower emerging cloud computing applications ranging from high performance networking and web servers to high volume home server products like Network Attached Storage (NAS) and media servers.
Marvell’s introduction of a powerful solution for enterprise-class cloud computing applications is a very important milestone in the mobile Internet revolution—cloud computing mobile servers like those powered by the ARMADA XP are the key link in what I envision to be a seamless, unified ecosystem of mobile connected devices, information appliances and smart ‘furnishings,'” said Weili Dai, Co-Founder of Marvell. “Marvell’s leadership in mobility, consumer, storage, enterprise networking and Wi-Fi products completes the circuit, delivering a powerful end-to-end total solution to anyone connected to the new global mesh, from consumers to small business and the enterprise.”
The ARMADA XP supplements the Marvell Plug Computer initiative by enabling a new class of mobile servers to serve the growing performance demands of connected consumer devices like smartphones and tablets. The new quad-core further builds on the success of the Marvell® Discovery™ Innovation and ARMADA 300 series by maintaining software compatibility to offer existing Marvell customers significant advantage in terms of reduced development cost and faster time to market.
About ARMADA XP series
The ARMADA XP is based on Marvell-designed ARM v7 MP compatible CPU offering 1.6GHz processing per core performance, delivering 16,600 DMIPS to make high performance computing affordable for mainstream applications. It integrates 2 MB of L2 cache and supports 64-bit DDR3 memory interface with ECC at 800MHz clock rate to enable a high throughput memory sub-system design. The ARMADA XP is a highly integrated System-on-a-Chip (SoC) that combines quad x4 PCI-express (PCI-e) interfaces, multiple USB ports, Gigabit Ethernet ports, SATA ports, security engine and other I/O peripherals to make system designs simple and economical. With ARMADA XP’s advanced power management architecture, it offers the industry’s best performance per watt to alleviate the challenges of energy and cooling costs faced by enterprise and server class systems.
Key Features
- Industry’s first quad-core ARM processor for enterprise applications
- Up to 1.6GHz processing performance for each ARM v7 compliant core
- 16,600 DMIPS performance at less than 10 watts
- “Heterogeneous multiprocessing” (SMP/AMP/Mixed) with “hardware-based Cache Coherence”
- Up to 2MB system level two cache
- Supplements the ARMADA family of single, dual-core and tri-core processors for easy software porting
- 64-bit DDR2/DDR3/DDR3L memory interface with ECC support at up to 800MHz clock rates
- 4 PCI-e Gen 2.0 units
- 4 enterprise class Gigabit networking ports
- Up to 16 high speed Marvell SERDES lanes with multi functionality (PCI-e, SATA, SGMII, QSGMII)
- Multiple USB ports
- Ultra low power consumption with advanced power management capabilities
Comprehensive Development Tools Marvell offers complete development platforms for the ARMADA XP enabling customers to start system development without waiting for hardware. Development platforms are available including software drivers and board support package.
Availability
Marvell’s ARMADA XP will be on display at the ARM Technology Conference on November 9-11, 2010 at booth #200. The ARMADA XP is currently sampling to customers.
Official Marvell product catalogues
- ARMADA 100 the current Sheeva PJ1*-based, “cost sensitive” application processor products (showed along with the preceding PXA3xx – Monahans which came with Intel’s XSCale aquisition). Frequency: 0.4 – 1.2 GHz (the upper limit is missing on the ARMADA* overview page). Instruction Set Architecture (ISA): ARMv5 / XScale.
- ARMADA 300 the Sheeva PJ1*-based embedded processors, currently the ARMADA 300/310SoCs (high-performance ARMADA 300 with frequency of 1.6 – 2.0 GHz, low power ARMADA 310 with frequency of 0.8 GHz and 1.0 GHz), which came in addition to the “pre-ARMADA” Discovery (single core with frequency of 0.6 – 1.2 GHz and dual core with frequency of 0.8 – 1.0 GHz) and Kirkwood (single core with frequency of 0.6 – 2.0 GHz and dual core with frequency of 0.8 – 1.0 GHz) series of PJ1* based products. Instruction Set Architecture (ISA): ARMv5.
- ARMADA 500: the high-end Sheeva PJ4*-based application processors for high performance consumer devices such as netbooks and smartbooks (now would rather be called slates, tables etc.). Currently the ARMADA 510 SoC with frequency of 1.2 GHz. Instruction Set Architecture (ISA): ARM v6 / ARM v7. One of the industry’s first running ARM v7 instruction set.
- ARMADA 600: the high-end Sheeva PJ4*-based application processors bringing high performance to the most compact form factors, such as smartphones and embedded mobile devices. Currently the ARMADA 610 and ARMADA 618 of 1.0 GHz SoCs (check for documentation on ARMADA 628). Instruction Set Architecture (ISA): ARM v6 / ARM v7. One of the industry’s first running ARM v7 instruction set.
- ARMADA 1000: for digital entertainment application processors, i.e. “the next generation of connected full-HD consumer devices, delivering immersive viewing experiences and offering a variety of networked applications at mainstream price points”. Currently the Sheeva PJ1*-based two-core ARMADA 1000/88DE3010 High-Definition Media Processor with frequency of upto 1.2 GHz. Instruction Set Architecture (ISA): ARMv5 / XScale. TDP: 5W (see: Nixeus Fusion XS Brings Marvell into the DMA Market [March 11, 2011]
- ARMADA XP: for enterprise-class cloud computing applications, i.e. “empower emerging cloud computing applications ranging from high performance networking and web servers to high volume home server products like Network Attached Storage (NAS) and media servers”. Quite probably based on Sheeva PJ4* cores, with one, two and four core versions, and a corresponding variety of I/O peripherals on the same chip.
- In the ARMADA XP Product Brief (quite worth to look at) you can find a broader target market defined as: “With its broad offering of 5 pin compatible chips, the ARMADA XP is ideally suited for applications ranging from high-performance networking, wireless infrastructure and web servers to high volume products like NAS, home servers, laser printers and other embedded applications.” This is also showing that the ARMADA XP series SoC is a companion to Marvell’s embedded products as well (shown under embedded for this reason as well), i.e. the Sheeva PJ1* based Discovery, Kirkwood and ARMADA 300 series. The ARMADA XP series of Instruction Set Architecture (ISA) is ARM v7 (ARM v6 is not mentioned why it could — quite probably — have an enhanced Sheeva PJ4* core).
Microsoft to lead standards compliance and implementation? … or how Microsoft is aiming to create a radically new Windows client platform via a set of “whole computer capable rich web” standards.
This is the question and the final conclusion I came to after studying all the details related to the announcement of Microsoft and Top Sites Celebrate the Beauty of the Web With Internet Explorer 9 Beta Release [Sept 15]. Let’s see the accompanying fact sheet Windows Internet Explorer 9 New Features at a Glance which has the following grouping and the related to my question major statement excerpts about the IE9 Beta (emphasis used within the excerpted text detail is mine):
Hardware-accelerated graphics
As an example of how Internet Explorer 9 takes advantage of the power of the whole computer, the rendering of graphics and text has been moved from the central processing unit (CPU) to the graphics card (the graphics processing unit or GPU), using the Direct2D and DirectWrite sets of Windows application programming interfaces (APIs). Hardware-accelerated text, video and graphics mean that your websites perform like applications installed directly on your Windows-based computer.
New DOM and new JavaScript engine
The newly optimized document object model (DOM) in Internet Explorer 9 provides dramatic speed improvements by interacting more efficiently with Chakra, the new JavaScript engine. Chakra interprets, compiles and executes code in parallel by taking advantage of multiple CPU cores. Although each of these is significant on its own, combining these changes, along with using hardware-accelerated graphics, makes the browser all-around fast.
F12 developer tools
Clean site-centric design makes sites shine and integrates them with Windows 7:
Clean browser user interface Pinned Sites JumpLists Windows Aero Snap for your websites Thumbnail preview controls Icon overlays Notification Bar New tab page One Box Address Bar Top Result Feel the confidence and trust that you are in control with Internet Explorer 9:
Download Manager with SmartScreen filter integration Add-on Performance Advisor Hang recovery Compatibility View Automatic updates Group Policy support Write interoperable markup with HTML5 and Internet Explorer 9:
Extensive support for HTML5, Scalable Vector Graphics (SVG), Cascading Style Sheets Level 3 (CSS3), ECMAScript5 and DOM provides a new set of capabilities that will help enable developers to write one set of markup and know that it will work and look the same in all modern browsers. Internet Explorer 9 was designed with support for industry standards built in to help ensure that the same markup works the same across browsers.
• HTML5 support. Internet Explorer 9 builds on the work done to implement HTML5 features in Internet Explorer 8, and adds several compelling features. Support for the video and audio elements enables native, hardware-accelerated video and audio content on a Web page without the need for a plug-in. Developers can now insert a video or audio clip onto their page as easily as they do images. Plus, support for the canvas element enables easy and dynamic graphics rendering, all while taking advantage of hardware acceleration through Windows and the graphics card. In addition, support for the selection APIs enables programmatic selection of text on a page, and HTML parsing improvements help make HTML authoring more versatile.
• DOM Level 2 and Level 3. Internet Explorer 9 adds support for more of the Document Object Model Level 2 (DOM L2) and Level 3 (DOM L3), and improves DOM L2 support over existing implementations. These DOM additions are taken from several DOM specifications, including DOM L2 and L3 Core, DOM L2 Views, DOM Element Traversal, DOM L2 and L3 Events, DOM L2 HTML, DOM L2 Style, DOM L2 Traversal and Range, and WebIDL (interactive data language).
• SVG. As the SVG standard has developed, developers have been requesting native support in Internet Explorer, and it is available in Internet Explorer 9. Support for SVG in Internet Explorer 9 enables powerful, attention-grabbing visuals with incredible detail, all without the need for a separate download or plug-in. Like all the graphics, text and media features in Internet Explorer 9, SVG in Internet Explorer 9 takes advantage of hardware-accelerated graphics.
• CSS3. Building on the work that was done in Internet Explorer 8, which is fully compliant with the Cascading Style Sheets Level 2.1 (CSS2.1) specification, Internet Explorer 9 adds support for many components of CSS3, enabling even more flexibility and functionality for Web designers and developers. Internet Explorer 9 introduces features from several CSS3 modules, including the Backgrounds & Borders Module, Color Module, Fonts Module, Media Queries Module, Namespaces Module, Selectors Module, the Values & Units Module, and support for the Web Open Font Format (WOFF).
• ECMAScript 5. The JavaScript implementation in Internet Explorer 9 is enhanced with many features defined by the latest edition of the ECMAScript standard. New ECMAScript 5 features introduce significant improvements to the JavaScript language and increase developer productivity. In addition, the Internet Explorer 9 DOM is designed to natively support ECMAScript 5, providing a consistent and natural programming model for developers when programming the Internet Explorer 9 DOM from JavaScript.
With things like that it is clear that Microsoft is aiming at a radically new Windows client platform creation which is based on the latest “rich web” standards capable of taking advantage of the power of the whole computer. In that sense what has debuted now as Internet Explorer 9 Beta is not less than:
- the fist implementation of that new Windows client platform, and also
- the live laboratory of platform development alongside with the development of new “rich web” standards.
The final questions are certainly how efficient is the current implementation and how much the latest “rich web” standards are covered by IE9 Beta?
To answer those two question let’s turn to the technology media leaders on the web having the opportunity to analyze the new release not less than week before it has been released by Microsoft:
Engadget Internet Explorer 9 Beta review [Sept 15] concentrated on completely redesigned fuctionality and performance, not advancements in the standards space (btw a pretty complicated issue):
IE9 bested Firefox 3.6 in lots of the tests, but Chrome still won out in them all. … What doesn’t really come through in those benchmarks is the browser’s hardware accelerated graphics. … There isn’t all that much in terms of graphics-heavy HTML5 sites at this point in time and Flash 10.1 already relies on the GPU, but we did try Microsoft’s Test Drive suite of sites in a number of different browsers. The JavaScript-based Amazon Shelf demo … is pretty stunning; on the M11x with the GPU activated, the demo ran at 60fps (about 55fps when we turned a page in a book). With the GPU off, the experience was a bit more sluggish – it ran at 16fps and 9fps when turning a page. … How does that Amazon Shelf demo work in other browsers? Both Chrome 6.0.4 and Firefox 3.6 don’t take advantage of the GPU, so even when it was turned on it notched 6fps. The results were much better in Firefox 4 Beta 5 which is optimized for GPU acceleration — it hit the 60fps mark
ZDNet Internet Explorer 9 beta review: Microsoft reinvents the browser [Sept 15] tried to answer questions readers were typically asking: Is it fast enough? Is it compatible enough? Is it cool enough to win back former IE users who have switched to other browsers, first to Firefox and more recently to Google Chrome? And will this shiny new browser be able to rehabilitate the tarnished Internet Explorer brand? From ZDNet’s review there was again answer only to my performance question (emphasis in the quoted text is again mine):
The single biggest performance boost in IE9 comes from its support for hardware acceleration. Because IE9 runs only on Windows Vista SP2 and Windows 7, it can be tuned to offload some rendering tasks to modern graphics hardware, which often has more raw processing power than the rest of the PC. (Microsoft claims that current browsers use only 10% of a PC’s power, which might be a bit of hyperbole.) It’s clear from daily use, though, that hardware acceleration really does make a difference in rendering text, images, and graphics. As a result, Microsoft finds itself in an unaccustomed position, out in front of other browsers, which are furiously trying to play catch-up.
I tested the IE9 beta alongside Firefox 4 beta 5, which was released in September 2010 and is the first Mozilla offering to support hardware acceleration. I also tested it against the most recent beta of Google Chrome 6, which doesn’t use the GPU for rendering. (Google has reportedly placed that feature on its roadmap for Chrome 7.) … The biggest performance differences, not surprisingly, were apparent on Microsoft’s own graphics-oriented tests at its IE Test Drive site. On the FishIE Tank example, which uses the new HTML5 Canvas tag, here’s how the three browsers compared: … IE9’s frame rates stayed high as I kicked up the number of animated fish in the virtual tank. Performance remained smooth and glitch-free even when I moved the window across multiple monitors and docked it to the side of the display using Aero Snap. Firefox 4, by contrast, was able maintain high frame rates for short bursts, but moving the browser frame caused performance to plummet and even froze the display for long periods. Using Firefox, frame rates plummeted dramatically when I selected the most demanding settings (500 and 1000 fish). …
For a more independent performance test, I enabled all three browsers for YouTube’s HTML5 channel and tried playing a handful of high-definition videos at 720p and 1080p resolution. All three browsers performed admirably within a window and at full-screen resolution. IE9 and Chrome 6 were able to maintain full-fidelity playback even when tearing a tab out of the browser pane and dragging it to its own window. Firefox 4, on the other hand, failed this test, stopping the playback and starting the clip over when it landed in a new window.
The other new performance-enhancing component in IE9 is the new Chakra JavaScript engine, … ran the SunSpider benchmark using only the most recent beta releases of IE9, Firefox, and Chrome. The difference between each browser is only about one-tenth of a second, and that composite result includes dozens of complex operations. The independent JSBenchmark test produced similar results, with the IE9 beta running 21% faster than the latest Firefox 4 beta but 29% slower than the latest test build of Google Chrome 6. The conclusion? JavaScript performance isn’t a significant differentiator between modern browsers, and IE9 can hold its own with any Webkit-based browser on this score.
Based on these two indepedendent reviews (and a lot of others with similar findings) I can conclude that performance-wise Microsoft is on track to create the radically new Windows client platform. From the point of view of upcoming “rich-web” standards, however, I should do my own investigation. That will come in the next post in this blog.
Be aware of mainland China and Taiwan stronger manufacturing links in ICT
A current article from Knowledge@Wharton is drawing our attention to the fact that mainland China and Taiwan are fast becoming essentially one in the important ICT sector, and this will further increase their global influence. See: Computer Compatriots: Taiwan and China Draw Economically Closer [Sept 1]
We are talking here about things like the already existing fact that more than 90% of notebook and netbook computers manufactured worldwide are from the combined manufacturing bases of Taiwan and mainland China. Things like that will be significantly increased in the future due to the Economic Cooperation Framework Agreement (ECFA) that Taiwan and China signed in June.
Important updates:
– ECFA to Further Drive China`s Procurements in Taiwan [Jan 5, 2011]: “Last year China organized 13 official buy-Taiwan groups, which together purchased products worth about US$20 billion. This year the projected groups may focus their purchases of 539 duty-free items as on the ECFA fast-track gainer list. The first group to arrive in 2011 will be from Liaoning, said to be led by provincial governor Chen Zhenggao and slated to arrive in Taiwan in mid-February.”
– Taiwanese Banks Lend Over US$31 B. to Clients in China [Oct 11]: “Some say that large enterprises in Taiwan borrow from domestic banks and then transfer part of the loan to affiliates in China. So, including such transferred loans, China would have borrowed more than NT$1 trillion or US$31.25 billion from Taiwan. However, some say that the value is unknown for the exact amount of loans flowing to Taiwanese firms in China through their parent companies in Taiwan is not tallied.”
– ECFA to Bring Strong Boost to Taiwan`s Economy [Sept 21]: “In its initial period, ECFA calls for the removal of tariffs on more than 800 items on an “early-harvest” list in three stages over a two-year period. The list includes 530 items included at Taiwan’s request (worth an annual US$13.8 billion in shipments to China) and 267 items included at China`s request (worth an annual US$2.86 billion in shipments to Taiwan). ECFA will also further open up cross-straits financial dealings. … Given Taiwan`s growing dependence on the Chinese market, the institutionalization of economic relations across the Taiwan Straits is vital to the continued development of the island`s economy.”
– Cross-Strait Investment Protection Agreement May Be Signed by Year End [Sept 17] — With cross-Taiwan Strait Economic Cooperation Agreement (ECFA) having taken effect, Taiwan and China can start talk on investment protection agreement.
– Taiwan should emulate investment strategies of local governments in China, says Acer founder [Sept 29]: “The Taiwan government has been encouraging international enterprises to set up regional headquarters and R&D centers in Taiwan for a while, but has said more than it has actually done for incentives it offered and promises made. … Shih also recommended that Taiwan-based manufacturers shift some of their resources to own-brand business operations and those related to proprietary intellectual property so as to avoid price competition.”
ECFA is prompting both sides to capitalise on this opportunity. Only the Taiwanese interest is, certainly, well manifested in the public media:
– Ministry [of Economic Affairs in Taiwan] works to attract major ICT and auto firms [July 4]
– Ma expects Taiwan to become a regional trade hub after ECFA [Aug 18]
This is all despite of the current political and economic interests, ties:
– Inconvenient impacts of Taiwan-PRC ECFA [Aug 5]
3d Parties are aiming to capitalise on the opportunity as well, which could only enhance the role of new mainland China and Taiwan cooperation further:
– Beyond Geopolitics – The Case for a Free Trade Accord between Europe and Taiwan [July 8, 70 pages]
– ECFA could open way for EU-Taiwan FTA [July 8]
– ECFA could be Taiwan’s window for trade deals: [a European] think tank [July 29]
– Israel welcomes ECFA as good for business [Aug 26]
Keep in mind, however, that ECFA is quite controversial from general social and public interest point of view. There are even some unexpected contradictions for outside observers unaware of the peculiarities of the “internal” situation. To demonstrate that, here is a notable exerpt from Taiwan: Let’s go poking around under the rock of ECFA [Aug 24] blog post, written after the Legislation Yuan of Taiwan approved the ECFA agreement on August 17:
After Hong Kong signed CEPA, it has become the city with the largest wealth inequality in the world. We don’t want an ECFA that let the rich get richer and the poor get poorer.
Some general background information on the “players” and the “play” (not ECFA in particular):
– Taiwan: About Taiwan [April 11, 2008]
Over the past 20 years, Taiwan has become the world’s fourth largest ICT (Information and Communication Technology) hardware producer and holds #1 ranking WW ICT product market share with more than 10 particular products (e.g. Notebook PC, Desktop PC, Computer Motherboard, Computer Server, CDT monitor, LCD monitor, Optical Disk Drive, Digital Still Camera, WLAN, Mobile Phone, and PDA etc.) By estimation, 75% of PCs installed in the world with Windows OS are produced by Taiwanese IT companies. Today, the entire production base has mostly migrated to mainland China due to low cost in labor and land factors. However, the Taiwanese entrepreneurs are dominating at least 75% of the ICT hardware production value produced in the PRC (People’s Republic of China). Impressively, Taiwan is also the fourth largest semiconductor industry in the world. Naturally, Taiwan is a major procurement center (One-Stop-Shopping) for global ICT companies.
– Taiwan [from wikipedia]
– Distinctive Characteristics of China’s Path of ICT Development: A Critical Analysis of Chinese Developmental Strategies in Light of the Eastern Asian Model [May 12, 2008]
– Adapting to the China Challenge: Lessons from experienced multinationals [Sept 20, 2007]
– Sons of The Yellow Emperor Go Online: The State of the Chinese Digital Diaspora [July 2, 2010]
Experiencing the Cloud 
