All the latest quantum computer articles

See the latest stories on quantum computing from eeNews Europe
Showing posts with label data center. Show all posts
Showing posts with label data center. Show all posts

Sunday, July 24, 2016

Farewell to the ITRS semiconductor roadmap

This month saw the last report of the international technology roadmap for semiconductors, the official publication that tracked the march of Moore's Law over the last 25 years. 
By Nick Flaherty www.flaherty.co.uk

The final installment highlights the challenges over the next 15 years as the advances of silicon technology become more difficult to maintain. With 7nm systems being produced this year, that's still quite a long way to go, and there are major challenge such as extreme ultraviolet (EUV) that have still to be addressed.

That is to to say its the end of the silicon business. If anything, the roll out of the Internet of Things and mass data centres demonstrates the longevity of silicon as the substrate of the modern world. Imagine what you could do in ten years' time with 7nm or 5nm silicon providing with billions of transistors for a dollar. Radio integration and power consumption will still be a major challenge.

Data from the report highlights the need for a change. For example, it sees data centres having nearly 5m processor  cores by 2021, doubling to over 10m by 2029, with significant increases in memory storage and networking performance and power consumption. But this may well start to be overtaken by photonics and quantum computing within that time. 
Predictions for data centre growth to 2029 Source: ITRS


The ITRS will no longer be tracking these developments directly. “For a quarter-century, the Roadmap has been an important guidepost for evaluating and advancing semiconductor innovation,” said John Neuffer, president and CEO of Semiconductor Industry Association. “The latest and final installment provides key findings about the future of semiconductor technology and serves as a useful bridge to the next wave of semiconductor research initiatives.”

While the final ITRS report charts a path for existing technology research, additional research is needed for that transition to an even more connected world, enabled by innovations like IoT where it is the integration and co-existence of technologies that is the challenge rather than the sheer performance. Some of these technology challenges were outlined in a recent SIA-Semiconductor Research Corporation (SRC) report, “Rebooting the IT Revolution."

“SIA appreciates the hard work, dedication, and expertise of those involved with the ITRS over the years and looks forward to continuing the industry’s work to strengthen semiconductor research and maintain the pipeline of semiconductor innovations that fuel the digital economy,” said Neuffer.

Key learning points from the report will be on the Embedded blog in the coming days. 

Monday, July 04, 2016

IoT drives software defined power into the data centre

By Nick Flaherty www.flaherty.co.uk

The increased demands from the Internet of Things (IoT) is driving a major shift in the industry at the infrastructure level, says Jeff Schnabel, vice president of global marketing at Portland, Oregon-based power supply maker CUI, in an interview with EETimes Power.
“One of our big plays is in infrastructure with data centres with zetabytes of data and that has huge implications. On the infrastructure side is the idea of software defined power (SDP) – it’s the next evolution of power. With digital power you have these islands of power – SDP ties that together at a higher level. We have partnered with Virtual Power Systems working on the software for SDP solution for data centres to drive the virtualization of power.”
One of the big issues is capacity utilisation, he says. Data centre operators over-provision the power to support peak use, so the infrastructure is only really used at 60% in order to cope with the spikes of power. Using virtualization to provide the power where and when it is needed can greatly reduce capex and opex spending and the amount of equipment to power the infrastructure – racks of power supplies can be replaced with racks of servers to boost the compute capacity. 
Within the hardware is a 4U power system with a battery backup built in. This provides the ability to use ‘peak fade’ so instead of building a 10kW because you need it at peak, you can build in 6kW and use the battery back up at peak times and recharge when the peak comes back down. This means operators are able to shave the peaks and remove 4kW of infrastructure. This has a huge benefit in capex when you are building the data centre but also gains in opex and reduce the thermal envelope as cooling is a huge issue.
“The great thing is these are plug and play so they can be installed into the existing infrastructure in weeks rather than months or years, he said. “It supports greenfield design but is also plug and play so an operator can quickly realise the gains.”
This move to software is also being driven by a reduction in the number of skilled power engineers. You can read more of this interview at EETimes Power

Tuesday, June 21, 2016

Next generation RapidIO targets 10Gbit/s for scalable infrastructure

By Nick Flaherty www.flaherty.co.uk

RapidIO.org has launched the next generation of its interconnect architecture, targeting 25xN 100Gbit/s performance for 5G wireless and data centre infrastructure as well as industrial embedded systems.

With more than 200 million RapidIO fabric ports deployed worldwide at speeds of 6xN and 10xN 10-50 Gbit.s in analytics, wireless infrastructure, industrial and military applications, the 25xN 100 Gbit/s specification provides an open standard interconnect fabric for systems requiring heterogeneous coherent computing infrastructure that can scale with non-volatile storage.

“With the public release of the specification, our focus within our RapidIO.org Technical Working Group now shifts toward our Coherent Scale Out Task Group and our Non-Volatile Storage Task Group,” said Rick O’Connor, Executive Director of RapidIO.org. “Work on the Coherent Scale Out and Non-Volatile Storage draft specifications is progressing and we invite members of industry to join us and participate in defining these key building blocks for flexible, heterogeneous, coherent scale out of performance critical, low-latency applications over RapidIO fabrics.”

“The public release of the 25xN 100 Gbps RapidIO Specification is the culmination of efforts and contributions of RapidIO.org member companies,” said Paul Carson, Chairman of RapidIO.org and Director of IP Development at Texas Instruments. “This work extends the performance roadmap of RapidIO fabrics for use in heterogeneous systems for many years to come.”

The emergence of heterogeneous systems and next-generation non-volatile storage technologies beyond NAND Flash such as Resistive RAM (ReRAM) means fabric clusters will be key to driving scalable performance, says Dr. Zvonimir Z. Bandić,
Director of Next Generation Platform Technologies at drive maker Western Digital. “We see the public release of the 25xN 100 Gbps RapidIO Specification as a key enabling building block for these coherent, non-volatile storage based systems,” he said.

The technology will be integrated into embedded system-on-chip devices with cores from ARM and MIPS as a way to link different types of computing resources together with low latency using standalone switches from suppliers such as IDT.

“Open standard interconnects are key to ongoing energy efficiency and performance improvements across the entire SoC,” said Phil Bourekas, director of segment marketing, ARM. “The public availability of the 25xN 100 Gbps RapidIO Specification addresses ARM ecosystem developer requirements for coherent network infrastructure SoCs based on the ARMv8-A architecture, providing an available standard for systems at 5G and beyond.”

“IDT has a strategic commitment to RapidIO that spans over a decade and the public release of the 25xN 100 Gbps RapidIO Specification is a key enabling platform for wireless, HPC, cloud and industrial systems our customers are building”, said Ron Jew, General Manager, Wireless Infrastructure Products at IDT. “Through key R&D collaborations with global technology experts, such as CERN openlab, we’re able to ensure that our heterogeneous coherent scale out and storage over 100 Gbps RapidIO fabrics delivers the performance our customers in 5G infrastructure and advanced data analytics are looking for.”


“Heterogeneous computing is becoming increasingly prevalent for our customers’ next generation systems and our MIPS processor technology plays a key role in these systems,” said Jim Nicholas, EVP MIPS Processor IP at Imagination Technologies. “The release of the industry open standard 25xN 100 Gbps RapidIO Specification is a welcome addition to our ecosystem and we’ll look forward to future developments from the Coherent Scale Out Task Group.”



Monday, May 23, 2016

Xilinx expands its 16nm UltraScale+ FPGA roadmap for new data centre technology

By Nick Flaherty www.flaherty.co.uk

Xilinx has added new acceleration technologies for the data centre to its 16nm FinFET-based UltraScale+ product roadmap.

The integrated High-Bandwidth Memory (HBM) is combined with the recently announced Cache Coherent Interconnect for Acceleration technology (CCIX). CCIX is initially driven by a group of seven companies to enable an acceleration framework that works with multiple processor architectures. These acceleration enhanced technologies will enable efficient heterogeneous computing for the most demanding data center workloads. The new products will also be highly leveraged in many other compute intensive applications requiring high memory bandwidth.

CCIX is backed by AMD, ARM, Huawei, IBM, Mellanox and Qualcomm alongside Xilinx to provide a high-performance open acceleration framework to data centers. For the first time in the industry, a single interconnect technology specification will ensure that processors using different instruction set architectures (ISA) can coherently share data with accelerators and enable efficient heterogeneous computing – significantly improving compute efficiency for servers running data center workloads.

Applications such as big data analytics, search, machine learning, NFV, wireless 4G/5G, in-memory database processing, video analytics, and network processing, benefit from acceleration engines that need to move data seamlessly among the various system components. CCIX will allow these components to access and process data irrespective of where it resides, without the need for complex programming environments. This will enable both off-load and bump-in-the-wire inline application acceleration while leveraging existing server ecosystems and form factors, thereby lowering software barriers and improving total cost of ownership (TCO) of accelerated systems.

“AMD strongly supports development of open standards to make heterogeneous computing more pervasive,” said Gerry Talbot, AMD corporate fellow and vice president of I/O and circuit technologies. “By joining with others in the industry to develop new interconnect specifications to accelerate performance, AMD continues its commitment to open, heterogeneous computing.”

“A ‘one size fits all architecture’ approach to data center workloads does not deliver the required performance and efficiency,” said Lakshmi Mandyam, director server systems and ecosystems, ARM. “CCIX enables more optimized solutions by simplifying software development and deployment of applications that benefit from specialized processing and hardware off-load, delivering higher performance and value to data center customers.”

"IBM Power Systems have recently demonstrated a total commitment to openness as a catalyst for industry innovation, creating benefits in cost and performance to clients in a post-Moore's Law era," said Brad McCredie, IBM Fellow and Vice President of POWER Development. "IBM is committed to working with like-minded industry leaders to expand our efforts around open coherency to help meet our clients' growing cognitive needs."

Built on the TSMCs proven CoWoS process, the HBM-enabled FPGAs will improve acceleration capabilities by offering 10X higher memory bandwidth relative to discrete memory channels for the lowest possible latency. The CCIX technology promotes efficient heterogeneous computing by allowing processors with different instruction-set architectures to coherently share data with accelerators such as the HBM-enabled FPGAs.

“Having already delivered 19 billion transistors on a chip at 20nm leveraging our second generation 3D IC technology, we are creating a third generation 3D IC breakthrough for data center acceleration and other compute intensive designs,” said Victor Peng, executive vice president and general manager, Programmable Products at Xilinx. “When combined with next generation CCIX acceleration framework and our software defined SDAccel development environment, this technology will enable a new breed of high-density, flexible platforms for accelerating compute, storage and networking applications.”

“CCIX enables greater performance and connectivity capabilities over existing interconnects, and actually paves the road to the next generation CPU – Accelerator – Network standard interface,” said Gilad Shainer, vice president of Marketing at Mellanox. “With an anticipated broad eco-system support of the CCIX standard, data centers will now be able to optimize their data usage, thereby achieving world-leading applications efficiency and scale.”

“Qualcomm Technologies is excited about the development of a new technology enabling efficient, high-performance architectures in an open, ISA-agnostic platform,” said Vinay Ravuri, vice president of product management, Qualcomm Technologies, Inc. “The data center of the future demands open architectures enabling choice of compute, acceleration and interconnect technologies, and this is a significant step forward in delivering on that goal.”

“CCIX will leverage existing server interconnect infrastructure and deliver higher bandwidth, lower latency, and cache coherent access to shared memory,” said Gaurav Singh, vice president of Architecture at Xilinx. “This will result in a significant improvement in the usability of accelerators and overall performance and efficiency of data center platforms.”
Xilinx is already collaborating with leading hyperscale data center customers to create optimized configurations and products.

Tuesday, March 08, 2016

Linaro launches 64bit ARM-based developer Cloud to go from server to node

By Nick Flaherty www.flaherty.co.uk

IF the sign of a maturing technology is to move from a hetereogeneous development envinrnment to a homogeneous one, then the announcement today by Linaro is significant.

As a collaborative engineering organization developing open source software for the ARM architecture, Linaro's ARMv8 based Developer Cloud brings together silicon, server and software members to provide a cloud-based native ARM development environment to design, develop, port and test server, cloud and IoT applications without substantial upfront hardware investment.

The Developer Cloud is the combination of ARM based silicon vendors’ server hardware platforms, emerging cloud technologies, and many Linaro member driven projects, including server class boot architecture, kernel and virtualization. These projects have been under development since the formation of the Linaro Enterprise Group (LEG) and Linaro has already been enabling key developers via remote access to bare metal ARM servers for the last year.

Having high performance ARM servers able to service the ARM chips in IoT nodes in theory should provide a more optimised and efficient eco-system to improve performance, bring down costs and drive more innovation. 

“Linaro’s reference software platform already gives developers easy access to open source project contributions that advance highly efficient ARM technologies across a variety of markets,” said Jeff Underhill, Director of Server Programs at ARM. “The Linaro Developer Cloud builds on this with a rich new development environment supporting a range of applications but it’s particularly exciting to see the immediate value this will bring to the growing ARM-based server ecosystem.”

The Developer Cloud is based on OpenStack, using both Debian and CentOS as the underlying cloud OS infrastructure. It will use ARM based server platforms from Linaro members AMD, Cavium, Huawei and Qualcomm Technologies and will expand with demand, and as new server platforms come to market. These platforms will include both single socket and dual socket configurations as well as 10/40Gb networking, scalable storage and integrated accelerators that ARM SOC partners are bringing to market.

“Qualcomm Technologies is pleased to be supporting the Linaro Developer Cloud,” says Elsie Wahlig, principal engineer, Datacenter Group, Qualcomm Technologies, Inc. “As the software ecosystem for ARM servers gains momentum, the Developer Cloud presents a new vehicle for friction-less access to standards-based hardware and software to enable accelerated development by our software partners and open source community participants.”

Linaro has established two Developer Cloud facilities, one based in Cambridge, UK and the second based in Austin, Texas. Over time the Developer Cloud will expand through participating Linaro member and member partner data-centers providing cloud regions in China, North America and Europe.

“Linaro works with its members to provide reference open source software to accelerate the development of innovative applications taking advantage of ARM based platforms”, said George Grey, CEO of Linaro. “As the adoption of ARM based servers accelerates and IoT applications rapidly evolve, software developers need access to hardware and easy to use software reference platforms. The Linaro Developer Cloud is designed to broaden the availability of the latest hardware to developers globally, and to enable commercial and private cloud providers to utilize the implementation to accelerate deployment of their own offerings. Linaro will publish the end to end open source code for the implementation of the Developer Cloud”.
Access to the Developer Cloud will be provided via the linaro.cloud web portal. Through the portal, developers can request cloud access and may report bugs and performance issues. The portal will also provide a developer forum to share development and porting knowledge, as well as best practices for ARM servers.

“AMD has been working closely with Linaro and its members to make available a complete software stack for installations like the Linaro Developer Cloud for AMD’s ARM based server processors,” said Suresh Gopalakrishnan, Corporate Vice President, Enterprise Solutions Engineering at AMD. “The Opteron A1100 processor is built with these workloads in mind. We are providing the necessary support so the Opteron A1100 is readily available in the Linaro Developer Cloud and in a range of commercially available systems.”

“Data Center customers are continuing to drive demand for ARMv8 server solutions and the requirements for optimized software and applications are growing dramatically, said Larry Wikelius, Vice President Software Ecosystem and Solutions Group at Cavium. “Cavium has partnered with OVH to deliver a Public Cloud solution based on ThunderX Cavium’s ARMv8 Workload Optimized Processor. We expect this would accelerate development of optimized software on ARMv8 servers and enhance services offered by the Cloud providers and enterprises.”

Linaro has over 200 engineers working on consolidating and optimizing open source software for the ARM architecture, including developer tools, the Linux kernel, ARM power management, and other software infrastructure. Linaro is distribution neutral: it wants to provide the best software foundations to everyone by working upstream, and to reduce non-differentiating and costly low level fragmentation. The effectiveness of the Linaro approach has been demonstrated by Linaro’s growing membership, and by Linaro consistently being listed as one of the top five company contributors, worldwide, to Linux kernels since 3.10.

To ensure commercial quality software, Linaro’s work includes comprehensive test and validation on member hardware platforms. The full scope of Linaro engineering work is open to all online at http://www.linaro.org and http://www.96Boards.org.



Wednesday, February 03, 2016

Membrane lasers embedded in CMOS for high speed communications

By Nick Flaherty www.flaherty.co.uk

A new type of ultra-thin semiconductor laser under development at The University of Texas at Arlington can be embedded into mainstream same silicon substrates to provide increased capacity and energy efficiency.
Weidong Zhou, UTA electrical engineering professor, is working to develop a new ultrathin semiconductor laser that will increase the chip's speed and capacity. A three year grant from the US Army Research Office will develop a membrane laser less than one micron thick that is compatible with planar CMOS platforms. The key innovation is the integration of certain compound semiconductor material with a silicon photonic crystal cavity, which allows a laser to be built directly on a silicon chip next to other electrical components. This leads to higher speed and higher efficiency.
The first application of Zhou’s laser is in data centres, and the group is pursuing various membrane laser architectures for extreme energy efficient computing and communication systems. “We are looking for devices and components to be integrated on a chip,” said Zhou. “As we address electrical injection, integration with other devices on the chip and increased power capabilities, we can begin to apply this technology to products in the medical field or in the consumer arena. These applications could include portable electronics, sensing and imaging equipment, bio applications and wearable electronics.” 
“Big companies like IBM and Intel are using this technology for high-performance computing centres,” Zhou said. “The big push now is for the next big thing: smaller, faster, and less and less power consumption.”

Quantum dot photonics array open up 80THz of new bandwidth for the data centre

By Nick Flaherty www.swinnovation.co.uk

Researchers in the UK and Japan unlocked 80THz of fibre-optic bandwidth that will enable future exascale data centres and transform 5G networks.
The collaboration between the University of Bristol’s Department of Electrical and Electronic Engineering, Keio University and numerous Japanese industrial partners, have designed, developed and prototyped an all-optical router that can unlock 80 THz of bandwidth across a newly defined frequency band named T-Band (thousand band) and O-Band (original band). The adjacent bands span from 1.0 μm (300 THz) to 1.36 μm (220 THz) and are able to support 1600 channels at 50 GHz spacing.
t-band-article
The technology fabricated and tested is based on cascaded arrayed waveguide gratings (AWGs) and is designed to potentially construct a 1600 x 1600 wavelength router that can guide data through. This uses quantum dot chips for the light sources which were originally developed by NICT in Japan.
“The technology and system proposed and prototyped will unlock the new frequency band and networks to support future exascale data centres, ‘zero-latency’ tactile optical internet, internet of everything, smart cities, fog computing and big data infrastructure among others," said Dr Georgios Zervas, Senior Lecturer in Optical and High Performance Networks in the Department of Electrical and Electronic Engineering.
A single passive optical system can interconnect over one million end points such as, broadband home users, IoT devices, data centre servers, while offering at least ten Gb/s per end point. It is also future proof since it’s transparent to any communication signal and it can also potentially consume zero power due to its passive nature says Hiroyuki Tsuda, Professor of Faculty of Science and Technology at Keio University. “The enabling technologies for the new frequency band are the quantum dot based optical devices and the silica planar lightwave circuits designed for the new band," he added.