Pages

Friday, December 22, 2017

IBM Q Network – moving Quantum Computing from science to problem solver

By Rich Ptak


Image Courtesy of IBM, Inc.
On December 14th, IBM announced the IBM Q Network, a worldwide collaborative effort to create a connected community of quantum computing involved individuals and organizations. Three relationship options are available to meet the varying interests and needs of potential members that range from start-ups to F500 enterprises, also including universities and research institutions with provisions for interested individuals in engineering, science and business.
The globe spanning cooperative network is linked by the IBM Cloud and enabled by IBM Q[1] to advance and accelerate progress in the race to realize Quantum Advantage (QA). QA occurs when quantum computing can deliver commercial value with demonstrably better, quicker and more accurate solutions than classical computing for substantive, real-life problems. This is an overview and our comments on the IBM Q Network announcement.



IBM Q Experience – building the “Market” for Quantum 

For decades, quantum computing existed primarily as an esoteric exercise in theories and physics. Activities focused on developing quantum science theories were restricted to universities, research institutions and theoretical scientists. Quantum science was a necessary precursor to quantum computing as a technology that could be applied to problem-solving.
More recently, a combination of events (potential exhaustion of Moore’s Law, escalating problem complexity) drove investigation into alternatives to classical computing techniques. Some vendors, including IBM, made the decision to pursue commercially viable quantum computing. The race to develop quantum technology began.
In May 2016, IBM made basic quantum computers available in the IBM Cloud. This, along with other IBM contributions (discussed here[2]), contributed to accelerate the evolution of quantum computing science, including IBM’s proposal of Quantum Volume as a more functional metric of computational value than qubits.
More significant was when IBM became the first vendor to provide widespread, free, public access to quantum computers via IBM Q Experience[3]. This enabled a larger, diverse audience to acquire knowledge about and experience with quantum science methodologies, modeling, etc. Today, it provides free access to 5- and 16-qubit quantum computing prototypes residing in the IBM Cloud.
Additional available services, support and tools include QISKit[4]; an open source software development kit with examples tailored for specific types of problems, e.g. chemical structure modeling and simulation. These helped to advance quantum computing away from what was primarily a scientific exercise to a publicly accessible ‘sand-box’ where interested individuals, engineers, businesses and other groups could learn and experiment.
But, the overall quantum “market”, while intensely competitive, remained dispersed, unorganized and unfocused. It was difficult to dispassionately assess progress, compare machines or get a coherent sense of the state of quantum in its progress from science to technology.
With the introduction of IBM Q Network, IBM provides an organizing model that allows full flexibility for creativity and innovation while helping to focus collaborative efforts undertaken by a global community to achieve Quantum Advantage.


The IBM Q Network

IBM Q Network will be a global community of individuals and organizations acting as independent, but collaborating units focused on achieving common goals. It is designed to operate as an ecosystem of loosely linked, coordinated organizations without the constraint and difficulty of imposed overarching management. This is very similar (consciously, or not) to an organizational model first implemented by the early RAND organization, a uniquely successful incubator of innovation.
Relationships are grouped in three categories (Hub, Partners and Members) with specific relationship, responsibility and activity interests. See Figure 1 below. Hubs (IBM Research, Oak Ridge National Laboratory, Keio University, University of Oxford, University of Melbourne) act as regional centers for education, research, development and commercialization of quantum computing. Partners (SAMSUNG, JPMorgan CHASE & Co., DAIMLER) focus on a specific industry or academic field as pioneers in applying quantum computing. Members (Barclays, Honda, Materials Magic (Hitachi), NAGASE) build their own general knowledge of quantum computing while developing a strategy to become quantum ready.
All participants have unique access to the latest IBM Q quantum systems (now on a 20-qubit device, followed shortly by a 50-qubit device) via the IBM Cloud.
                            Figure 1 IBM Q Network - Organizational Collaborators         Image Courtesy of IBM, Inc.
Specific levels of support, involvement with IBM on projects vary with category. For example, Hubs provide access to IBM Q systems, tech support, educational and training resources, community workshops and events.
Partners have direct access to IBM Q system and work with IBM on joint training, development and other projects.
Members can access IBM Q and IBM Q Network community resources through IBM Research. Additional details about resources committed, project activities, funding, etc. are worked out with IBM. These may vary based on agreements between participants and IBM. More details available at the IBM Q Network website[5].      
One more point to be emphasized, the IBM Q Network is also targeted at and provides support for individuals wanting to share ideas, research plans, projects and proposals. It is to be an open network for creative innovation, collaboration and communication.


What does this all mean?

It’s our opinion that IBM Q Network marks a significant advance for quantum computing in general. To date, the quantum space has been chaotic with few standards or consistent benchmarks against which to evaluate competing offerings and claims.
The network leverages existing communications technology to create a global “commons’ for sharing knowledge aimed at putting the science of quantum to work. The resulting global knowledge community resembles Republic of Letters[6], the Europe-wide shared community of thinkers, scientists and innovators which led to an explosion in scientific development and entrepreneurial activities from the 1600s through 1800s.  
IBM’s move to build a global community, loosely coupled but, with a clear set of goals along with a definite, expanding support infrastructure will be a significant advance on the road to fully commercialized quantum computing. It provides a much needed structure for extensive collaboration with minimal constraints dedicated to advancing quantum technology.
IBM has invited all interested parties, including competitor vendors to participate; subject only to an interest in advancing the commercialization of quantum computing and use of the IBM Q resources. Undoubtedly, there will be contractual issues over intellectual property rights, information sharing, ownership, financial arrangements, investment, etc. These will be worked out individually with participants. IBM understands the issues; they are willing to be very flexible in setting terms and conditions. 
Finally, even at its birth, IBM Q Network carries impressive weight. IBM provided some Day 1 statistics about the IBM Q Network. These include: the IBM Q Experience provides access to the first Quantum Computers on the cloud. It has over 60,000 users worldwide located on all 7 continents (including Antarctica). There are over 35 external papers already published. Day 1, the network includes F500 companies, research institutions in Europe, US and Asia, as well over 150 colleges and universities.
We encourage quantum-interested individuals and enterprises to examine the IBM Q Network. The network itself is a major effort to provide a working environment that will attract a diverse community of thought-leaders to advance quantum computing to a productive technology. It is a major gamble by IBM. But we think it will prove to be a decisive one on the road to a successful commercial quantum computing infrastructure. Look for yourself; you’ll see the benefit.





[4] IBM Quantum Information Software, go to link: https://www.qiskit.org/
[6] Mokyr, Joel 2016. A Culture of Growth: The Origins of the Modern Economy, Princeton, N.J.: Princeton University Press

Monday, December 11, 2017

IBM POWER9 breaks barriers that hamper AI solutions

By Bill Moran and Rich Ptak



On December 5, IBM announced POWER9, its newest Power System. The POWER9 title might be taken to imply it is just another POWER8 iteration with a performance boost and a few new features thrown in. Not so. POWER9 is a significant generational advance, providing much more than a minor turn-of-the-crank. These next-generation Power Systems embed leading edge new technologies, such as PCI-Express 4.0, next-gen NVIDIA NVLink 2.0 and OpenCAPI more about these later. The new server, AC922 is the base platform for the CORAL collaboration, the world’s most powerful supercomputer. 
With this announcement, IBM marks a major change-of-direction as it targets compute intensive the super-computing and AI workloads used for modeling, research, credit risk analysis, etc.  Workloads requiring LOTs of memory, extremely high processing speeds and analyze vast amounts of data. We comment on this and its implications.

POWER9 enhancements

First, a description of the improvements over prior iterations. Skipping the “speeds and feeds”, here are a few important points.
·         POWER9 chips are 14nm technology; a significant advance over last generation’s 22nm. IBM no longer controls a chip foundry having sold it to Global Foundry. However, the Global Foundry – IBM alliance is clearly working effectively and delivering products in a timely manner
·         POWER9 architectural changes yield many improvements. These include a new implementation of OpenCAPI 2.0, that delivers a major improvement in I/O capacity as it speeds bandwidth by a factor of 4 over CAPI[1] in POWER8. Implementation of PCI-Express 4.0 and next-gen NVIDIA NVLink 2.0 means that data flows in and out of the system more quickly. Complex data analysis, simulations and model building/evaluations complete faster. Programming is simplified.
·         POWER9 enhances links between system CPUs and the GPUs. Experience has proven that pairing GPU devices with the CPU can yield dramatic operational improvements. POWER9’s new links increase bandwidth speeds by a factor of 7 - 10 times to benefit data manipulation and analysis.
Connections between the GPUs and system memory is simplified and improved. Thus, AI models run faster, programming is simpler, and this permits quicker creation and evaluation of more complex and larger models for AI, data analytics, research etc. Learning times are also dramatically reduced.
These features and new architecture mean the POWER9 is a strong competitor as it delivers performance improvements that are much needed in AI and supercomputing market segments.

A new Server

But, chip level and design specification improvement tend to be of limited interest to many potential customers. They tend to evaluate a new chip or processor in the context of the product they will purchase. They want to know how they or their projects benefit from the AC922 POWER9 processor-based server.
Detailed specs for the AC922 server appear in IBM’s material.  But, IBM also provided some benchmark runs comparing AC922 to an Intel X86 server. Two AI workloads, Caffe and Chainer[2]  were run. For both workloads, AC922 out-performed the X86 system by approximately 3.7 times. The X86 system is a standard environment. We expect Intel will be enhancing x86 with AI capabilities at some point.
We like the benchmarks that IBM ran. They effectively demonstrate the impact of system improvements in actual application. Paper and pencil comparisons are fine; but nothing equals the actual performance a system delivers with a real workload. The initial air-cooled server will be followed in 2018 by a faster, water-cooled version. The air-cooled servers have a maximum of 4 GPUs; the follow-on water versions allow up to 6 GPUs.

Supercomputer Heaven

CORAL is a supercomputer that is being built for the US DOE with various Oak Ridge, Argonne and Livermore labs. CORAL will be the most powerful supercomputer in the world when deployed in 2018. It is expected to deliver 10X the power of Titan, today’s supercomputer leader. Very impressively, the building block for Coral is a standard AC922, now available for purchase. This provides normal customers the ability (if not the resources) to build their own version of a CORAL-type supercomputer around multiple AC922s. We believe many customers, e.g. weather bureaus, modeling researchers, etc., will be interested in constructing such systems. These deployments will verify AC922’s increased operational and programming simplicity, versatility, robustness and scalability.

IBM’s new direction 

IBM has changed the direction of its Power System marketing. In the past, Power Systems were promoted as a general-purpose Linux server in direct competition with Intel servers. Intel dominated the distributed server market (albeit with Windows) for decades. Windows-based systems would have to convert from Windows to Linux to use Power. As such conversions are generally viewed as risky, customers were far more likely to just continue to upgrade to the latest version of Intel. Even customers installing Linux were more likely to do so on an Intel platform, supplied by HP or Dell. Thus, IBM Power Systems, despite significant advantages in processing performance, capacity and I/O handling faced powerful resistance to change which worked against achieving significant market penetration.
Now, with Power9, IBM sees in the new area of AI an opportunity that plays directly to their architectural and performance advantages. Power9 systems were designed to deliver maximum performance with AI workloads and models. They will still compete with Intel, but on a more level playing field in a rapidly growing and diverse market. Both companies will have to compete with very attractive Cloud offerings. IBM believes there exists sufficient demand for on-premise computing to support a profitable business.  Although many, if not most, Cloud servers are x86 based, IBM believes they can deliver a sufficient performance edge to justify keeping AI projects on-premise.  Initial benchmarks suggest that they may be right, although maintaining that edge will remain a challenge.

Summary

We think that IBM has delivered a powerful new answer for anyone searching for a production AI platform. It has the right combination of hardware and software technology to succeed. It has other strengths including Open Power foundation support, enhanced CAPI and GPU interfaces. This support has been critical in the creation of CORAL, as IBM acknowledges. IBM is covering key basics very well.
Finally, a significant messaging advantage they have neglected to mention is the powerful boost that this new architecture and system provides to Watson.  The Watson Marketing group is rightly and understandably focused on marketing segment specific benefits and features.
IBM Watson lays claim to having the best AI solution system in the marketplace. Today, competition in the AI platform space is rapidly growing. Vendors large and small, much x86-based, are effectively competing against IBM. It appears to us that there exists a powerful message in how POWER9 System’s bespoke (for AI) IBM infrastructure meshes with, enables and drives IBM’s showcase AI application. 
The POWER9 architecture represents a significant advance in its offering of AI-specific features, capabilities and performance enhancements. Combined with a solid existing ecosystem, it should increase the market penetration of Power Systems.


[1] CAPI itself was a very significant improvement, see With Redis Labs, CAPI goes Mainstream, Big Time! at https://www.ptakassociates.com/content/
[2] Caffe and Chainer are both open source frameworks. See https://chainer.org/ and http://caffe.berkeleyvision.org/ for more information.