Chirag Dekate, Ph.D.

Devising compute strategies for AI applications can be challenging as it involves navigating complex design considerations. I&O leaders can use this research to deliver highly efficient infrastructures for compute-intensive machine learning and deep neural network-based applications.

I&O leaders selecting infrastructure for AI workloads involving machine learning and deep learning must comprehend the unique requirements of these emerging workloads. We analyze the impact of AI workloads on data infrastructure and outline best practices for storage selection and implementation.

Many organizations are struggling to architect infrastructures for machine learning and deep learning workloads that are significantly different than traditional enterprise applications. I&O leaders can use three best practices to deliver high-performance, scalable machine learning infrastructures.

Advances in the maturity of hyperconverged and OS containers are among the key 2018 trends affecting compute infrastructure on-premises and in the cloud. Here, we describe the maturity and benefits of 28 compute innovations to enable infrastructure and operations leaders to decide which to deploy.

In the server market, a shift in deployment to a mix of on-premises, hosted, public cloud and edge is disrupting what's bought and who buys it. Infrastructure and operations leaders should use this research as a detailed overview of server vendors by technology category and application use case.

The market for AI semiconductors is rapidly evolving, and many vendors are competing for software developers' attention. This report analyzes the Intel Artificial Intelligence Products Group's strategy and advises technology product management leaders on how to work with or compete against Intel.

Next-generation application architectures in emerging areas such as AI, microservices and IoT are reshaping computing ecosystems. I&O leaders must balance agility with risk while delivering effective compute and infrastructure strategies to support disruptive business models.

Google has responded to changes in its core advertising businesses and invested in its enterprise-focused offerings, while delivering valuable new services, such as deep machine learning. CIOs and IT leaders should expect Google to continue on this path.

Public cloud is becoming a preferred platform for building and deploying AI applications. I&O leaders can use this research to identify key requirements for AI applications involving deep neural networks and select the right cloud-based infrastructure delivery platform.

Successful AI initiatives require the adoption of new technologies, processes and governance models. I&O leaders should use the best practices outlined in this research to devise effective AI strategies.

Intel has pivoted its focus away from the PC and toward all facets of the data center as a way to reinvigorate growth. Enterprise CIOs and IT leaders should leverage Intel's technology leadership and closely watch its execution strategy in data-centric areas, such as AI, IoT and autonomous vehicles.

AI technologies, especially deep learning, are poised to diffuse rapidly through cloud services, APIs and the Internet of Things, driven by growing consumer use of virtual assistants in smartphones and smart homes. CIOs should start now to lay their organization's AI foundation.

The compute infrastructure market is undergoing major churn; traditional data center vendors face existential threats from off-premises computing, new Asian competitors and margin pressures. This research helps I&O leaders faced with how and where to address future workload requirements.

Organizations need to carefully assess the trade-offs involved before deciding whether to run AI workloads on-premises or in the cloud. I&O leaders should use this decision framework to determine the right deployment model for machine learning and deep learning use cases.

Disruptive server providers have changed the server market over the past 10 years. Technology business unit leaders must leverage effective disruptive provider approaches if they hope to achieve optimal business results.

Building AI infrastructure strategies can be challenging, especially for compute-intensive deep learning workloads. I&O leaders must choose the right accelerators for devising effective deep learning compute infrastructure strategies.

Advances in HPC, data analytics and AI, plus growing importance of edge computing and HCIS, are among the key 2017 compute infrastructure trends. This Hype Cycle evaluates business impact, adoption rate and maturity level of 26 technologies to help I&O leaders decide where and when to invest.

Many organizations are struggling to architect infrastructures for machine learning and deep learning workloads that are significantly different than traditional enterprise applications. I&O leaders can use three best practices to deliver high-performance, scalable machine learning infrastructures.

Storage products supporting Nonvolatile Memory Express and NVMe over Fabrics deliver performance improvements for I&O leaders modernizing their infrastructures. This research provides action-oriented insight on where, when and whether I&O leaders should take advantage of these emerging technologies.

As enterprises grapple with the right mix of on-premises, off-premises and native cloud, choosing a cloud infrastructure vendor becomes more critical. I&O leaders should look to vendors like those in this research for innovative ways to support varying workload deployments across delivery models.

Compute infrastructure is rapidly evolving from hardware-centric silos to software-driven ecosystems where application agility drives infrastructure architectures. I&O leaders must plan now to address these trends impacting computing to evolve and modernize their infrastructure for digital business.

HPE's acquisition of SGI creates new market dynamics that will influence acquisition of high-end x86 server infrastructure through 2017 and beyond. Infrastructure and operations leaders focused on server procurement and life cycle decisions can use this research to reassess their current plans.

The market for compute platforms and software-defined anything is fragmenting between legacy systems and emerging technologies, creating the need for a bimodal IT approach. This IT Market Clock will guide I&O leaders in budgeting for hardware churn and an increase in compute platform investment.

I&O leaders spearheading data center modernization seek to balance performance, agility and value from high-performance computing ecosystems. Use these Gartner best practices to improve HPC deployments and maximize business value.

1. Understand the Resource Requirements of Critical Workloads
2. Identify the Right Technologies to Maximize Value
3. Explore Alternative Models to Deliver HPC Ecosystem Agility

"From telescopes to microscopes, from vacuums to hyperbaric chambers, from sonar waves to laser beams, scientists have perpetually strived to apply technology and invention to new frontiers of scientific advancement. Along the way, they have used abacuses, slide rules, calculators, computers, and - today - supercomputers, to crunch the increasingly complicated calculations used to understand and predict natural phenomena. In the course of practicality, science begets engineering, and… Show more

"From telescopes to microscopes, from vacuums to hyperbaric chambers, from sonar waves to laser beams, scientists have perpetually strived to apply technology and invention to new frontiers of scientific advancement. Along the way, they have used abacuses, slide rules, calculators, computers, and - today - supercomputers, to crunch the increasingly complicated calculations used to understand and predict natural phenomena. In the course of practicality, science begets engineering, and engineering begets innovation. Science, and therefore supercomputing, finds its way into industry around the world, affecting our lives in ways that are simple (cleaner clothes, faster speedboats, and more spin on a seven-iron) and profound (fewer drug interactions, safer cars, and new energy sources). Altogether, industry consumes more than half of all high performance computing usage worldwide. This book tells that story." Show less

Over the past few years, Asian supercomputing capability has catapulted Japan and China to the forefront of leadership supercomputing. Critical investments by DOE have enabled the United States to develop leadership-class supercomputing

This IDC update summarizes the presentations on high-performance computing (HPC) and data-intensive computing ("Big Data") that took place at the 48th HPC User Forum meeting.

This IDC update summarizes the presentations on the economic value of HPC in science and industry that took place at the 48th HPC User Forum meeting.

Creating the next generation of power-efficient parallel computers requires a rethink of the mechanisms and methodology for building parallel applications. Energy constraints have pushed us into a regime where parallelism will be ubiquitous rather than limited to highly specialized high-end supercomputers. New execution models are required to span all scales, from desktop to supercomputer.

Productivity is an emerging measure of merit for high-performance computing. While pervasive in application, conventional metrics such as flops fail to reflect the complex interrelationships of diverse factors that determine the overall effectiveness of the use of a computing system. As a consequence, comparative analysis of design and procurement decisions based on such parameters is insufficient to deliver highly reliable conclusions and often demands detailed benchmarking to augment the more… Show more

Productivity is an emerging measure of merit for high-performance computing. While pervasive in application, conventional metrics such as flops fail to reflect the complex interrelationships of diverse factors that determine the overall effectiveness of the use of a computing system. As a consequence, comparative analysis of design and procurement decisions based on such parameters is insufficient to deliver highly reliable conclusions and often demands detailed benchmarking to augment the more broad system specifications. Even these assessment methodologies tend to exclude important usage factors such as programmability, software portability and cost. In recent years, the HPC community has been seeking more advanced means of assessing the overall value of high-end computing systems. One approach has been to extend the suite of benchmarks typically employed for comparative examination to exercise more aspects of system operational behavior. Another strategy is to devise a richer metric for evaluation that more accurately reflects the relationship of a system class to the demands of the real-world user workflow. One such measure of quality of computing is ‘productivity’, a parameter that is sensitive to a wide range of factors that describe the usage experience and effectiveness of a computational workflow. Beyond flops count or equivalent metrics, productivity reflects elements of programmability, availability, system and usage cost and the utility of the results achieved, which may be time critical. In contrast to a single measure, productivity is a class of quantifiable predictors that may be adjusted to reveal best understanding of system merit and sensitivity to configuration choices. Show less

The storing of data and configuration files related to scientific experiments is vital if those experiments are to remain reproducible, or if the data is to be shared easily. The prescence of historical (observed) data is also important in order to assist in model evaluation and development. This paper describes the design and implementation process for a data archive, which was required for a coastal modelling project.
The construction of the archive is described in detail, from its design… Show more

The storing of data and configuration files related to scientific experiments is vital if those experiments are to remain reproducible, or if the data is to be shared easily. The prescence of historical (observed) data is also important in order to assist in model evaluation and development. This paper describes the design and implementation process for a data archive, which was required for a coastal modelling project.
The construction of the archive is described in detail, from its design through to deployment and testing. As we will show, the archive has been designed to tolerate failures in its communications with external services, and also to ensure that no information is lost if the archive itself fails, i.e. upon restarting, the archive will still be in exactly the same state. Show less