PUBLISHER: Mind Commerce | PRODUCT CODE: 1236145
PUBLISHER: Mind Commerce | PRODUCT CODE: 1236145
The High-Performance Computing market includes computation solutions provided either by supercomputers or via parallel processing techniques such as leveraging clusters of computers to aggregate computing power.
HPC is well-suited for applications that require high performance data computation and analysis such as high frequency trading, autonomous vehicles, genomics-based personalized medicine, computer-aided design, deep learning, and more. Specific examples include computational fluid dynamics, simulation, modeling, and seismic tomography.
This Mind Commerce report evaluates the HPC market including companies, solutions, use cases, and applications. Analysis includes HPC by organizational size, software and system type, server type, price band, and industry verticals.
The report also assesses the market for integration of various artificial intelligence technologies in HPC. It also evaluates the exascale-level HPC market including analysis by component, hardware type, service type, and industry vertical.
High Performance Computing (HPC) may be provided via a supercomputer or via parallel processing techniques such as leveraging clusters of computers to aggregate computing power. HPC is well-suited for applications that require high performance data computation such as certain financial services, simulations, and various R&D initiatives.
The market is currently dominated on the demand side by large corporations, universities, and government institutions by way of capabilities that are often used to solve very specific problems for large institutions. Examples include financial services organizations, government R&D facilities, universities research, etc.
However, the cloud-computing based "as a Service" model allows HPC market offerings to be extended via HPC-as-a-Service (HPCaaS) to a much wider range of industry verticals and companies, thereby providing computational services to solve a much broader array of problems. Industry use cases are increasingly emerging that benefit from HPC-level computing, many of which benefit from split processing between localized devices/platforms and HPCaaS.
In fact, HPCaaS is poised to become much more commonly available, partially due to new on-demand supercomputer service offerings, and in part as a result of emerging AI-based tools for engineers. Accordingly, up to 54% of revenue will be directly attributable to the cloud-based business model via HPCaaS, which makes High-Performance Computing solutions available to a much wider range of industry verticals and companies, thereby providing computational services to solve a much broader array of problems.
We conducted interviews with major players in the market as well as smaller, lesser known companies that are believed to be influential in terms of innovative solutions that are likely to drive adoption and usage of both cluster-based HPC and supercomputing. In an effort to identify growth opportunities for the HPC market, we investigated market gaps including unserved and underserved markets and submarkets. We uncovered a market situation in which HPC currently suffers from an accessibility problem as well as inefficiencies and supercomputer skill gaps.
Stated differently, the market for HPC as a Service (e.g. access to high-performance computing services) currently suffers from problems related to the utilization, scheduling, and set-up time to run jobs on a supercomputer. We identified start-ups and small companies working to solve these problems.
One of the challenge areas identified is low utilization but (ironically) also high wait times for most supercomputers. Scheduling can be a challenge in terms of workload time estimation. About 29% of jobs are computationally heavy and 41% of jobs cannot be defined very well in terms of how long jobs will take (within a 3-minute window at best). In many instances, users request substantive resources and don't actually use computing time.
In addition to the scheduling challenge, we also identified a company focused on solving additional problems such as computational planning and engineering. We spoke with the principal of a little-known company called Microsurgeonbot, Inc. (doing business as MSB.ai), which is developing a tool for setting up computing jobs for supercomputers.
The company is working to solve major obstacles in accessibility and usability for HPC resources. The company focuses on solving a very important problem in HPC: Supercomputer job set-up and skills gap. Their solution known as `Guru` is poised to make supercomputing much more accessible, especially to engineers in small to medium-sized businesses that do not have the same resources or expertise as large corporate entities.
|
|