EXECUTIVE SUMMARY
Intersect360 Research surveyed the High Performance Computing (HPC) user community on their software usage for HPC. The complete study analyzes users’ current operating systems, middleware, developer tools, and primary applications across 11 different application domains, including:
- the overall mix of in-house, open-source, and commercial software
- the mix of public cloud versus on-premises computing, by application
- scalability and performance on various processing elements, by application domain
- the overall adoption of machine learning.
Our goal in this analysis is to provide an in-depth analysis of how HPC software trends. Intersect360 Research surveyed members of the HPC user community worldwide across industry, government, and academia. The survey was conducted in the fourth quarter of 2021 and received responses from 238 qualified sites, yielding tens of thousands of data points for analysis with respect to HPC software.
Intersect360 Research defines HPC as the use of servers, clusters, and supercomputers—plus associated software, tools, components, storage, and services—for scientific, engineering, or analytical tasks that are particularly intensive in computation, memory usage, or data management. Intersect360 Research reports available in this HPC Technology Survey report series include the following segmentations:
- Operating Systems and Middleware: including operating systems, HPC job management tools, workflow and workload management tools, and software for cluster management, monitoring, and provisioning
- Compilers and Developer Tools: including languages, compilers, and programming environments for HPC software development
- AI Adoption: including machine learning frameworks and current levels of organizational AI engagement
- Primary Application Software: 11 distinct reports covering applications used, typical and maximum job scalability, affinity to cloud computing, and impressions of performance by processor type, for 11 application domains, as follows:
- Primary Application Software—Biosciences: including pharmaceuticals, biochemistry, genomics, agricultural engineering, cryo-EM, etc.
- Primary Application Software—Business Analytics
- Primary Application Software—Chemistry
- Primary Application Software—Computational Fluid Dynamics (CFD)
- Primary Application Software—Engineering: including computer-aided engineering (CAE), finite element analysis (FEA), noise-vibration-harshness testing (NVH), materials science, etc.
- Primary Application Software—Finance: including banking, trading, pricing, risk management and value-at-risk (VaR), econometric modeling, insurance, etc.
- Primary Application Software—Geophysics: including seismic modeling, reservoir simulation, geology, etc.
- Primary Application Software—Physics: including pharmaceuticals, biochemistry, genomics, agricultural engineering, cryo-EM, etc.
- Primary Application Software—Scientific Research: including general fields of scientific inquiry not covered elsewhere
- Primary Application Software—Visualization
- Primary Application Software—Weather and Climate Modeling: including earth science and ocean modeling
This report provides a detailed examination of primary application software in Scientific Research. Scientific Research is a diverse catch-all category of general research applications that did not fit into other domains. 138 respondents provided data on applications in the Scientific Research category, the most commonly selected domain in the total survey. Due to an error in the survey mechanism, only 93 of these 138 were able to give additional detailed information about their applications. This resulted in 73 unique application mentions, indicating a high degree of specificity tying researchers to individual applications. Additionally, we received five write-ins of additional applications; most of which were mentioned by only one respondent each. Intersect360 Research clients should carefully heed the data and insights in this report in planning their future developments for all HPC technologies and services.
TABLE OF CONTENTS 4
INTRODUCTION 6
Methodology 6
What Is HPC? 8
PRIMARY APPLICATION SOFTWARE: SCIENTIFIC RESEARCH 9
Overview: Primary Applications 9
Figure 1: Percent of Respondents Reporting Primary Applications in Each Domain 9
Figure 2: Proportion of Commercial, In-House, and Open-Source Software by Sector 10
Figure 3: Proportion of Commercial, In-House, and Open-Source Software Across Survey Population, 2010 vs. Today 11
Primary Applications: Scientific Research 12
Table 1: Primary Applications: Scientific Research 12
Job Scalability: Scientific Research 12
Figure 4: Typical and Maximum Job Sizes: Scientific Research 13
Table 2: Typical Job Scalability by Application 14
Table 3: Maximum Job Scalability by Application 14
Cloud Affinity: Scientific Research 15
Figure 5: Cloud Affinity: Scientific Research 15
Table 4: Cloud Affinity by Application 16
Processing Architecture Affinity: Scientific Research 16
Figure 6: Processor Architecture Affinity: Scientific Research 17
CONCLUSIONS 18
APPENDIX A: SURVEY DEMOGRAPHICS 19
Figure A1: Economic Sector of Respondent Organization 19
Figure A2: Number of Employees for Respondent Site and Organization 19
Figure A3: Annual HPC Budget for Respondent Site and Organization 20
Figure A4: Role of Survey Respondent in Organization 20
Figure A5: HPC Software Purchasing Responsibility of Respondent 21
APPENDIX B: SCIENTIFIC RESEARCH: SELECTED QUALITATIVE COMMENTS 22