HPC Budget Allocation Map: Industry Averages
Intersect360 Research surveyed the High Performance Computing user community to complete its seventh Site
Budget Allocation Map, a look at how HPC sites divide and spend their budgets. We surveyed users on their
spending in seven top-level categories: hardware, software, facilities, staffing, services, cloud/utility/outsourcing
computing, and other. Each category was further divided into constituent subcategories, resulting in 25 unique
items included in the analysis.
This report provides the average budget distribution for the responding sites within each category and presents
an entire view of HPC budget distribution and IT product spending (excluding facilities and staffing).
Highlights from this study include:
- Hardware represents the largest overall budget item, accounting for 45% of the total HPC budget in 2014. Servers, followed by storage, continue to lead spending within the hardware segment. When asked for future budget trends, hardware is the primary driving force behind budget change. About 50% of sites that provided qualitative input on budget trends reported a change in hardware spending as their reason. About two-thirds expect hardware spending to increase, and one-third expect a decrease.
- Staffing continues to be the second-largest overall expense, accounting for 22% of sites. Average spending on staffing has stabilized over the last three surveys, after a few years of declining share. System management and operations, followed by application programmers, and user support and services, are the top three categories within staffing and account for 71% of the staffing budget.
- Software as a share of the overall HPC budget accounts for 13% of that budget. System software and software tools represent the largest share of expenditures within software, accounting for a combined share of 47% of the HPC software budget.
- Cloud/Utility/Outsourcing computing is still a very small percentage of overall HPC expenditures, with about 3% share. We have seen fluctuations year-to-year in spending on cloud computing but noconsistent or significant movement either upwards or downwards.
- Budgets are expected to increase by 1% or more for 59% of all respondents, and 43% expect budgets to increase by more than 5%. Commercial sites showed the most optimism with 78% expecting budgetsto increase by at least 1% and 66% expecting budgets to increase by at least 5%. About 19% ofacademic sites are expecting a decline in their budgets, which is a higher percentage than either commercial or government sites.
- HPC system elements
- Systems, clusters
- HPC clusters
- Server technologies
- Processor elements
- System processors
- Storage elements
- Storage systems
- Network-attached storage (NAS)
- Storage area networks (SAN)
- Direct-attached storage (DAS)
- Cloud storage
- Storage components
- Storage software
- Software elements
- Operating systems
- Cluster management
- Job scheduling
- Load balancing
- Cloud management
- Other middleware
- Developer tools
- Programming models (i.e., MPI, OpenMP)
- Compilers and languages
- Developer environments (IDEs)
- Accelerator-specific optimization tools (i.e., CUDA, OpenCL)
- Parallel programming tools
- Other development and optimization libraries and tools
- Application software
- Independent software vendor (ISV) or third-party (purchased/licensed) applications
- In-house developed applications
- Facilities-level technologies
- Programming services
- Actionable Market Intelligence for High Performance Computing
- Maintenance services
- Cloud computing, grid computing, utility computing
- Public cloud technologies
- Private cloud technologies
- XaaS models, such as SaaS, IaaS, PaaS, and HPC as a Service (“HPCaaS”)
- Cloud bursting