Future Computing
Technologies

Meeting challenges with
new computational capabilities

Our computing research is enabled, in part, by the Constance computing cluster. The system is a workhorse for parallel applications, including those in molecular dynamics, climate, and fluid flow calculations, or for high-throughput computing uses such as in high-energy physics or machine learning.

Andrea Starr | Pacific Northwest National Laboratory

At Pacific Northwest National Laboratory (PNNL), future computing technologies span fundamental computer science research areas across hardware and software stacks. Research in this area has the potential to transform AI, quantum computing, HPC, edge computing, and distributed computing, and it is leading to emerging computing paradigms, such as quantum, analog, and neuromorphic computing.

What are future computing technologies?

Future computing technologies are planned or ongoing innovations in the field of advanced computing. They stem from research in high-performance scientific computing, quantum computing and artificial intelligence (AI), heterogeneous computing (systems utilizing multiple processor types), and advanced computer systems.

The following are a few examples:

AI-specialized computers
quantum computers
near-memory computation
neuromorphic computing technologies
optical computing technologies
zettascale computing technologies
analog computing technologies
brain/computer interfaces
advanced data storage systems
advanced networks from chip to system level.

Future computing technologies will achieve greater speed, energy efficiency, and problem-solving capabilities than traditional computing technologies. Examples of future computing technologies include the organization of high-performance computing (HPC) systems, quantum computers, and dedicated accelerators for AI and domain science. Equally important are advancements in the constituent components of future computers, including novel processing elements, memory, and networks.

There are far more possibilities than those listed here. PNNL is currently making strides in bringing high-performance scientific computing technologies from the lab into the real world, shaping the future of scientific computing.

The building blocks: Advanced computing and scientific computing

Advanced computing refers to computational systems, processes, and techniques that extend the capabilities of traditional computers. Ultimately serving as the foundation of advancements in computing technology, advanced computing encompasses HPC as well as the state-of-the-art in quantum computing, AI, and machine learning.

Scientific computing is the use of advanced computing technologies, simulations, data analytics, and machine learning to further scientific research and discovery. Researchers use scientific computing to plan and conduct experiments, model physical systems, and analyze data.

Why do future computing technologies matter?

In the years to come, scientific computing technologies will not only benefit researchers but also lead to benefits for the nation as a whole.

Potential benefits include the following:

new scientific discoveries
quicker and more efficient computing capabilities
quicker communications
more automation in research and industry
stronger digital security and cybersecurity.

How will future computing technologies be used?

Future computing technologies will accelerate scientific discovery, reduce the burden on researchers, speed up computational problem-solving and data analysis, and drive the development of new innovations for industry and consumers.

The integrated computational tools being developed at PNNL will enable domain science researchers to analyze, model, simulate, and predict complex phenomena in areas ranging from molecular, biological, subsurface, and atmospheric sciences to complex networked systems.

Future computing technologies at PNNL

a technician in safety goggles and ear protection works on a server rack

PNNL is leading the next generation of computing for scientific discovery.
Explore our Computing & AI story

PNNL provides science, technologies, and leadership in creating and enabling new computational capabilities to solve challenges using extreme-scale simulation, data analytics, and machine learning. PNNL delivers the computer science, mathematics, computer architecture, and algorithmic advances that enable the integration of extreme-scale modeling and simulation with knowledge discovery and model inference from petabytes of data.

Research areas

Research at PNNL covers a multitude of areas, including the following:

advanced computer systems and architectures
advanced memory systems
system software and runtime systems
performance modeling and analysis
quantum computing
high-performance data analytics
machine learning techniques at scale.

Current projects

PNNL has recognized expertise in evaluation and capability prediction for both current and emerging large-scale system architectures. Our researchers also lead efforts to prepare the Department of Energy (DOE) for future eras of advanced computing, including through the development of new software tools.

AMAIS: Advanced Memory to Support AI for Science (AMAIS) explores novel large disaggregated shared memory systems.
ENCODE: The End-to-end Co-design for Performance, Energy Efficiency, and Security in AI-enabled Computational Science (ENCODE) project continues the success of CENATE by designing, deploying, and operating advanced architecture test beds for the assessment of novel hardware designs.
DeCoDe: Democratization of Co-design for Energy-Efficient Heterogeneous Computing (DeCoDE), with roots in the Data-Model Convergence Initiative, has helped to establish expertise and develop the cross-disciplinary acumen required to implement hardware-software co-design while meeting energy-efficiency challenges spanning traditional high-performance computing, to edge computing, to the integration of sensor data from DOE’s experimental user facilities.
Global Arrays Toolkit: The Global Arrays Toolkit is a programming model that provides primitives for one-sided communication and atomic operations.
Lamellar HPC Runtime: The Lamellar HPC Runtime is an asynchronous tasking runtime for HPC systems.
SHAD: The Scalable High-Performance Algorithms and Data-Structures library (SHAD) provides scalability and performance to support different application domains, including graph processing, machine learning, and data mining.

PNNL is also innovating in areas of data–model convergence and charting a new path in integrating the elements of HPC with data analytics to enable new scientific discoveries and computational capabilities.