High Performance Computing (HPC)

Bringing Super Computing, Climate and Weather to the Desktop
Imagine running a detailed climate analysis for anywhere in the world on your desktop at the click of a mouse
Imagine running a detailed climate analysis for anywhere in the world on your desktop at the click of a mouse

Cloud Computing

NSAP has developed expertise in cloud computing to achieve the goals of cost-effectiveness, accessibility, security, and flexibility. The Bihar World Bank project was required by the sponsor to set up a web interface that can be commonly accessed by NSAP scientists and developers, as well as the sponsor and interested parties worldwide. The web application hosted on the Amazon Web Services (AWS) provides precipitation climatology display only on demand, thus making it cost-effective for cloud implementation. In addition, NSAP has tested running Numerical Weather Prediction (NWP) models on Penguin HPC cloud and is in the process of developing the capability to run NWP models on demand on AWS.

Big Data

Data production at NCAR has grown dramatically in recent years, and based on feedback gathered during several user surveys, data storage is now the primary concern of scientists, software engineers, and managers, and is expected to remain at the top of the list for the coming years. With the introduction of new high-resolution cameras, telescopes, sensors, and models, the amount of data produced by a single project can easily reach tens of PetaBytes, which represents an unsustainable workload for affordable storage and memory systems. NSAP is currently developing solutions involving generative models based on modern AI technologies to reduce the need for big in-memory datasets during the post-processing analysis.

New Frontiers

  • Exascale Computing Effort: the advent of new heterogeneous and domain-specific hardware dictates changes in the way software access and manages data. Furthermore, the expected increase in parallelism and the number of cores in future HPC platforms will make fault-tolerance and code resilience more important. For these reasons, NSAP is actively researching new solutions and technologies to make models, pre- and post- processing codes run efficiently on the modern HPC hardware, even in the presence of failures.
  • HPC for Urgent Decision Making: HPC in the weather and climate community is usually used to simulate disasters and concerning atmospheric conditions for human security. What is currently missing is the possibility of using High-Performance Computing capabilities in order to perform urgent decision-making tasks with guaranteed time constraints. The unreliable nature of current (and future) HPC platforms has forced researchers and practitioners to find reliable alternatives, like cloud computing or dedicated hardware.

Search through all publications in NCAR's OpenSky Library.

High Performance Computing (HPC)