Viskores Update

May 5, 2025 | HPSF Conference

Kenneth Moreland

Oak Ridge National Laboratory

ORNL IS MANAGED BY UT-BATTELLE LLC �FOR THE US DEPARTMENT OF ENERGY

1

Acknowledgements

This work was supported by the U.S. Department of Energy (DOE) RAPIDS SciDAC project under contract number DE-AC05-00OR22725.

This research was supported by the Exascale Computing Project (17-SC-20-SC), a collaborative effort of two U.S. Department of Energy organizations (Office of Science and the National Nuclear Security Administration) responsible for the planning and preparation of a capable exascale ecosystem, including software, applications, hardware, advanced system engineering, and early testbed platforms, in support of the nation’s exascale computing imperative.

This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research, under Award Numbers 10-014707, 12-015215, and 14-017566.

Thanks to many, many partners in labs, universities, and industry.

2

What is Viskores?

3

3

Viskores: Visualization on Accelerators

Contouring and features

Rendering

And much more…

4

Why Viskores?

5

5

Distributed

Parallelism

6

Distributed

Parallelism

7

8

9

Viskores Scaling on Frontier

• Science Problem
• Frontier, the world’s first exascale machine, achieves its scale with next-generation processors and unprecedented parallelism.
• The bandwidth of computation on modern supercomputers like Frontier exceeds the bandwidth of storage, making complete captures of simulation data impossible.
• Technical Solution
• Viskores, a visualization library designed for modern accelerators, was enhanced with multi-node communication to scale rendering to the entirety of Frontier.

Viskores rendering of 79.5 trillion cell synthetic data set using 74,088 GPUs on Frontier. This image was rendered in 0.3 seconds.

• The combined rendering algorithm was applied to a synthetic data set that was generated, processed, and rendered on the majority of Frontier.
• Science Impact
• Viskores is instrumental in providing in situ visualization capability to multiple ECP simulation codes, both directly and via ECP ALPINE’s ParaView, VisIt, and Ascent tools.
• This demonstrated functionality is an important feature of in situ processing, and this low time overhead minimizes the overall impact on the simulation performance.

10

11

Why Join the HPSF?

12

12

Helping Joining Viskores with Customers

Users

13

Providing Consistent Support for Software Stewardship

Continuous Integration (CI) Testing Support

Web Design and Hosting

Graphic Design, Logos

14

HPSF Plans

15

15

HPSF Next Steps

Support the open, collaborative effort for usable HPC software

Transition from Emerging to Established project

Interest in benchmark workgroup

User Group and Hackathon community events

16