This title appears in the Scientific Report :
2023
Please use the identifier:
http://dx.doi.org/10.34734/FZJ-2023-05321 in citations.
Please use the identifier: http://dx.doi.org/10.1145/3624062.3624249 in citations.
Advancing the distributed Multi-GPU ChASE library through algorithm optimization and NCCL library
Advancing the distributed Multi-GPU ChASE library through algorithm optimization and NCCL library
As supercomputers become larger with powerful Graphics Processing Unit (GPU), traditional direct eigensolvers struggle to keep up with the hardware evolution and scale efficiently due to communication and synchronization demands. Conversely, subspace eigensolvers, like the Chebyshev Accelerated Subs...
Saved in:
Personal Name(s): | Wu, Xinzhe (Corresponding author) |
---|---|
Di Napoli, Edoardo | |
Contributing Institute: |
Jülich Supercomputing Center; JSC Center for Advanced Simulation and Analytics; CASA |
Imprint: |
ACM New York, NY, USA
2023
|
Physical Description: |
1688–1696 |
DOI: |
10.34734/FZJ-2023-05321 |
DOI: |
10.1145/3624062.3624249 |
Conference: | SC-W 2023: Workshops of The International Conference on High Performance Computing, Network, Storage, and Analysis, Denver, CO (USA), 2023-11-12 - 2023-11-17 |
Document Type: |
Contribution to a conference proceedings |
Research Program: |
Simulation and Data Laboratory Quantum Materials (SDLQM) Domain-Specific Simulation & Data Life Cycle Labs (SDLs) and Research Groups |
Link: |
Get full text OpenAccess |
Publikationsportal JuSER |
Please use the identifier: http://dx.doi.org/10.1145/3624062.3624249 in citations.
As supercomputers become larger with powerful Graphics Processing Unit (GPU), traditional direct eigensolvers struggle to keep up with the hardware evolution and scale efficiently due to communication and synchronization demands. Conversely, subspace eigensolvers, like the Chebyshev Accelerated Subspace Eigensolver (ChASE), have a simpler structure and can overcome communication and synchronization bottlenecks. ChASE is a modern subspace eigensolver that uses Chebyshev polynomials to accelerate the computation of extremal eigenpairs of dense Hermitian eigenproblems. In this work we show how we have modified ChASE by rethinking its memory layout, introducing a novel parallelization scheme, switching to a more performing communication-avoiding algorithm for one of its inner modules, and substituting the MPI library by the vendor-optimized NCCL library. The resulting library can tackle dense problems with size up to , and scales effortlessly up to the full 900 nodes—each one powered by 4 × A100 NVIDIA GPUs—of the JUWELS Booster hosted at the Jülich Supercomputing Centre. |