A multi-GPU QSGW implementation in ecalj reaches over ten times speedup on one GPU node versus four CPU nodes, and a mixed-precision mode triples that speed with meV-level accuracy changes.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.comp-ph 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Efficient implementation of the quasiparticle self-consistent $GW$ method on GPU
A multi-GPU QSGW implementation in ecalj reaches over ten times speedup on one GPU node versus four CPU nodes, and a mixed-precision mode triples that speed with meV-level accuracy changes.