use massively parallel sundials solvers for running many solves with different input parameters #2644
Labels
difficulty: hard
Will take several weeks
feature
in-progress
Assigned in the core dev monthly meeting
priority: medium
To be resolved if time allows
solver
Description
Sundials has a few "massively parallel" solvers that can solve many independent models in parallel on a GPU (SUNLinSol_OneMklDense and SUNLinSol_MagmaDense for dense systesm, SUNLinSol_cuSolverSp_batchQR for sparse). These could be used for parameter sweeps, data generation or inference.
See also #2217
Motivation
enable the use of GPUs to solver a large number of PyBaMM models in parallel
Possible Implementation
No response
Additional context
No response
The text was updated successfully, but these errors were encountered: