routines for performing rapidly-parallelized linear regression on nvidia GPUs
Switch branches/tags
Nothing to show
Clone or download
Pull request Compare This branch is even with tommysprague:master.
Latest commit 897e332 Feb 17, 2017
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
README.md Create readme.md Feb 17, 2017
gpuregress.m initial commit Feb 17, 2017
gpusse.m initial commit Feb 17, 2017
gridFit_vars.mat initial commit Feb 17, 2017
gridfitgpu.m initial commit Feb 17, 2017
run_gridfitgpu.m initial commit Feb 17, 2017

README.md

gridfitgpu

routines for performing rapidly-parallelized linear regression on nvidia GPUs

Rapidly determine the best model predictions to account for each of many measurements simultaneously. Requires a compatible NVidia GPU (see MATLAB documentation), which requires Windows or Linux.

Approximately 6-20x speedup, depending on compute capabilities. Identical results to using similar functions on CPU.

Required for GPU-accelerated functions in vistasoft (https://github.com/tommysprague/vistasoft_ts/tree/sprague_gridfit_updates)