emiwar / caJulia

Fast CNMF-E implementation with Julia and CuArrays
0 stars 0 forks source link

Parallell oasis #5

Open emiwar opened 2 years ago

emiwar commented 2 years ago

How slow would it be to run OASIS on the GPU? If it's possible each GPU kernel could run with one set of gamma/lambda. Maybe that's faster than running on CPU? Otherwise at least make it thread-parallel on the CPU...