Example of converting GPU algorithm to Triton custom backend
Opened this issue · 1 comments
kpedro88 commented
- pick a classical algorithm (e.g. HCAL reconstruction) that has already been ported to CUDA
- get GPU code for that algorithm
- set up custom backend using library for that algorithm
- set up associated Sonic producer to handle inputs/outputs
- document procedures for steps 3 and 4
assigned to: @violatingcp, ?