Skip to content

Latest commit

 

History

History
33 lines (24 loc) · 948 Bytes

README.md

File metadata and controls

33 lines (24 loc) · 948 Bytes

An exploratory package for calling NVIDA’s CUDA runtime API

2026-05-29

Infrastructure for calling CUDA from R

Currently under construction…

This package skeleton is meant to serve as a starting point for a hackathon project. Code chunks present here are guides for using NVIDIA’s brev platform for deploying testing and development environments. A docker container built from rocker’s RStudio container has been created to facilitate remote development on deployed resources. NVIDIA’s brev resource also supports remove VScode access.

Initial brev interactions.

# Using defaults should pick the cheapest GPU
brev create testinstance

# log in to the instance you asked to created
brev ssh testinstance

# set up port forwarding to your device
brev port-forward testinstance --port 8787:8787

# check the instances you've created
brev ls

# delete the instance you created
brev delete testinstance