This site contains documentation for v0.2 of the Intrepydd programming system being developed in the DDARING research project with contributors from Georgia Tech, UIUC, U.Michigan, and USC. Intrepydd is a Python-based domain-specific programming language for data analytics kernels, suitable for ahead-of-time compilation on current and future hardware platforms and accelerators. Intrepydd is not intended for writing complete/main programs. This release (v0.2) contains an implementation of an early version of the Intrepydd language. We welcome feedback from testers of this release; your feedback will influence future enhancements to the Intrepydd language, its implementation, and its documentation.
You can use Intrepydd from within Jupyter. The following collection of notebooks shows you how.
Video Preview. For a quick video overview, which goes through the k-means case study below, click here (YouTube).
Functions Reference. For a Jupyter-based summary and demo of many of the Intrepydd-supported functions, refer to the functions reference or its preview. Some are not yet included (e.g., sparse matrix functions), so do refer to the complete Intrepydd built-in functions and libraries list.
Below, the title links directly to a Jupyter notebook. If, instead, you just want to preview the notebook, click the “Preview” links, which redirect you to the GitHub versions.
Since Jupyter is usually run on a web browser on your local machine, the following three steps need to be performed to enable Jupyter to connect with a Docker instance running Intrepydd. In the following, we will assume that the Docker instance is running on a server named ubuntu@54.198.41.116, that is representative of a typical EC2 instance, and key_CqPMRtDoyTYq.pem is the ssh identification file located in the current directory. The actual path names may vary depending on the server instance that you use.
Step 1: ssh to the server and run Jupyter via Docker
$ ssh -i key_CqPMRtDoyTYq.pem ubuntu@54.198.41.116
$ sudo docker run -p 8888:8888 --privileged -it --rm --net=host -v /data:/data sdhph1-eval1 /bin/bash -c "/opt/conda/bin/jupyter notebook --ip=0.0.0.0 --port=8888 --no-browser --allow-root"
Copy the URL returned by the “docker run” command, e.g., http://127.0.0.1:8888/?token=0628d41265d1fa85e7d6c489ad3834317e9a2e6234bd9e29
Step 2: forward port 8888 from the server to your local machine
$ ssh -i key_CqPMRtDoyTYq.pem -L 8888:localhost:8888 ubuntu@54.198.41.116
Step 3: Open the URL copied at the end of step 1 in a web browser, e.g., http://127.0.0.1:8888/?token=0628d41265d1fa85e7d6c489ad3834317e9a2e6234bd9e29