Is there something like Jupiter for C/C++

5 points by bdjsiqoocwk ↗ HN
Is there something like Jupyter for C++? Specifically, something that is a long lived process, that keeps its long lived state over this time, and allows me to give it some code, have it JIT compiled, and executed.

Use case: imagine large files in disk and I want to read and parse into memory (imagine this step will take a long time, say tens of minutes), compute some metrics, and generate a plot, but I'm not sure what metrics I want yet, and I would need to iterate on it. Without Jupyter I'm going to have to read and parse the files every time I want to test what some metrics look like. With Jupyter I don't need to read and parse each time, the data is already in memory.

5 comments

[ 3.1 ms ] story [ 26.5 ms ] thread
There is a Jupyter kernel for the C++ interpreter cling. I tried it once but my experience was that C++ doesn't work very well in a notebook environment. Honestly I would suggest getting the data into a format compatible with NumPy or something similar and using Python, or at least do it at a smaller scale in Python to figure out your metrics.
Alternatively, put the metrics into a relational database. They tend to be optimized for IO and concurrent access.
It sounds like the issue is IO not interface.