3
$\begingroup$

Currently I am using Firefox as default browser, I have no issue to naviguate with it, but I have a little limited memory (8GB, 6.8 available) and Firefox takes a lot of RAM for nothing big, the more I use it in the time, the more it "eats" the RAM. I want to know which is the lightest browser to use jupyter notebook to save my RAM for computations on my dataset for instance.

$\endgroup$
  • 2
    $\begingroup$ You can try Google Chrome or Microsoft Edge. $\endgroup$ – Shubham Panchal Feb 24 '19 at 4:20
  • $\begingroup$ @Shubham Panchal I did try with Chrome and it consumes very less than Firefox clearly (about -2GB). I know the constraint to find a browser which supports Javascript because Jupyter needs it, I did read articles about it (in Quora) but it never precises which is the lightest. $\endgroup$ – AvyWam Feb 24 '19 at 12:16
  • $\begingroup$ There are no more browsers out there. Firefox, Chrome, that's it. Opera switched to Chrome engine long time ago, Edge will switch soon. Safari on Mac and iOS share common ancestor with Chrome (the Webkit engine). $\endgroup$ – tomash Mar 5 '19 at 14:42
3
$\begingroup$

Try 32-bit Chrome or Firefox, it may significantly lower memory usage.

Remember, it is surprisingly difficult to estimate physical memory consumption of given app: https://dzone.com/articles/windows-process-memory-usage-demystified

Also - unused RAM is wasted RAM. Do not blame browser for taking and using, what's available. When overall memory usage will raise, browser can drop some internal cache and free that memory.

| improve this answer | |
$\endgroup$
1
$\begingroup$

I have tried Midori and the RAM usage is lower than Firefox, but it seem a little bit slow with window motion.

| improve this answer | |
$\endgroup$
0
$\begingroup$

You might be better off restructuring your code to not run heavy computations within notebooks.

I found https://towardsdatascience.com/introducing-jupytext-9234fdff6c57 this helful in the past, or https://github.com/minodes/pynb to create a simple .py file from a notebook on the fly.

That way i can switch out the dataset / some parameters and run heavy computation separately from jupyter and not launched from a browser.

| improve this answer | |
$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.