Python Forum

Full Version: Memory consumption grows during execution only on single machine
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
I have a strange problem. I've spent the whole weekend trying to find the root cause of the unrestricted growth of memory consumption of my Python program until it turned out, that it happens only on one PC...
I run the same code on 3 machines: Ubuntu PC with Intel processor, Macbook, and Ubuntu with AMD Ryzen CPU and observe the linear memory consumption during operation only on the last one.
I install everything in the same way on each machine: same Python version (set up with pyenv) same packages (using pip freeze) and only on Ryzen CPU based Ubuntu machine I get the growth of memory usage during operation until the process terminates or I run out of memory (please have a look on plots below).
The only thing which maybe differs Ubuntu machines apart from CPU is the kernel version, on one I am on 4.15.0-76 (where the memory consumption is stable) and on Ryzen CPU based, where I have linear growth of memory consumption I have 5.3.0-28. But I am not sure if that can be an issue...

Memory increase on Ubuntu with kernel 5.3 and Ryzen
[Image: dltDH6r.png]

Stable memory on another machine (here Mac)
[Image: IGzNXJu.png]

I will be thankful for any hints what can be the source of the problem here.
If this is a new kernal for the Rizen, I'd suspect that, and send the data you provide here to the writer of the 5.3.0-28 kernal to see what they have to offer.
There's a blogon that kernel and AMD Rizen here:
This appears to not be related to the symptoms you are seeing, but who knows.
You can report here:
Thank you! I will mark thread as solved as soon as I get it sorted out.