npmemory

NumPy memory editing for efficient array calculations.

Results for sample routine:

Box Averaging	Pure Python (ms)	Python with memory editing (ms)	Speed factor
Windows 10	2015 ms	18 ms	112x
Linux	1422 ms	13 ms	107x

Note: Tested on Intel i9-9880

Graph showing speed increase

Windows Setup

Step 1: Install the 64-bit TDM GCC compiler from the following link: https://jmeubank.github.io/tdm-gcc/ Ensure that you select 'TDM GCC MinGW w64'

Step 2: git clone https://github.com/Fletcher-Climate-Group/npmemory

Step 3: Run 'python cross_compile.py'

Linux Setup

Step 1: Install the latest version of gcc for your distribution

Step 2: git clone https://github.com/Fletcher-Climate-Group/npmemory

Step 3: Run 'python cross_compile.py'

Mac OSX Setup

Step 1: Install the latest version of gcc using Homebrew

Step 2: git clone https://github.com/Fletcher-Climate-Group/npmemory

Step 3: Run 'python cross_compile.py'

Quick Start (all operating systems)

Once you have compiled the npmemory module for your OS, run the following example report to ensure the memory editing works correctly.

cd examples
python report.py

This should return a report which compares the speed differential between a pure Python routine and the C-augmented memory editing routine.