Ta repozitorij vsebuje implementacijo izbranih algoritmov množenja matrik in je del moje diplomske naloge.
This repository contains implementation of next fast matrix multiplication algorithms:
- classic matrix multiplication
- Strassen's algorithm (
<2, 2, 2>, tensor rank 7) - Laderman's algorithm (
<2, 2, 2>, tensor rank 23) - Bini's approximative algorithm (
<2, 2, 3>, tensor rank 10) - Schonhage's approximative algorithm (
<3, 3, 3>, tensor rank 21)
Algorithms are far from optimized, the main goal was to see the diffenences in time complexitiy.
Clone this repository
git clone [email protected]:matejm/fast-matrix-multiplication.git
Go to root folder of the repository and do the usual compile & run cycle:
mkdir -p build
cd build
cmake ..
make
Run demo of all algorithms or time algorithm execution for different matrix sizes:
./../bin/demo
./../bin/time_multiplication
Google's testing framework is used. To install it on Ubuntu, run:
sudo apt install libgtest-dev
To test the algorithms, run:
make tests
./../bin/tests