https://github.com/mtsokol/fortran-course

https://github.com/mtsokol/fortran-course

fortran unit-test

Last synced: 11 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/mtsokol/fortran-course
• Owner: mtsokol
• Created: 2018-04-12T09:25:58.000Z (almost 7 years ago)
• Default Branch: master
• Last Pushed: 2018-08-04T12:31:23.000Z (over 6 years ago)
• Last Synced: 2024-11-13T19:07:06.054Z (2 months ago)
• Topics: fortran, unit-test
• Language: Fortran
• Size: 119 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Fortran Course

Completed assignments for Fortran Course

## Essentials

- lab1 basic operations

- lab2 numerical integration example

## FDM

Finite Difference Method implementation with error measurement.  

For given problem:

![eq1](https://latex.codecogs.com/gif.latex?-k\frac{d^2u}{dx^2}=0)

with boundary conditions:

![eq2](https://latex.codecogs.com/gif.latex?u(0)=0)  

![eq3](https://latex.codecogs.com/gif.latex?u(1)=1)

program creates system of equations using FDM and solves it with gauss method.

Having returned solution, program computes error connected with real precision.

Concretely single (kind = 4), double (kind = 8) and fourfold (kind = 16) precision.

### Usage

Change directory to /FDM and:

```sh

$ make bulid

```

For concrete kind (4, 8, 16) use option:

```sh

$ make build KIND=kind16

```

And then:

```sh

$ make run

```

Output shows sum of errors for every iteration.  

To inspect system of equations before and after solving change directory to concrete precision and type:

```sh

$ make verbose-run

```

### Results

Plots below show error values output on logarithmic scale.

![plot](https://image.ibb.co/nuqtbx/plot.png)

To generate these plots by yourself:

```sh

$ python3 generate_plot.py

```

## Tests

Simple hello world pFunit test. 

## Matrices

Testing matrices multiplication with pFunit. Module contains simple loop implementation and also improved version with dot product.

### Usage

(FIRST TASK) Change directory to /matrices and:

```sh

$ make bulid

```

And then:

```sh

$ make run

```

(SECOND TASK) To test improved version use:

```sh

$ make run-improv

```

To run tests start with:

```sh

$ rm main.o

```

Then build tests:

```sh

$ make test

```

And finally:

```sh

$ ./test

```

### Results

(THIRD TASK) Time and matrix size dependence for all cases:  

Input for plots generated with:

```sh

$ make time-measure

```

![plot](https://image.ibb.co/gnADNd/plot.png)

## MPI 

Examlpe usage of coarrays. Combining Fortran with Python (f2py). Doxygen documentation. 

### Usage

(TASK ONE) Python usage

```sh

$ make build  

$ make run

```

Generating docs

```sh

$ make docs

```  

(TASK TWO) MPI example

```sh

$ make build-mpi

$ make run-mpi

```  

### Results

(TASK THREE) Time execution comparison

```sh

$ make time-compare

```  

For given (simple) example execution of MPI version is significantly faster.