I had never realized how fascinating sorting algorithms could be, until I saw this amazing video. It motivated to implement some of them in my favorite language, Go.
I also tried to slightly improve the performance of some of them using concurrency (see mergesort.go and quicksort.go), since it's so easy and pleasant to work with in Go. It can lead to up to 50% and 25% performance improvment respectively.
Finally, I wrote a CLI that benchmarks them and generates the data in a .csv or .json file.
- Selection Sort
- Insertion Sort
- Bubble Sort
- Merge Sort ❤️ / Merge Sort Conc
- Quick Sort / Quick Sort Conc
- Radix Sort
- Go Sort
- Counting Sort
- Cocktail Shaker Sort
The CLI tool can generate some comparative data on demand. It takes three parameters:
| flag | usage | default value |
|---|---|---|
-a |
Names of algorithms to be tested(*) | "" (all) |
-s |
Sizes of the arrays to be sorted | "100 1000 10000" |
-o |
Filename output (.csv or .json) | "results.csv" |
Each parameter accepts several values, in that case they must be quoted and separated by a space.
(*) Full names list (not case-sensitive):
- BubbleSort
- SelectionSort
- InsertionSort
- MergeSort
- MergeSortConc
- QuickSort
- QuickSortConc
- RadixSort
- GoSort
- CountingSort
go-sorting-algorithms$ go run .
go-sorting-algorithms$ go run . -a "MergeSort MergeSortConc QuickSort QuickSortConc" -s "1000 1000000" -o "excel.csv results.json"
This is a no-go. 💩
Well, in fact the CLI application is cross-platform compiled and accessible without any additionnal installation through the executables in /bin directory.
path/to/go-sorting-algorithms/bin$ ./gsa -a "MergeSort QuickSort" -s 1000000C:\path\to\go-sorting-algorithms\bin>gsa.exe -a "MergeSort QuickSort" -s 1000000Of course you should never run executables you can't 100% trust, so maybe don't do this.
- Implement more funky algorithms
- Add output
none/cli(no file generated) as default output value - Add output
localhost:PORTto get the json served on local http server