This simple program uses Geospatial Partitioning to find the postal code of a specific geographical point in Germany. I use this program to compare the performance of four popular programming languages. There are some points regarding this benchmark:

• I didn't use any third-party library for this benchmark, and all the implementations remain based on standard data structures and libraries available in each programming language.
• The benchmark result is based on sample data containing polygons that cover all the german region.
• To keep the system as simple as possible, I am using a custom file format for importing the polygons data. you can download the original data from https://www.suche-postleitzahl.org/plz-karte-erstellen as a Geojson file
• the algorithm contains three operation modes:
  • HIGH
  • MEDIUM
  • LOW

• The results are related to the following compiler versions :
  • gcc 9.3.0
  • clang 10.0.0
  • jdk-17
  • go version 1.17.1
  • rustc 1.55.0

cd cppgeo
mkdir __buid
cmake ..
make -j

uncomment set(CMAKE_C_COMPILER "clang") set(CMAKE_CXX_COMPILER "clang++") in the CMakeLists.txt file
cd cppgeo
mkdir __buid
cmake ..
make -j

cd java
export PATH=$PATH:{JDKPATH}
export PATH=$PATH:{GRADLEPATH}/bin
gradle build

cd go
go build

cd rust/polygon_search
cargo build --release

For viewing the available options for each version, use -h switch

The results are based on 1000000 times lookup of the specified point and are in Microsecond. To consider the effect of the java jit loop predicate, I executed the tests in two modes, In the first mode I searched a constant point and in the second mode I added some random noises to it

The first Scenario considers the location of the Freiburg central church.

Freiburger Münster Lookup results (High and low accuracy)

Random points around Freiburger Münster

A location in a big polygon (High and low accuracy)

A location in a big polygon ( Random Points )