vanklompf / perf-challenge6

Performance Challenge #6

See the announcement post here.

The task

The task is to split a text and count each word's frequency, then print the list sorted by frequency in decreasing order. Ties are printed in alphabetical order.

Example:

$ echo "apple pear apple art" | ./wordcount.exe
apple   2
art     1
pear    1

See additional rules and assumptions in wordcount.cpp.

How to set up the environment

Here is the list of tools you absolutely have to install to build the benchmark:

• CMake 3.13 or higher
• Google benchmark.

Next steps depend on your platform of choice. So far we support native builds on Windows and Linux. Check out the instructions specific to each platform (Windows) (Linux).

How to build & run

Build:

cmake -E make_directory build
cd build
cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_CXX_COMPILER="clang++" -DCMAKE_C_COMPILER="clang" -G Ninja ..
cmake --build . --config Release --parallel 8

Validate:

cmake --build . --target validate

Benchmark:

cmake --build . --target benchmark

Benchmark against baseline:

python3 check_speedup.py -challenge_path path/to/repo -bench_lib_path ~/workspace/benchmark/benchmark -num_runs 3

Keep in mind, the large input file is very large. We recommend having at least 16GB RAM.

Acknowledgments

• Inspired by https://github.com/juditacs/wordcount/
• Idea for the challenge was suggested by Karthik Tadinada.