MathBench Performance Suite

A shitty C++ benchmarking tool for testing mathematical operations performance. Primiarly to test ARM and RISC-V Math perfomance

Features

• 12 mathematical benchmarks
• "Multi"-threaded support
• UI designed for 80x24 display
• "Real-time" progress tracking

Project Structure

MathBench/
├── src/               # Source files
│   ├── main.cpp       # Entry point
│   ├── MathBench.h    # Main benchmark class header
│   ├── MathBench.cpp  # Benchmark implementations
│   ├── UI.h           # Terminal UI header
│   └── UI.cpp         # Terminal UI implementation
├── build/             # Build artifacts (object files)
├── external/          # External dependencies
│   └── picosha2.h     # SHA-256 hashing library
├── Makefile           # Build configuration
└── mathbench          # Compiled executable

Dependencies for Debian based systems

Native Build

apt update && apt install make g++ gcc build-essential 

If you don't wanna add sudo manually

sudo apt update && sudo apt install make g++ gcc build-essential 

Cross-Compilation Toolchains

For cross-compiling to different architectures:

sudo apt install \
    gcc-arm-linux-gnueabi g++-arm-linux-gnueabi \
    gcc-arm-linux-gnueabihf g++-arm-linux-gnueabihf \
    gcc-aarch64-linux-gnu g++-aarch64-linux-gnu \
    gcc-riscv64-linux-gnu g++-riscv64-linux-gnu

Building

Native Build

make

This will:

• Create the build/ directory if it doesn't exist
• Compile all source files
• Link the final executable mathbench

Cross-Compilation

Build for specific architecture:

make armv6      # ARMv6 (Raspberry Pi Zero, Pi 1)
make armv7      # ARMv7
make armhf      # ARM Hard Float (ARMv7)
make arm64      # ARM64/AArch64 (Raspberry Pi 3/4/5, most modern ARM)
make riscv64    # RISC-V 64-bit (Milk-V, StarFive)

Build all architectures at once:

make all-cross

This will create binaries:

• mathbench-armv6
• mathbench-armv7
• mathbench-armhf
• mathbench-arm64
• mathbench-riscv64

Running

Run with default settings (1 thread):

./mathbench

Run with multiple threads:

./mathbench 4

Run cross-compiled binary on target device:

# Transfer binary to target device, then:
./mathbench-arm64 4

Benchmarks Included

1. Basic Arithmetic - Addition, multiplication operations
2. Trigonometry - Sin, cos, tan calculations
3. Logarithm - Natural logarithm computations
4. Exponential - Exponential function calculations
5. Square Root - Square root operations
6. SHA-256 Hashing - Cryptographic hash operations
7. Array Sorting - std::sort on large arrays
8. Matrix Multiplication - Dense matrix operations
9. Prime Numbers (Sieve) - Sieve of Eratosthenes algorithm
10. Fibonacci - Recursive Fibonacci calculation
11. Monte Carlo Pi - Pi estimation using random sampling
12. Fourier Transform (DFT) - Discrete Fourier Transform

Cleaning

Remove build artifacts:

make clean

Requirements

• C++17 compatible compiler (g++, clang++)
• POSIX-compatible terminal for UI features
• Linux/Unix-like operating system

License

See LICENSE file for details.