Instruction file imported from Y-55/ICS507-Matrix-Multiplication-Project (.cursor/rules/ics-matrix-multiplication.mdc). Copyright stays with the author.

ICS 507 Matrix Multiplication Project Rule

Follow these project-specific conventions when assisting in this repository:

• Build with ./scripts/build.sh (preferred and default).
• Do not invoke g++ manually unless the user explicitly asks or the build script fails.
• Generate tests with python3 scripts/generate_input.py <n> (optional flags: --output, --seed, --min, --max).
• Run with: ./build/matrix_mult <input-file> [threads] [strassen-base-case] [mode]
• Respect positional CLI arguments:
  1. input-file, 2) threads, 3) strassen-base-case, 4) mode.
• If setting mode, always include threads and strassen-base-case first.
• Valid mode values are exactly: all, Sequential, ParMtrixMult, Strassen, ParStrassen.

Preserve algorithm API signatures:

Matrix multiplySequential(const Matrix& a, const Matrix& b);
Matrix multiplyParMatrixMult(const Matrix& a, const Matrix& b, int threads);
Matrix multiplyStrassen(const Matrix& a, const Matrix& b, int baseCase);
Matrix multiplyParStrassen(const Matrix& a, const Matrix& b, int baseCase, int threads);

Treat these as coordination-sensitive files; avoid changing them unless the task explicitly requires it:

• src/matrix.hpp
• src/matrix.cpp
• src/io.hpp
• src/io.cpp
• src/algorithms.hpp
• src/main.cpp

When running experiments or documenting results, keep reporting aligned with project expectations: machine specs, tested matrix sizes, tested thread counts, tested Strassen base cases, runtime tables, correctness vs sequential baseline, and observed performance differences.