Design a cache-locality measurement experiment

[devkyato]

Prepare the Computer Architecture roadmap's cache-locality experiment without claiming results before measurement.

Acceptance criteria

• state a testable hypothesis comparing row-major and column-major traversal
• control matrix size, data type, iteration count, build mode, and warm-up
• use a monotonic clock and report multiple runs rather than one timing
• record compiler, optimization flags, CPU, OS, and limitations
• commit source, raw results, and a journal interpretation
• keep conclusions limited to the measured machine and configuration

Open a design PR before collecting final measurements so the methodology can be reviewed.