Configurable benchmark runner

[codingwithmagga]

Description

Build a benchmark runner that accepts configuration parameters controlling execution behavior.

Acceptance Criteria

• Thread count, data size, and iteration count are configurable at runtime
• Configuration can be passed programmatically and via CLI arguments
• Runner executes a given kernel repeatedly for the configured number of iterations
• Invalid configuration values are rejected with clear error messages

Notes

TDD — tests are written before implementation code.