BIZNIQUE / LINEAR ALGEBRA / 06
Benchmark the work you actually run
Benchmark the work you actually run
Compare storage assumptions before making claims about runtime.
Embedded Files
THE IDEA
Read the example.
Read the example.
This calculator compares a dense float64 matrix with a simplified compressed sparse row representation: eight-byte values, four-byte column indices and four-byte row pointers. It reports estimated storage, not elapsed time.
Dense storage = n × d × 8 bytes
Use it well
Use it well
A useful runtime benchmark repeats equivalent work, checks equal outputs, separates setup from execution and records the environment. Sparse storage helps only when enough entries are zero.
TRY IT YOURSELF
A quick check.
A quick check.
For 1,000 rows and 100 columns, how much space do dense float64 values occupy?
Check your reasoning
Check your reasoning
800,000 bytes, before object overhead. That is about 0.76 MiB.
Page updated
Google Sites
Report abuse