Next: Naive Fibonacci Up: Benchmarks Previous: Benchmarks

Naive Reverse

The standard benchmark, naive reverse, was tested in three environments:

  1. in REL, compiled into the WAM without indexing
  2. in REL with indexing
  3. in LL, automatically generated from the REL specification (see section

For naive reverse applied to a list with 50 elements, the following internal run times were measured:

The speed up achieved by indexing and by transformation into LL (compared with the indexed version) are both approximately 2.

Harold Boley & Michael Sintek (