String compare - Timing information

[andresriancho]

Research story

As a researcher I would like to know how long it (usually) takes to run a == b in python, c, java and ruby (for different string sizes and string similarities/differences).

Tasks

• Create python2.7 and python3k, c, java and ruby programs to compare two strings
• Use this code for C tests
• Use this code for Java
• Read source for python3 strcmp and link from wiki
• Commit all sources to a pico-string-compare-local repo
• Create graphs for x) string length y) time it took to compare string where strings are in Ax, Ay shape and A length increases
• Test with different strings for each language:
  • A vs. A
  • A vs. B
  • A * 128 vs. A * 128
  • A * 32 vs. A * 32
  • A * 127 + B vs. A * 127 + C
  • A * 31 + B vs. A * 31 + C
  • A * 64 + B + A * 64 vs. A * 64 + C + A * 64
• Use the best profiler for each language to measure 1M string compares and then divide to get how much it takes to compare one
• Write results to wiki
  • Include python/java/ruby versions
  • Include OS information (kernel, etc.)
  • Include CPU info (64bit)
• Compare results with previous research

References

• Min measurable delay vs. time it actually takes to compare strings
• time trial talk also discusses this issue
• Also here a table with the time it takes to compare 1 byte in different programming languages.