ChangeLog

Differences Between Caliper 0.5 and 1.0

• 0.5 benchmarks extended SimpleBenchmark while 1.0 benchmarks extend Benchmark
• 0.5 only measured runtime by default. 1.0 measures runtime and allocation.
• Runtime results from 0.5 were subject to corruptions from factors like GC and JIT occurring mid-timing. 1.0 now detects and reports such events.
• 0.5 produced incomplete and possibly misleading console output. 1.0 does not.
• 0.5 required an API key to upload results to the webapp. Caliper 1.0 uploads every run automatically, but an API key is still required to associate a run with a particular user.
• 1.0 stores the JSON result of every run in ~/.caliper/results/. 0.5 required an explicit flag.
• 0.5 was configured using ~/.caliperrc while 1.0 uses ~/.caliper/config.properties.
• Many bug fixes