These benchmarks measure the overhead of various implementations of a
conduit between a source: () => Future[Buf]
and sink: Buf =>
Future[Unit]
.
These benchmarks measure the overhead of various implementations of a
conduit between a source: () => Future[Buf]
and sink: Buf =>
Future[Unit]
.
The steps are as follows:
1. connect the source to the conduit 2. drain three items from the conduit into the sink
The control benchmark measures the baseline when there is no conduit, i.e., the source feeds directly into the sink.
Measure Scheduler.submit/run time when there are multiple threads.
Measure Scheduler.submit/run time when there are multiple threads. Previous to changing SampleScale from 50 to 1000, there was sometimes contention for the global threads lock in getCurrentThreadCpuTime().