OpBench

This is a basic client to show achieved throughput. It is designed to be modifiable, so that a consumer can substitute a command that is more representative of their workload. It is not designed to be a fully featured benchmarking tool.

The work to perform is specified via Work; optionally, Init can be supplied as a once-only setup step.

Variables:

• Connections: the number of actual Redis connections to use. Defaults to 1. Most SE.Redis scenarios do not benefit from multiple connections, due to the internal multiplexing.
• Clients: the number of effective parallel clients to simulate. Defaults to 1 per connection.
  • each client represents a concurrent call path in your application code; high concurrency is very normal in .NET applications (especially server scenarios)
• PipelineDepth: the number of commands to pipeline (batch) per client. Defaults to 1.
  • Pipelining is a very effective way of improving throughput, but it is not always applicable; in particular, it is not applicable if you need the result of one command before issuing the next. Pipelining is especially suitable for write-heavy workloads.
  • Note that SE.Redis internally pipelines commands from parallel callers (clients) on the same connection.
• IterationsPerClient: the number of operations to perform per client. Defaults to 1000.
• RepeatCount: the number of times to repeat the test