I've been doing some primitive benchmarks on test runners lately, and I found
that the R# test runner is almost twice a slow as the NUnit-gui test runner.
The test case was 1000 trivial tests (Assert.IsTrue(true)). I know this test
case is very trivial, but most unit tests should run almost this quickly,
based on the act of creating lots of small, focused, fast tests.
For 1000 tests, NUnit-gui ran them in 15 seconds, nunit-console ran them
in 7 seconds, and R# ran them in 30 seconds.
R# was run in hierarchical mode, not following the currently running tests.
Tests were organized into 10 classes, 100 tests per class.