A test harness that comprises one or more tests for a program in one example is executed. An output from the test harness is received. The output comprises one or more respective test results for the one or more tests. A verification that the one or more respective test results comprise one or more expected test results for the one or more tests is received from a user. The one or more expected test results are stored in a benchmark file. A test harness for benchmark file generation that comprises one or more tests for a program in a further example is created. The test harness comprises one or more calls to one or more subroutines that are employable for one or more of: a definition of one or more expected test results for the one or more tests; and/or a verification of the one or more expected test results.