Tomasz Grabiec 121fa4ff46 test: Introduce test for in-memory CQL query performance ...

The schema and queries resemble that used by cassandra-stress.

Results on my laptop:

$ build/release/tests/perf/perf_simple_query -c1
Concurrency = 100 x 1
Creating 1000 partitions...
Timing single key selects...
452146.50 tps
449365.24 tps
457650.58 tps
460334.78 tps
458281.51 tps
Timing random key selects...
439181.03 tps
449899.67 tps
405146.90 tps
440228.84 tps
440889.50 tps

$ build/release/tests/perf/perf_simple_query -c3
Concurrency = 100 x 3
Creating 1000 partitions...
Timing single key selects...
302615.08 tps
301471.02 tps
303040.67 tps
302114.77 tps
302465.13 tps
Timing random key selects...
627516.46 tps
628978.04 tps
623664.15 tps
624098.48 tps
614549.85 tps

2015-03-25 12:59:25 +01:00

..

perf_cql_parser.cc

tests/perf: CQL statement parser performance test

2015-02-26 14:14:53 +01:00

perf_mutation.cc

Use column_definition via const reference

2015-03-24 12:03:00 +01:00

perf_simple_query.cc

test: Introduce test for in-memory CQL query performance

2015-03-25 12:59:25 +01:00

perf.hh

tests: Add time_parallel() utility

2015-03-25 12:59:25 +01:00