----------------------------------------------------------------------------------- 1 row ----------------------------------------------------------------------------------- zilence@mx2:~/install/shadow-pgsql$ java -cp target/shadow-pgsql-0.8.0-SNAPSHOT-standalone.jar clojure.main -m shadow.pgsql.benchmark.runner 1 java ===== shadow-pgsql via java WARNING: Final GC required 1.038441572186623 % of runtime amd64 Linux 2.6.32-5-amd64 8 cpu(s) Java HotSpot(TM) 64-Bit Server VM 25.60-b23 Runtime arguments: Evaluation count : 691080 in 60 samples of 11518 calls. Execution time sample mean : 117.328437 µs Execution time mean : 117.002069 µs Execution time sample std-deviation : 37.673531 µs Execution time std-deviation : 38.022484 µs Execution time lower quantile : 52.197156 µs ( 2.5%) Execution time upper quantile : 150.862433 µs (97.5%) Overhead used : 1.289648 ns zilence@mx2:~/install/shadow-pgsql$ java -cp target/shadow-pgsql-0.8.0-SNAPSHOT-standalone.jar clojure.main -m shadow.pgsql.benchmark.runner 1 pgjdbc ===== pgjdbc via java amd64 Linux 2.6.32-5-amd64 8 cpu(s) Java HotSpot(TM) 64-Bit Server VM 25.60-b23 Runtime arguments: Evaluation count : 683880 in 60 samples of 11398 calls. Execution time sample mean : 86.035748 µs Execution time mean : 86.026940 µs Execution time sample std-deviation : 8.143039 µs Execution time std-deviation : 8.162961 µs Execution time lower quantile : 66.213094 µs ( 2.5%) Execution time upper quantile : 91.012787 µs (97.5%) Overhead used : 1.283264 ns Found 10 outliers in 60 samples (16.6667 %) low-severe 10 (16.6667 %) Variance from outliers : 66.9828 % Variance is severely inflated by outliers ----------------------------------------------------------------------------------- 10 rows ----------------------------------------------------------------------------------- jmc profiler flight recorder or jvisualvm profiler not good enough to find further performance issues should probably also use JMH, might be more accurate than criterium majority of time is IO though, but pgjdbc is still consistently 50usec faster ----------------------------------------------------------------------------------- zilence@mx2:~/install/shadow-pgsql$ java -cp target/shadow-pgsql-0.8.0-SNAPSHOT-standalone.jar clojure.main -m shadow.pgsql.benchmark.runner 10 java ===== shadow-pgsql via java WARNING: Final GC required 1.11641275631245 % of runtime amd64 Linux 2.6.32-5-amd64 8 cpu(s) Java HotSpot(TM) 64-Bit Server VM 25.60-b23 Runtime arguments: Evaluation count : 447480 in 60 samples of 7458 calls. Execution time sample mean : 146.153368 µs Execution time mean : 145.822312 µs Execution time sample std-deviation : 30.180719 µs Execution time std-deviation : 30.452910 µs Execution time lower quantile : 58.668432 µs ( 2.5%) Execution time upper quantile : 164.723268 µs (97.5%) Overhead used : 1.364173 ns Found 6 outliers in 60 samples (10.0000 %) low-severe 5 (8.3333 %) low-mild 1 (1.6667 %) Variance from outliers : 91.1649 % Variance is severely inflated by outliers zilence@mx2:~/install/shadow-pgsql$ java -cp target/shadow-pgsql-0.8.0-SNAPSHOT-standalone.jar clojure.main -m shadow.pgsql.benchmark.runner 10 pgjdbc ===== pgjdbc via java amd64 Linux 2.6.32-5-amd64 8 cpu(s) Java HotSpot(TM) 64-Bit Server VM 25.60-b23 Runtime arguments: Evaluation count : 741060 in 60 samples of 12351 calls. Execution time sample mean : 92.078353 µs Execution time mean : 91.958761 µs Execution time sample std-deviation : 10.647631 µs Execution time std-deviation : 10.704883 µs Execution time lower quantile : 79.493350 µs ( 2.5%) Execution time upper quantile : 104.790863 µs (97.5%) Overhead used : 1.301604 ns ----------------------------------------------------------------------------------- 100 rows ----------------------------------------------------------------------------------- same situation, as 10 ----------------------------------------------------------------------------------- zilence@mx2:~/install/shadow-pgsql$ java -cp target/shadow-pgsql-0.8.0-SNAPSHOT-standalone.jar clojure.main -m shadow.pgsql.benchmark.runner 100 java ===== shadow-pgsql via java WARNING: Final GC required 1.2957706321569629 % of runtime amd64 Linux 2.6.32-5-amd64 8 cpu(s) Java HotSpot(TM) 64-Bit Server VM 25.60-b23 Runtime arguments: Evaluation count : 219660 in 60 samples of 3661 calls. Execution time sample mean : 286.413909 µs Execution time mean : 286.386006 µs Execution time sample std-deviation : 9.174970 µs Execution time std-deviation : 9.200494 µs Execution time lower quantile : 268.894975 µs ( 2.5%) Execution time upper quantile : 294.696216 µs (97.5%) Overhead used : 1.540293 ns zilence@mx2:~/install/shadow-pgsql$ java -cp target/shadow-pgsql-0.8.0-SNAPSHOT-standalone.jar clojure.main -m shadow.pgsql.benchmark.runner 100 pgjdbc ===== pgjdbc via java amd64 Linux 2.6.32-5-amd64 8 cpu(s) Java HotSpot(TM) 64-Bit Server VM 25.60-b23 Runtime arguments: Evaluation count : 288360 in 60 samples of 4806 calls. Execution time sample mean : 222.998228 µs Execution time mean : 222.868850 µs Execution time sample std-deviation : 8.456342 µs Execution time std-deviation : 8.569100 µs Execution time lower quantile : 206.223484 µs ( 2.5%) Execution time upper quantile : 229.802610 µs (97.5%) Overhead used : 1.545911 ns Found 14 outliers in 60 samples (23.3333 %) low-severe 1 (1.6667 %) low-mild 13 (21.6667 %) Variance from outliers : 25.4140 % Variance is moderately inflated by outliers ----------------------------------------------------------------------------------- 1000 rows ----------------------------------------------------------------------------------- the more rows we get the more we pull ahead, not sure how or why, need better profiler ----------------------------------------------------------------------------------- zilence@mx2:~/install/shadow-pgsql$ java -cp target/shadow-pgsql-0.8.0-SNAPSHOT-standalone.jar clojure.main -m shadow.pgsql.benchmark.runner 1000 java ===== shadow-pgsql via java amd64 Linux 2.6.32-5-amd64 8 cpu(s) Java HotSpot(TM) 64-Bit Server VM 25.60-b23 Runtime arguments: Evaluation count : 104640 in 60 samples of 1744 calls. Execution time sample mean : 937.954996 µs Execution time mean : 934.023690 µs Execution time sample std-deviation : 198.838676 µs Execution time std-deviation : 200.583919 µs Execution time lower quantile : 514.750869 µs ( 2.5%) Execution time upper quantile : 1.111744 ms (97.5%) Overhead used : 1.542590 ns zilence@mx2:~/install/shadow-pgsql$ java -cp target/shadow-pgsql-0.8.0-SNAPSHOT-standalone.jar clojure.main -m shadow.pgsql.benchmark.runner 1000 pgjdbc ===== pgjdbc via java amd64 Linux 2.6.32-5-amd64 8 cpu(s) Java HotSpot(TM) 64-Bit Server VM 25.60-b23 Runtime arguments: Evaluation count : 49440 in 60 samples of 824 calls. Execution time sample mean : 1.196629 ms Execution time mean : 1.195805 ms Execution time sample std-deviation : 95.045787 µs Execution time std-deviation : 103.473317 µs Execution time lower quantile : 996.356451 µs ( 2.5%) Execution time upper quantile : 1.262201 ms (97.5%) Overhead used : 1.552005 ns Found 7 outliers in 60 samples (11.6667 %) low-severe 3 (5.0000 %) low-mild 1 (1.6667 %) high-mild 2 (3.3333 %) high-severe 1 (1.6667 %) Variance from outliers : 63.5305 % Variance is severely inflated by outliers