Module KMath-Benchmarks
BenchmarksResult
Report for benchmark configuration main
- Run on OpenJDK 64-Bit Server VM (build 21.0.9+10-LTS) with Java process:
- JMH 1.37 was used in
thrpt mode with 5 warmup iterations by 10 s and 5 measurement iterations by 10 s.
| Benchmark |
Score |
benchmarkArrayRead |
3.9E+06 ± 1.1E+06 ops/s |
benchmarkBufferRead |
4.0E+06 ± 2.2E+05 ops/s |
nativeBufferRead |
4.0E+06 ± 1.7E+05 ops/s |
| Benchmark |
Score |
jvmAdd |
2.9E+07 ± 2.8E+06 ops/s |
jvmAddLarge |
3.8E+04 ± 6.4E+03 ops/s |
jvmMultiply |
5.3E+07 ± 6.1E+06 ops/s |
jvmMultiplyLarge |
2.2E+02 ± 1.9 ops/s |
jvmParsing10 |
3.9E+06 ± 4.7E+05 ops/s |
jvmParsing16 |
3.1E+06 ± 4.6E+05 ops/s |
jvmPower |
24 ± 1.7 ops/s |
jvmSmallAdd |
4.7E+07 ± 4.6E+06 ops/s |
kmAdd |
2.3E+07 ± 5.1E+06 ops/s |
kmAddLarge |
2.6E+04 ± 3.0E+02 ops/s |
kmMultiply |
3.7E+07 ± 2.9E+06 ops/s |
kmMultiplyLarge |
34 ± 2.8 ops/s |
kmParsing10 |
2.5E+06 ± 1.5E+05 ops/s |
kmParsing16 |
4.0E+06 ± 2.4E+05 ops/s |
kmPower |
6.5 ± 0.69 ops/s |
kmSmallAdd |
1.6E+07 ± 8.0E+05 ops/s |
| Benchmark |
Score |
bufferViewReadWrite |
5.4E+06 ± 3.8E+05 ops/s |
bufferViewReadWriteSpecialized |
5.0E+06 ± 1.2E+06 ops/s |
complexBufferReadWrite |
2.2E+06 ± 5.7E+04 ops/s |
doubleArrayReadWrite |
6.9E+06 ± 1.2E+06 ops/s |
doubleBufferReadWrite |
6.6E+06 ± 1.1E+06 ops/s |
| Benchmark |
Score |
bufferedDot |
1.2 ± 0.20 ops/s |
cmDot |
0.36 ± 0.14 ops/s |
cmDotWithConversion |
0.80 ± 0.092 ops/s |
ejmlDot |
2.9 ± 0.61 ops/s |
ejmlDotWithConversion |
2.7 ± 0.15 ops/s |
multikDot |
23 ± 2.4 ops/s |
ojalgoDot |
11 ± 0.79 ops/s |
parallelDot |
9.4 ± 1.3 ops/s |
tensorDot |
1.0 ± 0.15 ops/s |
tfDot |
3.9 ± 0.90 ops/s |
| Benchmark |
Score |
asmGenericExpression |
15 ± 1.8 ops/s |
asmPrimitiveExpression |
27 ± 0.98 ops/s |
asmPrimitiveExpressionArray |
78 ± 14 ops/s |
functionalExpression |
4.4 ± 0.25 ops/s |
justCalculate |
79 ± 5.4 ops/s |
mstExpression |
4.2 ± 0.93 ops/s |
rawExpression |
25 ± 5.0 ops/s |
| Benchmark |
Score |
complexIntegration |
2.2E+03 ± 3.0E+02 ops/s |
doubleIntegration |
2.3E+03 ± 6.4E+02 ops/s |
| Benchmark |
Score |
cmLUPInversion |
2.0E+03 ± 1.1E+02 ops/s |
ejmlInverse |
1.2E+03 ± 29 ops/s |
kmathLupInversion |
3.9E+02 ± 92 ops/s |
kmathParallelLupInversion |
55 ± 5.0 ops/s |
ojalgoInverse |
1.7E+03 ± 35 ops/s |
| Benchmark |
Score |
kotlinArrayMin |
1.6E+03 ± 3.0E+02 ops/s |
minBlocking |
1.2E+03 ± 1.2E+02 ops/s |
| Benchmark |
Score |
boxingFieldAdd |
1.9 ± 0.089 ops/s |
multikAdd |
6.8 ± 1.0 ops/s |
multikInPlaceAdd |
32 ± 4.7 ops/s |
specializedFieldAdd |
6.7 ± 0.98 ops/s |
tensorAdd |
7.9 ± 1.1 ops/s |
tensorInPlaceAdd |
11 ± 3.4 ops/s |
viktorAdd |
6.4 ± 0.41 ops/s |
| Benchmark |
Score |
doubleFieldAddition |
7.3 ± 1.1 ops/s |
rawViktor |
6.0 ± 0.88 ops/s |
viktorFieldAddition |
6.7 ± 0.47 ops/s |
| Benchmark |
Score |
rawViktorLog |
1.3 ± 0.40 ops/s |
realFieldLog |
1.2 ± 0.34 ops/s |
viktorFieldLog |
1.3 ± 0.0073 ops/s |
