Improve performances: Remove the aggregateAsync call

[guizmaii]

To replicate the benchmark results:

sbt "zioKafkaBench/Jmh/run -wi 10 -i 10 -r 1 -w 1 -t 1 -f 5 -foe true zio.kafka.bench.ConsumersComparisonBenchmark"

The benchmark results BEFORE these changes (based on these changed: #749) are quite stable, and the average performance is quite good:

[info] # Fork: 5 of 5
[info] OpenJDK 64-Bit Server VM warning: -XX:ThreadPriorityPolicy=1 may require system level permission, e.g., being the root user. If the necessary permission is not possessed, changes to priority will be silently ignored.
[error] SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
[error] SLF4J: Defaulting to no-operation (NOP) logger implementation
[error] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
[error] SLF4J: Failed to load class "org.slf4j.impl.StaticMDCBinder".
[error] SLF4J: Defaulting to no-operation MDCAdapter implementation.
[error] SLF4J: See http://www.slf4j.org/codes.html#no_static_mdc_binder for further details.
[info] # Warmup Iteration   1: 2782.623 ms/op
[info] # Warmup Iteration   2: 2160.183 ms/op
[info] # Warmup Iteration   3: 2410.273 ms/op
[info] # Warmup Iteration   4: 2087.847 ms/op
[info] # Warmup Iteration   5: 1564.117 ms/op
[info] # Warmup Iteration   6: 1626.549 ms/op
[info] # Warmup Iteration   7: 2345.690 ms/op
[info] # Warmup Iteration   8: 2324.042 ms/op
[info] # Warmup Iteration   9: 2322.717 ms/op
[info] # Warmup Iteration  10: 1814.914 ms/op
[info] Iteration   1: 1800.236 ms/op
[info] Iteration   2: 1605.795 ms/op
[info] Iteration   3: 1887.400 ms/op
[info] Iteration   4: 1471.681 ms/op
[info] Iteration   5: 1265.121 ms/op
[info] Iteration   6: 1603.288 ms/op
[info] Iteration   7: 2387.570 ms/op
[info] Iteration   8: 1711.696 ms/op
[info] Iteration   9: 2411.515 ms/op
[info] Iteration  10: 2272.847 ms/op
[info] Result "zio.kafka.bench.ConsumersComparisonBenchmark.zioKafka":
[info]   1849.335 ±(99.9%) 185.495 ms/op [Average]
[info]   (min, avg, max) = (1170.867, 1849.335, 2703.178), stdev = 374.709
[info]   CI (99.9%): [1663.840, 2034.830] (assumes normal distribution)
[info] # Run complete. Total time: 00:06:39

The benchmark results AFTER these changes are quite unstable, but the peak performance is quite better:

[info] # Fork: 5 of 5
[info] OpenJDK 64-Bit Server VM warning: -XX:ThreadPriorityPolicy=1 may require system level permission, e.g., being the root user. If the necessary permission is not possessed, changes to priority will be silently ignored.
[error] SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
[error] SLF4J: Defaulting to no-operation (NOP) logger implementation
[error] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
[error] SLF4J: Failed to load class "org.slf4j.impl.StaticMDCBinder".
[error] SLF4J: Defaulting to no-operation MDCAdapter implementation.
[error] SLF4J: See http://www.slf4j.org/codes.html#no_static_mdc_binder for further details.
[info] # Warmup Iteration   1: 3443.871 ms/op
[info] # Warmup Iteration   2: 1770.033 ms/op
[info] # Warmup Iteration   3: 3526.743 ms/op
[info] # Warmup Iteration   4: 1888.485 ms/op
[info] # Warmup Iteration   5: 5033.633 ms/op
[info] # Warmup Iteration   6: 4413.383 ms/op
[info] # Warmup Iteration   7: 1503.804 ms/op
[info] # Warmup Iteration   8: 2806.786 ms/op
[info] # Warmup Iteration   9: 3364.834 ms/op
[info] # Warmup Iteration  10: 1269.350 ms/op
[info] Iteration   1: 686.590 ms/op
[info] Iteration   2: 3273.708 ms/op
[info] Iteration   3: 1413.702 ms/op
[info] Iteration   4: 1310.915 ms/op
[info] Iteration   5: 1307.911 ms/op
[info] Iteration   6: 2215.191 ms/op
[info] Iteration   7: 1435.898 ms/op
[info] Iteration   8: 1017.505 ms/op
[info] Iteration   9: 4297.892 ms/op
[info] Iteration  10: 4327.193 ms/op
[info] Result "zio.kafka.bench.ConsumersComparisonBenchmark.zioKafka":
[info]   1799.058 ±(99.9%) 515.830 ms/op [Average]
[info]   (min, avg, max) = (686.590, 1799.058, 5009.849), stdev = 1042.003
[info]   CI (99.9%): [1283.227, 2314.888] (assumes normal distribution)
[info] # Run complete. Total time: 00:07:31

Last run of the benchmarks with the latest version of this code:

[info] # Fork: 5 of 5
[info] OpenJDK 64-Bit Server VM warning: -XX:ThreadPriorityPolicy=1 may require system level permission, e.g., being the root user. If the necessary permission is not possessed, changes to priority will be silently ignored.
[error] SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
[error] SLF4J: Defaulting to no-operation (NOP) logger implementation
[error] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
[error] SLF4J: Failed to load class "org.slf4j.impl.StaticMDCBinder".
[error] SLF4J: Defaulting to no-operation MDCAdapter implementation.
[error] SLF4J: See http://www.slf4j.org/codes.html#no_static_mdc_binder for further details.
[info] # Warmup Iteration   1: 2049.754 ms/op
[info] # Warmup Iteration   2: 1481.757 ms/op
[info] # Warmup Iteration   3: 1578.479 ms/op
[info] # Warmup Iteration   4: 4836.508 ms/op
[info] # Warmup Iteration   5: 1980.513 ms/op
[info] # Warmup Iteration   6: 4928.249 ms/op
[info] # Warmup Iteration   7: 1472.769 ms/op
[info] # Warmup Iteration   8: 1263.501 ms/op
[info] # Warmup Iteration   9: 5014.016 ms/op
[info] # Warmup Iteration  10: 1258.164 ms/op
[info] Iteration   1: 1958.565 ms/op
[info] Iteration   2: 1735.329 ms/op
[info] Iteration   3: 1414.613 ms/op
[info] Iteration   4: 1224.270 ms/op
[info] Iteration   5: 1404.372 ms/op
[info] Iteration   6: 1236.202 ms/op
[info] Iteration   7: 3817.833 ms/op
[info] Iteration   8: 3835.710 ms/op
[info] Iteration   9: 1841.279 ms/op
[info] Iteration  10: 1332.895 ms/op
[info] Result "zio.kafka.bench.ConsumersComparisonBenchmark.zioKafka":
[info]   1636.533 ±(99.9%) 402.992 ms/op [Average]
[info]   (min, avg, max) = (783.816, 1636.533, 4721.287), stdev = 814.063
[info]   CI (99.9%): [1233.541, 2039.525] (assumes normal distribution)
[info] # Run complete. Total time: 00:07:09

[svroonland]

👍