diff options
| author | Roman Musin <995612+roman-r-m@users.noreply.github.com> | 2024-01-27 14:17:55 +0000 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-01-27 15:17:55 +0100 |
| commit | f9c58414da2c647e6d7df5cf49aa21cd30684f38 (patch) | |
| tree | 8e6ccdf57fb2fe861f238b860aff39c0a76e0c4d /calculate_average_roman-r-m.sh | |
| parent | c228633b5753e2565e93833cf0ef8af23c66ac77 (diff) | |
Next version (#596)
* cleanup prepare script
* native image options
* fix quardaric probing (no change to perf)
* mask to get the last chunk of the name
* extract hash functions
* tweak the probing loop (-100ms)
* fiddle with native image options
* Reorder conditions in hope it makes branch predictor happier
* extracted constant
Diffstat (limited to 'calculate_average_roman-r-m.sh')
| -rwxr-xr-x | calculate_average_roman-r-m.sh | 21 |
1 files changed, 8 insertions, 13 deletions
diff --git a/calculate_average_roman-r-m.sh b/calculate_average_roman-r-m.sh index acf9864..5ba132f 100755 --- a/calculate_average_roman-r-m.sh +++ b/calculate_average_roman-r-m.sh @@ -15,21 +15,16 @@ # limitations under the License. # -JAVA_OPTS="--enable-preview -XX:+UseTransparentHugePages" - -# epsilon GC needs enough memory or it makes things worse -# see https://stackoverflow.com/questions/58087596/why-are-repeated-memory-allocations-observed-to-be-slower-using-epsilon-vs-g1 -JAVA_OPTS="$JAVA_OPTS -XX:+UnlockExperimentalVMOptions -XX:-EnableJVMCI -XX:+UseEpsilonGC -Xmx1G -Xms1G -XX:+AlwaysPreTouch" - if [ -f target/CalculateAverage_roman_r_m_image ]; then - echo "Picking up existing native image 'target/CalculateAverage_roman_r_m_image', delete the file to select JVM mode." 1>&2 + echo "Running native image 'target/CalculateAverage_roman_r_m_image'." 1>&2 target/CalculateAverage_roman_r_m_image else - JAVA_OPTS="--enable-preview -XX:+UnlockExperimentalVMOptions -XX:+TrustFinalNonStaticFields -dsa -XX:+UseNUMA" - if [[ ! "$(uname -s)" = "Darwin" ]]; then - # On OS/X, my machine, this errors: - JAVA_OPTS="$JAVA_OPTS -XX:+UseTransparentHugePages" - fi - echo "Choosing to run the app in JVM mode as no native image was found, use additional_build_step_roman_r_m.sh to generate." 1>&2 + JAVA_OPTS="--enable-preview -XX:+UseTransparentHugePages" + JAVA_OPTS="$JAVA_OPTS -XX:+UnlockExperimentalVMOptions -XX:+TrustFinalNonStaticFields -dsa -XX:+UseNUMA" + # epsilon GC needs enough memory or it makes things worse + # see https://stackoverflow.com/questions/58087596/why-are-repeated-memory-allocations-observed-to-be-slower-using-epsilon-vs-g1 + JAVA_OPTS="$JAVA_OPTS -XX:+UnlockExperimentalVMOptions -XX:-EnableJVMCI -XX:+UseEpsilonGC -Xmx1G -Xms1G -XX:+AlwaysPreTouch" + + echo "Running on JVM" 1>&2 java $JAVA_OPTS --class-path target/average-1.0.0-SNAPSHOT.jar dev.morling.onebrc.CalculateAverage_roman_r_m fi |