asmhatre/zgc-jdk25-benchmarks
1
0
Fork 0
Code Activity

Add jdk27-memory: compact object headers and G1-everywhere on a Spring Boot service

Browse Source 
Spring Boot 4.1.1 catalog service measured on JDK 25, 26 and 27: live heap 320 MB to 273 MB with compact headers, per-class bytes per instance, container-sizing sweeps with real cgroup limits, 1-CPU Serial vs G1 with Native Memory Tracking, removed flags (MaxRAM, UseCompressedClassPointers), and a kernel OOM-kill check. Outputs in jdk27-memory/output.

Co-Authored-By: Claude Sonnet 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01TF9JWFvJSNm6HVzswzZU5a
This commit is contained in:
Claude
2026-09-24 15:09:55 +00:00
parent e189da79ba
commit e83c9949d8
29 changed files with 1041 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
jdk27-memory/scripts/classes.sh
Executable
+19
View File
@@ -0,0 +1,19 @@
#!/usr/bin/env bash
# Bytes per instance of the classes that dominate the heap, from GC.class_histogram of the runs made by
# footprint.sh (kept in target/hist). The point: which classes shrink, and which do not.
source "$(dirname "$0")/lib.sh"
H="$MOD/target/hist"
echo "== bytes per instance (= #bytes / #instances) from jcmd GC.class_histogram after loading 2,000,000 products"
printf '%-40s %14s %14s %14s\n' "class" "JDK 26 default" "JDK 27 default" "JDK 27 opt-out"
for cls in 'com.ankurm.memory.Product' 'java.util.HashMap$Node' 'java.lang.Long' 'java.lang.String' '[B'; do
printf '%-40s' "$cls"
for cfg in jdk26-default jdk27-default jdk27-optout; do
awk -v c="$cls" '$4==c {printf " %14.1f", $3/$2; found=1} END{if(!found) printf " %14s","?"}' "$H/$cfg.txt"
done
echo
done
echo
echo "== total bytes of those five classes"
for cfg in jdk26-default jdk27-default jdk27-optout; do
awk -v cfg="$cfg" '$4=="com.ankurm.memory.Product"||$4=="java.util.HashMap$Node"||$4=="java.lang.Long"||$4=="java.lang.String"||$4=="[B" {s+=$3} END{printf "%-16s %d bytes (%.1f MB)\n", cfg, s, s/1048576}' "$H/$cfg.txt"
done