How Tweaking IntelliJ IDEA Memory Settings Supercharges Java IDE Performance

Goal

The author plans to test different IntelliJ IDEA memory configurations in a realistic development scenario (loading a large monolith project, several micro‑service modules, and refreshing after git pull) and pick the setting that gives the best balance of memory consumption and speed.

Test machine and projects

MacBook Pro Retina, 2.3 GHz Intel Core i7, 16 GB RAM, SSD, OS X Yosemite

Monolith project: ~700 k lines of Java 8/Groovy code, 303 Gradle modules

Two micro‑service projects: each 10‑20 k lines of Java 8/Groovy, single Gradle module

Test scenario

Close all projects in IDEA

Apply a specific idea.vmoptions file

Reboot the computer

Close unrelated processes

Start IDEA and record launch time

Open the monolith and record load time

Run jstat -gcutil to capture GC metrics

Open the two micro‑service projects and record load time

Run jstat -gcutil again

Refresh the monolith (Gradle refresh) and record time

Run jstat -gcutil once more

jstat -gcutil

jstat is a JDK tool that reports real‑time JVM statistics. The -gcutil option prints utilization percentages for survivor spaces (S0, S1), Eden (E), Old (O), Metaspace (M), Compressed Class Space (CCS) and counts/times for young‑generation GC (YGC, YGCT) and full GC (FGC, FGCT).

Test settings

Four configurations were created:

Default (gray)

-Xms128m
-Xmx750m
-XX:MaxPermSize=350m
-XX:ReservedCodeCacheSize=240m
-XX:+UseCompressedOops

Big (red)

-Xms1024m
-Xmx4096m
-XX:ReservedCodeCacheSize=1024m
-XX:+UseCompressedOops

Balanced (blue)

-Xms2g
-Xmx2g
-XX:ReservedCodeCacheSize=1024m
-XX:+UseCompressedOops

Sophisticated (orange)

-server
-Xms2g
-Xmx2g
-XX:NewRatio=3
-Xss16m
-XX:+UseConcMarkSweepGC
-XX:+CMSParallelRemarkEnabled
-XX:ConcGCThreads=4
-XX:ReservedCodeCacheSize=240m
-XX:+AlwaysPreTouch
-XX:+TieredCompilation
-XX:+UseCompressedOops
-XX:SoftRefLRUPolicyMSPerMB=50
-Dsun.io.useCanonCaches=false
-Djava.net.preferIPv4Stack=true
-Djsse.enableSNIExtension=false
-ea

Results

IDEA startup time

All configurations took about 10 seconds to start; memory settings did not affect launch time.

Loading the large monolith

The default configuration was roughly three times slower than the custom ones. GC activity was much higher, especially Full GC time.

Opening two micro‑services

The sophisticated setting performed best, while the default lagged behind.

Refreshing the monolith

The default configuration caused IDEA to crash during Gradle refresh, confirming insufficient memory. The “Big” setting gave the shortest refresh time.

Final jstat observations

Full GC execution time was lowest with the “Big” configuration; larger Xmx values clearly improved responsiveness.

Conclusion

Even modest adjustments to IntelliJ IDEA’s JVM memory options can dramatically improve IDE responsiveness. For most projects, an Xmx between 2 GB and 3 GB offers a good trade‑off between speed and memory consumption. Further tuning can be explored with jstat or jvisualvm.

Discussion

Readers are invited to share their own idea.vmoptions and additional performance tips.