Performance Tuning for Large Result Sets
Overview
The number of rows returned in a result set is not limited by the Andi application, but will be limited by the memory installed in your computer.
Andi Configuration File
If your computer has free memory, you can adjust the Andi options file to allow the Andi application to consume more of your system memory when needed. You can find the file in the Andi product installation directory, for example "C:\users\<userid>\JReflectionsLLC\Andi\".
The following snippet is an example of the default Java Virtual Machine (JVM) configuration. The -Xmx1024m value limits the amount of memory that the Andi application can allocate to 1 gigabyte.
-Xms512m
-Xmx1024m
#-XX:+UseG1GC -XX:MaxGCPauseMillis=200Increasing Available Memory
Increasing the -Xmx value to -Xmx2048m for example, allows Andi to use up to 2 gigabytes of memory while processing, allowing for larger result sets. Note that only 512 megabytes will be allocated initially, additional memory will be allocated as needed up to the new 2 gigabyte limit. This allows Andi to run alongside of other applications but to allocate large amounts of memory when needed.
-Xms512m
-Xmx2048m
#-XX:+UseG1GC -XX:MaxGCPauseMillis=200Tuning the Java Virtual Machine
The Java Virtual Machine has additional parameters which vary between Java Virtual Machine (JVM) providers. The following values were tested using the JVM shown under Limitations. Uncomment (remove the #) the line as shown. This change will reduce pauses which occur due to JVM cleanup operations and can dramatically increase the number or rows per second retrieved.
-Xms512m
-Xmx2048m
-XX:+UseG1GC -XX:MaxGCPauseMillis=200Limitations
Not all Java Virtual Machines will support the "-XX:+UseG1GC -XX:MaxGCPauseMillis=200" settings. If Andi will not start, remove the settings and submit a support request for help. Be sure to include your Java Runtime Information by clicking on the Help->About menu item and then click "Java Information". Copy the Java Runtime Information and include it with your support ticket. Please reference "Performance Tuning for Large Result Sets".