# Parallel Processing
In heap dump analysis, the heap report includes two separate sections: the Troubleshooting Report and the Performance Report.
By default, the tool processes both sections in sequential order. First, it analyses your heap dump file to get the data for the Troubleshooting Report and then starts analyzing for the Performance Report. If you'd like to process both sections simultaneously, you can enable parallel processing for your heap dump analysis.
Keep in mind that if you opt for parallel parsing, the tool will use more memory as it will analyze both sections parallelly.
Note:
Please note that the parallel processing feature doesn't apply when analyzing your heap dump file using the REST API. Internally, the tool uses two separate threads for analyzing your heap dump concurrently, generating both troubleshooting and performance reports.
If you make a REST API request using tools like cURL or POSTMAN, these threads start their work at the same time. Each thread sends a response as soon as it finishes its task, without waiting for the other thread. So, if one thread completes first, the client will receive its response immediately, and it won't wait for the other thread.
If you're interested in a specific report, like troubleshooting or performance, you can use flags like onlyTroubleshootingReport and onlyPerformanceReport to instruct the tool to analyze your heap dump for that particular report.
# yCrash Enterprise Edition
Here are the steps to configure the flag in yCrash Enterprise Edition:
Open the launch file: If you're using a Unix-based machine, open the
launch-yc-server.shfile. For Windows machines, use thelaunch-yc-server.batfile.Add flag: Modify your launch file to include the
-DhdParallelParsing=truesystem property. For example,java -Xms2g -Xmx4g -DlogDir=. -DuploadDir=. -DhdParallelParsing=true ...1Restart yCrash server: Save the changes to your launch file and restart the yCrash server to apply the updates.
# yCrash Cloud
Here are the steps to configure the flag in yCrash Cloud:
Coming Soon…