Topaz for Java Performance Webcast
Download as PPTX, PDF1 like666 views
The document outlines the advantages of using Topaz for Java Performance, emphasizing improvements in Java running on mainframes, including cost benefits and performance enhancements. It discusses features such as visibility into Java performance, pinpointing CPU usage issues, and scheduled measurements for performance management. Additionally, it highlights the integration with Strobe and support for Java batch and WebSphere applications.
Topaz for Java Performance Webcast
- 1. 1 Topaz™ for Java Performance Steve Kansa, Product Manager Spencer Hallman, Product Manager Glenn Everitt, Technical Consultant August 27, 2015
- 2. 2 Spencer Hallman Product Manager Steve Kansa Product Manager Glenn Everitt Technical Consultant
- 4. 4 Why Move to Java? • IBM provides special MIPS pricing for Java running on Specialty Processors • Many programmers know Java and prefer it to COBOL • Immediately benefit from performance hardware and JDK upgrade performance improvements Cost Improvements Innovations
- 5. 5 A Couple Things are Different in Java… • Java Program Classes are – Loaded only when they are used – May actually be compiled just before use • Memory management is automatic – Garbage Collection reclaims unused memory – JVM Heap size settings are important • CPU can include both: – General CPU – Specialty Processors And some things aren’t … • You can still write bad code
- 6. 6 Why Topaz for Java Performance? • Mainframe tools available today don’t allow you to see “inside” of Java Virtual Machines (JVMs) • Existing developer Java profiling tools aren’t mainframe-centric • Continuously monitoring application performance management tools can be overkill • Current tools don’t work well with Java batch programs
- 7. 7 “Under the hood” visibility into the performance of Java on the mainframe • Measuring Java Classes CPU performance • Highlighting areas for code optimization • Identifying blocked threads • Spotting memory issues • Integrating with Strobe for full mainframe performance analysis Topaz for Java Performance
- 8. 8 DEMO
- 9. 9 iStrobe to Topaz for Java Performance Use Topaz for Java Performance icon to open Java detail in Topaz
- 10. 10 iStrobe to Topaz for Java Performance Use menu Reports | Java | Topaz for Java Performance to open Java detail in Topaz
- 11. 11 Identify CPU Usage Issues • Quickly see if CPU spikes exist in your JVM Very high CPU usage
- 12. 12 Identify Java Class and Method Using Most CPU • Don’t want to look through 10,000 – 14,000 different Classes to find performance problem • Look … all my classes loaded later than program started
- 14. 14 JVM Performance at a Glance Very high CPU usage Very high rate of Garbage Collection Max Heap Memory limit hit Loaded Classes jumped from 10361 to 13902 Blocked Threads Threads jump from 60 to 102
- 15. 15 Very high CPU usage Very high rate of Garbage Collection Max Heap Memory limit hit Loaded Classes jumped from 10359 to 14341 Threads jump from 60 to 102
- 17. 17 JVM Threads Time Thread was blockedTime Thread was runnableWAS Liberty Large ThreadPool Group
- 18. 18 Highlight Acceptable Values within Thresholds Grey area denotes value within Threshold Boundary Values outside of boundary indicate Threshold Breach
- 20. 20
- 21. 21 All Remote JVM Arguments listed All Remote System Properties listed JVM Version really matters on z/OS Unix – it affects how much program runs on zAAP
- 22. 22 Easily find Class or Method using lots of CPU
- 23. 23 Key Advantages • Efficient and targeted – Runs on Specialty Engine – Scheduled measurements rather than continuous monitoring • Integrated with Strobe • One, central location to measure all JVMs across LPARs – Maintains performance history • Initial release supports Java batch and WebSphere – One tool for all types of Java z/OS Topaz for Java Performance
- 24. 24 CONNECT WITH US Glenn Everitt: glenn.everitt@compuware.com Spencer Hallman: spencer.hallman@compuware.com Steven Kansa: steve.kansa@compuware.com
- 25. 25 © 2015 Compuware Corporation. All rights reserved.