0

In our applciation we use LMAX-Distuptor for pulishing events. It shows the Distuptor code is doing some intense CPU operation just to calculate the minimum sequence from an array of sequences, which takes most of the CPU. Click the below URL to fetch the JFR hot thread snap.

https://screenshot.net/ez5zztv

The ring buffer size we use is big, 2^16. and we use PhasedBackoffWaitStrategy as wait strategy.

Kindly put your suggestion on how to optimize this, to avoid this high cpu utilization.

Dhiman Ghosh
  • 57
  • 3
  • Your screenshots omits the most important thing, what is the main contributor to the hot method. Post the full picture, not just the few top rows. – Karol Dowbecki Feb 18 '19 at 14:22
  • @KarolDowbecki, Here are two more screenshots that should show in details: https://screenshot.net/r3rxguo https://screenshot.net/qx436u7 – Dhiman Ghosh Feb 18 '19 at 16:11

0 Answers0