https://bugs.kde.org/show_bug.cgi?id=523100
--- Comment #2 from Fred M <[email protected]> --- Hi Paul, The part I find time-consuming is the manual comparison between two Massif snapshots. My current workflow is roughly: Open the massif.out file in KCachegrind. Select the latest snapshot. Select a reference snapshot (for example after 5 minutes, once the application has reached a steady state). Compare the top allocators in both snapshots. Identify the allocation tree whose memory usage has increased the most. Expand that node in both snapshots and compare them again. Repeat this process recursively until I finally reach the allocation call stack responsible for the increase. This process is quite repetitive and requires manually expanding and comparing many nodes, which also makes it easy to miss the branch with the largest growth. I was wondering whether it would be possible to automate this analysis by producing: the full allocation call stack corresponding to the largest increase between two snapshots; and ideally, highlighting the point in the call stack where the increase starts to appear. The goal is not to replace Massif or Memcheck, but to automate the manual comparison that users currently have to perform. -- You are receiving this mail because: You are watching all bug changes.
