Week 5 of GSoC is finished, the communication was hard this week since some of the mentors were travelling, but being able to work without weekly guidance is a skill to learn.
Updates from Veronica, our GSoC admin and mentor
We are moving forward and learning a lot during this first GSoC. There have been some challenges from both sides. Slips is a tool that was created over 10 years ago, and it has been growing and maturing, but this is the first time we have so many contributors working on it at the same time - as part of GSoC and from our community. This brings some challenges for our team which needs to adjust their time and methodologies to make sure we all work harmoniously to make a better tool.
Updates from Daniel, our Slips Performance Contributor
This week, Alya introduced some changes to the modules which should be able to resolve the termination issue from viztracer. I attempted to run her merged changes on my branch but there were some discrepancies between the variable names and code section ordering so I had to resolve those conflicts. After resolving those conflicts, I attempted to run my cpu profiling feature and update_manager would not terminate even after running for 30 minutes. I tried again with profiling off and the issue would still happen, though inconsistently. This is normally no issue since you can Ctrl-C to stop the update_manager however pressing those keybinds while running viztracer will cause the profiling to terminate and the runtime data to be lost. I reverted the changes and attempted to merge again, carefully reading through each of the merge conflicts to make sure they were resolved correctly. However, this time a new problem emerged and the program output was getting directed to a file called “pipe” which was really strange and I didn’t know how to resolve it. After this, I decided to keep the cpu profiler feature in its current state where modules had to be disabled during profiling since I couldn’t figure out how to merge the changes to my branch. I then started researching memory profilers for the next portion of the project. I found some promising ones like memray, memory-profiler, pympler, and guppy3. I would have to do some further testing to check which ones are needed for live mode profiling and dev mode profiling.
Updates from Shubhangi, our Slips Web UI Contributor
Finally, I'm done with the designs of the 2 screens and have landed on a perfect library for visualization as well. The designs needed some corrections and the new designs are to be completely reviewed as well. Post this week, I'll start on coding the screens.
Before you go…
We have just published our video presentation of Slips at the last BlackHat Asia Arsenal, you can watch it here: