Stage 2 was all about profiling our project and get an idea of the usage that it is taking the software. To refresh I am working with zopfli which is an open-source from Google.
As I already mentioned in my post I already did some of the profiling during my Stage 1, but continue to do so with another command which is perf. During stage 1 the command that I used to do the profiling was gprof.
Both commands are different but the results that I gather were very similar. With gprof ZopfliFindLongestMatch, BoundaryPM, encodeLZ77and ZopfliUpdateHash were the ones that took the most time when I ran my test images. Then when I used the perf command and I got the same results. All four files took more than 15% usage time when I perform the test with the 1mb and 5mb image.
I decided to experiment with ZopfliFindLongestMatch since it is the one that in both cases take more than 15% of usage time. I will be posting the results that I gather during my Stage 3 blogpost.
Overall the Stage 2 of this part of the project was relatively easier and less stressful than Stage 1. I think stage 1 took me the most time since I needed to find a project that I like and feel comfortable working with.
Personal Note:
I did encounter some problems while trying to profile the project. In the beginning, I was working with the Aarch64 Israel server, but this server has a weird kernel and did not let me perform the perf command even if I try to download the tools.
I decided to move that part of the project to the bbetty server and everything work well there. I had to take some time to git clone and build the project on that server again. But the results were similar to when I was working on the Israel server.
Comentarios