New plot script: Run duration over time #1760
Conversation
| with open(sys.argv[1]) as json_file: | ||
| data_json = json.load(json_file)['benchmarks'] | ||
|
|
||
| # Build flat list of {name, begin, duration} entries for every benchmark item run |
There was a problem hiding this comment.
As is evident, I am not the biggest Python hacker. I can make things work, but there are probably various things that could be improved here. Please let me know!
|
They add 5 to 15 order lines, which is why you should see 11 "lines" in there. Bear in mind that everything is still super inefficient.
Added. |
|
That looks already much better. If we create a collection of these over time, I am sure we can sell it as modern art at some point. |
|
Anything else that needs to be done here? |
|
Closing this for now as #1766 will change the JSON format. Will re-open and change to new JSON format once that is done. |
mrks
dismissed
Bensk1’s stale review
Please re-review the changes made in 12c7bcc
| data.append({'name': name, 'begin': begin / 1e9, 'duration': duration}) | ||
| for run_json in benchmark_json['successful_runs']: | ||
| data.append({'name': name, 'begin': run_json['begin'] / 1e9, 'duration': run_json['duration'], 'success': True}) | ||
| for run_json in benchmark_json['unsuccessful_runs']: |
There was a problem hiding this comment.
I'm not really happy with the code duplication here, but I have no idea how to avoid it without making everything more verbose.
There was a problem hiding this comment.
I am sure we can handle these two duplicated lines
Adds a script that plots the duration of individual benchmark runs over time. Example:
Detailed version: