action #14976
closed
action #14804: encoder still stalls
coordination #14972: [tools][epic] Improvements on backend to improve better handling of stalls
Change the isotovideo backend to write ppm files
Added by szarate about 8 years ago.
Updated about 8 years ago.
Category:
Feature requests
Description
User story¶
As a user of isotovideo, i want the screenshots to be written in ppm, to reduce peaks in write times
acceptance criteria¶
- The isotovideo writes ppm to be used by the video encoder instead of png
- The isotovideo will update last.png when the worker has explicitly requested this.
Tasks¶
- Description updated (diff)
- Checklist item changed from to [x] Have the worker request last.png to the videoencoder so that becomes on demand
- Checklist item changed from to [ ] Have the worker request last.png to the videoencoder so that becomes on demand
- Checklist item changed from to [x] Have the videoencoder to write last.png
- Checklist item changed from to [x] Have the worker request last.png to the videoencoder so that becomes on demand
- Checklist item changed from to [x] Add debug timing marks so we can detect bottlenecks or peaks easier
- Checklist item changed from to [ ] Add debug timing marks so we can detect bottlenecks or peaks easier
- Checklist item changed from to [x] Change tinycv::write_img to support multiple formats
- Waiting for Several changes to the image handling to be merged which has my changes for tasks 2 and 3.
- Task 4 is already implemented by: Have the worker to touch live_log file on demand
- Task 1 is being developed at feature/poo#14976/timing_marks but will wait until changes done to baseclass.pm from pr 664 are merged to add timing to parts of the code there.
- Task 3, is was easily implemented by changing the extension of the screenshot and by pure luck, optipng doesn't care about the filetype to convert to png the images for the testresults.
- Task 5 should be an update soon to the videoencoder
When the aforementioned PR is merged, AC 1 and 2 of this ticket, will be sucessfully met (AC2 is already met by task 4).
- Checklist item changed from [ ] Add debug timing marks so we can detect bottlenecks or peaks easier, [x] Change tinycv::write_img to support multiple formats, [x] Have the videoencoder to write last.png, [x] Have the worker request last.png to the videoencoder so that becomes on demand to [x] Add debug timing marks so we can detect bottlenecks or peaks easier, [x] Change tinycv::write_img to support multiple formats, [x] Have the videoencoder to write last.png, [x] Have the worker request last.png to the videoencoder so that becomes on demand
- Status changed from In Progress to Resolved
the PR is merged - including the timemarks. We could use more of these - but I leave that to issue 14972
Also available in: Atom
PDF