DOC: add short inline summary for Mar / Apr Developer Meetings
..some conclusions from Scheduler performance testing
This commit is contained in:
parent
f7918bbfa3
commit
0d5f95704f
1 changed files with 19 additions and 0 deletions
|
|
@ -11,6 +11,25 @@ spending time on the task of writing a summary.
|
|||
Anyone interested in Lumiera development is also encouraged to read mailing list
|
||||
archives and other documentation.
|
||||
|
||||
13 Mar / 10 Apr 2024
|
||||
--------------------
|
||||
Status update and informal exchange regarding current development activities
|
||||
|
||||
* after factoring out further external influences, a _breaking point_ for
|
||||
processing complex dependent schedules could be established, which closely
|
||||
matches theoretical expectations
|
||||
* a further measurement series (linear regression) was necessary to establish
|
||||
performance characteristics of the scheduler under constant high load
|
||||
* as expected, a serious weakness pertaining contention could be confirmed +
|
||||
(yet contention is not expected to be a problem in practice)
|
||||
* as mitigation, concurrency is scaled down on very short (100µs) jobs
|
||||
* overall we are confident that the new scheduler is capable to handle the
|
||||
expected load pattern and does not suffer from major (unknown) weaknesses.
|
||||
|
||||
Moreover, the problems related to integration and use of external frameworks
|
||||
for video display were discussed further.
|
||||
|
||||
|
||||
14 Feb 2024
|
||||
-----------
|
||||
|
||||
|
|
|
|||
Loading…
Reference in a new issue