Sound Process - Processor usage rises with time
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Open Rails |
Fix Released
|
High
|
Peter Gulyas |
Bug Description
Whilst running activities on the Dorset Coast route, I've recently noticed that with V0.9 and at least a few previous X releases that the Sound Process seems to steadily increase its processor usage with time. I've found that if you save the activity and then resume it the Sound Process usage will drop, sometimes quite markedly, but will then gradually increase again as time goes on. I've had drops from about 70% usage before the save to 20% upon resume. This will then slowly increase until you do the next save / resume.
I have a 4 core processor so the Sound Process is running on its own. The other processes have the same CPU usage before and after the save / resume.
I did a test with a DMU from which I had removed all sounds and started it off in explore mode (no AI) from Waterloo. The Sound Process started at 40% and rose to 50% at Wimbledon ( about 10 mins ) where I did a save and resume. At resume the Sound Process was running at 20% and rose to 55% at Woking ( about another 15 mins ) where I once again saved and resumed. This time the Sound Process resumed at 15% and hardly changed for about another 5 mins when I stopped the test.
Another test was to restore the sounds to the DMU and run it for 50 mins on a section with no environmental sounds. I know there aren't any because I've constructed this section myself and haven't yet put any in. The test was in explore mode, once again no AI. For the whole of this test the Sound Process registered 0% usage although all train sounds were being clearly produced.
It therefore seems to me that the bulk of the upwards creep of processor usage is due to environmental and zoned sounds of which there are a very high number in the Dorset Coast route.
Changed in or: | |
assignee: | nobody → Peter Gulyas (pzgulyas) |
tags: | added: sounds |
Changed in or: | |
status: | New → Triaged |
importance: | Undecided → High |
status: | Triaged → In Progress |
Changed in or: | |
milestone: | none → 1.0 |
Changed in or: | |
status: | Fix Committed → Fix Released |
Dennis,
Could you please test r1668 (or higher) if the issue is still alive? I made some modifications, but there still might be some more mines in the code field.