573
edits
Troubleshooting performance issues: Difference between revisions
Jump to navigation
Jump to search
→For the geeks out there: moved tools for performance evaluation from Howto:Debugging FlightGear Crashes
m (fix links, +cat) |
(→For the geeks out there: moved tools for performance evaluation from Howto:Debugging FlightGear Crashes) |
||
| Line 81: | Line 81: | ||
=== FPS and frame spacing === | === FPS and frame spacing === | ||
Frames per second is an average by definition, and as such is not an optimal indicator of performance, because you could have half a second with 100 FPS and the other half with 0 FPS. You'd get a respectable 50 FPS but an unacceptable stutter of 0.5 seconds. A much better one is keeping track of the maximum frame spacing, that gives a measure in milliseconds of the worst performance during a fixed interval. | Frames per second is an average by definition, and as such is not an optimal indicator of performance, because you could have half a second with 100 FPS and the other half with 0 FPS. You'd get a respectable 50 FPS but an unacceptable stutter of 0.5 seconds. A much better one is keeping track of the maximum frame spacing, that gives a measure in milliseconds of the worst performance during a fixed interval. | ||
=== Tools for performance evaluation === | |||
The performance impact can be evaluated by using: | |||
* The frame rate counter (frames created per second) | |||
* The frame spacing indicator (latencies between rendered frames) | |||
* The [[Howto:Use the system monitor|built-in system monitor]] (can be inspected via telnet/http using [[Telnet usage|--telnet=5900]]) | |||
* The OSG on-screen stats | |||
* The [[Built-in Profiler]] | |||
=== Some more details on the inner workings === | === Some more details on the inner workings === | ||