Difference between revisions of "Slicer/Features/Events and communication"
From Slicer Wiki
(One intermediate revision by the same user not shown) | |||
Line 31: | Line 31: | ||
* Louis: in RTOS this would be taken care | * Louis: in RTOS this would be taken care | ||
* Steve P: but also rendering and other should be separate processes (in sens of thread), so the OS can kill it | * Steve P: but also rendering and other should be separate processes (in sens of thread), so the OS can kill it | ||
− | * S Barre: on vista our rendering gets | + | * S Barre: on vista our rendering gets killed by the OS if it takes more than 2 sec (it's a bug?) |
* Louis: can we just have one thread (currently 2 threads: tracker update, render update) | * Louis: can we just have one thread (currently 2 threads: tracker update, render update) | ||
* Steve P: No. But it can add an event. Or a socket to have an event loop (FIFO) - already in the Slicer daemon. | * Steve P: No. But it can add an event. Or a socket to have an event loop (FIFO) - already in the Slicer daemon. |
Latest revision as of 16:12, 19 May 2008
Home < Slicer < Features < Events and communicationContents
Title
Events, communication and data stream
Names
- Noby Hata
- Junichi Tokuda
- Steve Pieper
- Stephen Aylward
- Haiying Liu
- Sebastien Barre
- Louis Ibanez
- Peter Kazanzides
- Tina Kapur
- Csaba Csoma
- David Gobbi
- Ron Kikinis
- Patrick Cheng
- Ziv Yanev
Technology/Function description
Discussion
Junichi Tokuda: IGTDataStream / IGT_IGSTK_Stream presentation
- Steve P: kwWidgets - timer depends on TCL GUI (same thread).
- Noby: What is the ideal solution for this timer?
- Steve P: the way it is implemented now it's single threaded cooperative
- David: the VTK Render model also is good for just GUI (periodic abort check)
- Steve P: Consistent way of handling frame rate, level of detail etc. Progress events
- Louis: in RTOS this would be taken care
- Steve P: but also rendering and other should be separate processes (in sens of thread), so the OS can kill it
- S Barre: on vista our rendering gets killed by the OS if it takes more than 2 sec (it's a bug?)
- Louis: can we just have one thread (currently 2 threads: tracker update, render update)
- Steve P: No. But it can add an event. Or a socket to have an event loop (FIFO) - already in the Slicer daemon.
- Liu: what happens if the queue gets full?
Potential application in IGT
- Timer event: periodic update from tracker / robots / devices