Difference between revisions of "2012 Summer Project Week:EventOptimization"
From NAMIC Wiki
(Created page with '__NOTOC__ <gallery> Image:PW-MIT2012.png|Projects List </gallery> ==Key Investigators== * Isomics: Steve Pieper, Alex Yarmarkovich * Kitwa…') |
|||
Line 29: | Line 29: | ||
<h3>Progress</h3> | <h3>Progress</h3> | ||
+ | * Discussions of approach | ||
+ | ** Identify end-user cases where performance is critical: | ||
+ | *** OpenIGTLink, Editor, Startup, MultiVolume, CompareView... | ||
+ | *** Develop reproducible tests and profiling tools (probably Apple Instruments) | ||
+ | *** Identify/Fix hot spots | ||
+ | *** Refine code and re-run tests | ||
+ | ** Work on improved design to help developers create efficient and reliable code | ||
+ | *** Avoid duplication of code by moving event management to new superclasses | ||
+ | *** Review lowest level code implementations for maximum event efficiency | ||
+ | *** Review usage patters to see if functionality can be maintained with fewer event invokations | ||
Latest revision as of 11:53, 22 June 2012
Home < 2012 Summer Project Week:EventOptimization
Key Investigators
- Isomics: Steve Pieper, Alex Yarmarkovich
- Kitware: Julien Finet
Objective
Investigate performance bottlenecks in slicer and develop strategies for fixing them.
Approach, Plan
Our goal will be to test first and then develop solutions based on the evidence.
We will evaluate approaches to profiling the existing code. So far, the Time Profiler in the Apple Instruments Developer Tool has shown great potential for identifying the issues.
Progress
- Discussions of approach
- Identify end-user cases where performance is critical:
- OpenIGTLink, Editor, Startup, MultiVolume, CompareView...
- Develop reproducible tests and profiling tools (probably Apple Instruments)
- Identify/Fix hot spots
- Refine code and re-run tests
- Work on improved design to help developers create efficient and reliable code
- Avoid duplication of code by moving event management to new superclasses
- Review lowest level code implementations for maximum event efficiency
- Review usage patters to see if functionality can be maintained with fewer event invokations
- Identify end-user cases where performance is critical:
Delivery Mechanism
This work will be delivered to the NA-MIC Kit as a (please select the appropriate options by noting YES against them below)
- ITK Module
- Slicer Module - NO
- Built-in - YES (as improved behavior)
- Extension -- commandline
- Extension -- loadable
- Other (Please specify)