How to push the most performance possible out of Qt applications
summary
Shows concrete patterns and measured wins: thumbnails are generated with QtConcurrent::mapped so UI stays responsive while background threads scale multi-megabyte images, with the model only updated on the main thread to avoid mutexes; a bad example shows a race when workers touch the hash. A moveToThread controller-worker pattern queues long blocking jobs. Huge batch jobs drop from 25 hours to 3 hours by using plain-old arrays with cache-aligned chunks and per-thread data slices.
AI-written summary · about
chapters (9)
- 0:00Introduction and speaker background
- 1:09Performance challenges in Qt applications
- 3:39Background processing for smooth UI
- 8:37Thread-safe data handling in Qt
- 11:06Controller-worker pattern explained
- 13:17Optimizing CPU cache usage
- 15:28Data processing with filters
- 17:15Simplicity over complexity in performance
- 18:57When to avoid threading in Qt
related talks
01h 00mCreating multithreaded applications with Qt
01h 04mAtomics, Locks, and Tasks (part 2 of 2)
01h 01mAdvanced Qt - A Deep Dive 5/6 - ModelView Part 1
01h 21mAdvanced Qt - A Deep Dive 6/6 - ModelView Part 2
49mWriting better code with help from Qt and the compiler
30mDesigning UIs with Qt Quick, 2/6, Design-Workflow Part 2
01h 19mGetting Started - a Qt fast track 4/4 - Application Creation
28mImplementing Modern Design Patterns with Qt