Adds full frame implementation to this node operation.
No functional changes.
3x times faster than tiled fallback.
Details
Details
Diff Detail
Diff Detail
- Repository
- rB Blender
- Branch
- cmp-node-ellipse-mask (branched from master)
- Build Status
Buildable 15290 Build 15290: arc lint + arc unit
Event Timeline
| source/blender/compositor/operations/COM_EllipseMaskOperation.cc | ||
|---|---|---|
| 148 | Best to move this outside the for loops. Perhaps use a lambda expression? | |
Comment Actions
- Merge branch 'master' into cmp-node-ellipse-mask
- Move switch statement out of for loops
Comment Actions
I've done performance tests comparing with tiled implementation fallback, added to main post, I use this patch:
| source/blender/compositor/operations/COM_EllipseMaskOperation.cc | ||
|---|---|---|
| 148 | I expected something to add a method that loops over the rows and columns that will use a lambda expression. update_memory_buffer_partial would than contain a switch statement that calls the new method with the specific expression. This would reduce the number of if statements in the inner loop. | |
Comment Actions
- Merge branch 'master' into cmp-node-ellipse-mask
- Add a method that loops applying the lambda expression