Age | Commit message (Collapse) | Author |
|
|
|
|
|
Global; except for the 32-bit SSE blitter, which has some #DEFINEs
in not-very-nice places.
|
|
|
|
|
|
This reverts commit 25ab9c1997f770f4a8a66bb3ad4b82ba87e3a977.
|
|
|
|
When zooming out with a high res display, there can be about 150k sprites
to be sorted before displaying. With the O(n^2) complexity of the sprite
sorter, this can take several seconds.
This patch works around this by sorting the sprites by the xmin coordinate
first using QSort, which later allows an early bailout out of the inner
loop. This is enough to cut down the full unzoom time on a 4k display to a
fraction of second.
|
|
"unsafe" functions to prevent them from being used, and thus having to care about certain aspects of their return values
|
|
|
|
performance significantly (MJP)
For example with GCC 4.8, x86_64 Linux, Intel i5-3337U this patch improves the performance of Pile, Treham and Hamac test save games by about 10% in over-all run time at fast forward at 1920x1080 when zoomed out and when trees are not disabled.
|