Currently and optimization exists to only copy over the dynamic game information:
|
// These blocks account for *most* of the 33MB shared memory, |
However we still allocate a full (33MB) buffer of which we only use parts because of that optimization.
This can be reduced by pointing calls to the static information to the "main" buffer and only allocating as much as needed for the dynamic information (which is only 4-5MB).
Currently and optimization exists to only copy over the dynamic game information:
JBWAPI/src/main/java/bwapi/FrameBuffer.java
Line 182 in f854bcf
However we still allocate a full (33MB) buffer of which we only use parts because of that optimization.
This can be reduced by pointing calls to the static information to the "main" buffer and only allocating as much as needed for the dynamic information (which is only 4-5MB).