Closed toofishes closed 3 years ago
By avoiding as much 32-bit math and preferring 8-bit and 16-bit math, we can generate smaller code, even if more local variables are needed.
Before:
RAM: [======= ] 65.7% (used 1345 bytes from 2048 bytes) Flash: [========= ] 89.7% (used 29378 bytes from 32768 bytes)
After:
RAM: [======= ] 65.7% (used 1345 bytes from 2048 bytes) Flash: [========= ] 89.5% (used 29314 bytes from 32768 bytes)
This change saves 64 bytes of flash.
Note: I'm relatively confident I implemented identical logic between the old and new code, but I have no testbed hardware to verify this.
By avoiding as much 32-bit math and preferring 8-bit and 16-bit math, we can generate smaller code, even if more local variables are needed.
Before:
After:
This change saves 64 bytes of flash.
Note: I'm relatively confident I implemented identical logic between the old and new code, but I have no testbed hardware to verify this.