As of AVR-GCC 14.2.0, the code produced for unsigned 32-bit shifts with a compile time shift distance is very poor when using -O2 or -O3. E.g.

rpmDelta = (toothDeltaV << 10) / (6 * toothDeltaT);

The functions provided here implement optimised left and right shifting of uint32_t up to 31 places.

On a physical AtMega2560, up to:

• 35% increase in right shift performance
• 22% increase in left shift performance

See timing unit tests.

The library is available in both the Arduino Library and PlatformIO Library registries.

The library can also be cloned & included locally or included directly from GitHub (if your tooling supports it).

1. #include <avr-fast-shift.h>
2. Replace all 32-bit shift operations that use a compile time constant shift distance with the equivalent shift function. I.e.
   • a >> b -> rshift<b>(a)
   • a << b -> lshift<b>(a)
   • E.g.
   • rpmDelta = (toothDeltaV << 10) / (6 * toothDeltaT); -> rpmDelta = lshift<10U>(toothDeltaV) / (6 * toothDeltaT);