Firefox bitwise operation speed?

Post by ya » Sat, Apr 13, 2013, 6:35 AM UTC

Post by ya » Sat, Apr 13, 2013, 4:32 PM UTC

kevingadd wrote: post #2

It's called Loop Invariant Code Motion, and it's why jsperf benchmarks are mostly inaccurate. A smart JIT can hoist all the 'work' in a benchmark out of the benchmark loop and only run it once.