Is wy2u01 mistakenly not making use of all 53 bits?

[redoak]

Current implementation:

wy2u01(uint64_t r){ const double _wynorm=1.0/(1ull<<52); return (r>>12)*_wynorm;}

I believe this should be <<53 and >>11 in order to get the highest number of possible results (while still being evenly spaced).

See:

• Daniel Lemire's blog: How many floating-point numbers are in the interval [0,1]?
• Taylor R Campbell: Uniform random floats: How to generate a double-precision floating-point number in [0, 1] uniformly at random given a uniform random source of bits.
• hash4j's nextDouble implementation (Java): return (nextLong() >>> 11) * 0x1.0p-53;