QuarkBurst is the ultra-efficient chaotic PRNG (non-cryptographic) that has a period of at least 2⁶⁴ (from a Weyl sequence), excellent statistical randomness quality test results, low-footprint implementation, reversibility (state rewinding) and ultra-fast speed.
When independent parallel streams are required, a counter-based PRNG (LeptonFlurry) is the alternative.
QuarkBurst was created by William Stafford Parsons as a product of Eightomic.
QuarkBurst is licensed with BSD-3-Clause.
Each mention of QuarkBurst refers to each of the 3 following variants individually (quarkburst1x64, quarkburst2x64, quarkburst4x64 and quarkburst8x64) implemented in C (requiring the stdint.h header to define a 64-bit, unsigned integral type for uint64_t).
The quarkburst1x64 function modifies the state in a struct quarkburst1x64_state instance to generate a deterministic pseudorandom uint64_t integer as the return value.
Each state variable (a, b and c) in a struct quarkburst1x64_state instance must be seeded before generating a deterministic quarkburst1x64 sequence (that must discard the first 3 quarkburst1x64 results as a state warmup).
The quarkburst2x64 function modifies the state in a struct quarkburst2x64_state instance to generate 2 deterministic pseudorandom uint64_t integers in the output array.
Each single-letter state variable (a, b and c) in a struct quarkburst2x64_state instance must be seeded before generating a deterministic quarkburst2x64 sequence (that must discard the first 3 quarkburst2x64 results as a state warmup).
The quarkburst4x64 function modifies the state in a struct quarkburst4x64_state instance to generate 4 deterministic pseudorandom uint64_t integers in the output array.
Each single-letter state variable (a, b, c and d) in a struct quarkburst4x64_state instance must be seeded before generating a deterministic quarkburst4x64 sequence (that must discard the first 3 quarkburst4x64 results as a state warmup).
quarkburst8x64 doesn't exist.
QuarkBurst yielded the following excellent results (from a stream that iterated through a QuarkBurst instance) from statistical randomness quality tests.
QuarkBurst passed dieharder -Y 1 -a -g 200 -k 2 (each ambiguous result resolved to a PASSED result).
QuarkBurst passed assess 1000000 (using 100 bitstreams).
QuarkBurst passed RNG_test stdin -tlmin 1KB -tlmax 8TB.
Further testing to 32TB is in progress.
QuarkBurst passed smokerand full stdin64 (4/4 quality score).
QuarkBurst passed BigCrush.
Each of the following results log the fastest process execution speed (in milliseconds) among several repetitions of a speed benchmark (using gcc -O3 from an AMD A4-9120C) that generates 1 billion pseudorandom uint64_t integers in a blocking #pragma GCC unroll 0 loop.
| PRNG | Elapsed |
|---|---|
quarkburst8x64 |
?ms |
quarkburst4x64 |
561ms |
quarkburst2x64 |
743ms |
***shishua_avx2 (-mavx2) |
866ms |
**aesdec2 (-maes -msse4) |
905ms |
**shishua_sse4 (-msse4) |
978ms |
quarkburst1x64 |
1072ms |
**shishua_sse3 (-msse3) |
1147ms |
**shishua_sse2 (-msse2) |
1154ms |
biski64 |
1292ms |
sfc64 |
1320ms |
*xoshiro128_plus |
1541ms |
xoshiro256_plus |
1546ms |
xorshiftr128_plus |
1654ms |
jsf64_2rotate |
1718ms |
xoroshiro128_plus |
1733ms |
*xoshiro128_plusplus |
1740ms |
*xoroshiro64_star |
1749ms |
*xoshiro128_starstar |
1759ms |
mrsf64 |
1833ms |
jsf64_3rotate |
1841ms |
mrc64 |
1862ms |
romu_trio |
1894ms |
wob2m |
1928ms |
xoshiro64_starstar |
1945ms |
mwc192 |
1997ms |
wyrand |
2033ms |
xorshift64 |
2135ms |
shishua |
2251ms |
xorshift128_plus |
2260ms |
*xorwow |
2882ms |
romu_mono |
2982ms |
*pcg32_minimal |
2983ms |
*pcg_oneseq_64_xsh_rr_32 |
2987ms |
mwc128 |
2998ms |
*lehmer_mcg32 |
3402ms |
*pcg_oneseq_64_xsh_rs_32 |
3404ms |
*lcg32 |
3409ms |
lehmer_mcg64 |
3413ms |
lcg64 |
3416ms |
*threefry4x32 |
3532ms |
*philox4x32 |
3626ms |
**aes_ni_ctr_128 (-maes -msse4) |
3796ms |
pcg_oneseq_64_xsl_rr_rr_64 |
3928ms |
*isaac |
4099ms |
splitmix64 |
4385ms |
cwg64 |
4680ms |
cwg128 |
4757ms |
*threefry2x32 |
4959ms |
*philox2x32 |
5139ms |
**sfmt (-msse2) |
5525ms |
lxm_xbg128 |
5863ms |
wanghash64 |
5983ms |
pcg_oneseq_128_xsh_rr_64 |
6833ms |
threefry4x64 |
7093ms |
mt19937_64 |
7126ms |
*squares32 |
7552ms |
pcg64_dxsm |
7604ms |
pcg_oneseq_128_xsh_rs_64 |
7676ms |
philox4x64 |
9171ms |
*****google_randen (-maes -msse4) |
9206ms |
squares64 |
9596ms |
threefry2x64 |
9907ms |
philox2x64 |
10622ms |
*chacha8 |
13230ms |
tinymt64 |
16081ms |
****mrg32k3a |
19356ms |
*chacha20 |
26402ms |
*rand (stdlib.h) |
46083ms |
* Each n-bit output integer was casted to a uint64_t integer.
** Each 128-bit output integer was extracted as 2 uint64_t integers.
*** Each 256-bit output integer was extracted as 4 uint64_t integers.
**** Each output integer was returned as a uint64_t integer (omitting the double integer conversion).
***** Each block of 4 uint8_t output integers were merged into a uint32_t integer that was casted to a uint64_t integer.