taus88 RNG

Author: WrathfulSpatula

FindAFactor/_find_a_factor.cpp

@@ -52,6 +52,7 @@
 
 #include <boost/dynamic_bitset.hpp>
 #include <boost/multiprecision/cpp_int.hpp>
+#include <boost/random.hpp>
 
 #include <pybind11/pybind11.h>
 #include <pybind11/stl.h>
@@ -764,8 +765,8 @@ struct Factorizer {
   ForwardFn forwardFn;
 
   Factorizer(const BigInteger &tfsqr, const BigInteger &tf, const BigInteger &tfsqrt, const BigInteger &range, size_t nodeCount, size_t nodeId, size_t w, size_t spl, size_t bsv,
-             size_t lm, const std::vector<size_t> &p, ForwardFn fn)
-    : dis(0U, p.size() - 1U), wordDis(0ULL, -1ULL), toFactorSqr(tfsqr), toFactor(tf), toFactorSqrt(tfsqrt), batchRange(range), batchNumber(0U), batchOffset(nodeId * range),
+             size_t lm, size_t bn, const std::vector<size_t> &p, ForwardFn fn)
+    : dis(0U, p.size() - 1U), wordDis(0ULL, -1ULL), toFactorSqr(tfsqr), toFactor(tf), toFactorSqrt(tfsqrt), batchRange(range), batchNumber(bn), batchOffset(nodeId * range),
     batchTotal(nodeCount * range), wheelRadius(1U), wheelEntryCount(w), smoothBatchLimit(spl), batchSizeVariance(bsv), ladderMultiple(lm), isIncomplete(true), primes(p), forwardFn(fn)
   {
     for (size_t i = 0U; i < primes.size(); ++i) {
@@ -805,7 +806,7 @@ struct Factorizer {
     return 1U;
   }
 
-  BigInteger monteCarlo(std::mt19937_64& gen) {
+  BigInteger monteCarlo(boost::random::taus88& gen) {
     // This function enters only once per thread.
     size_t batchPower = 0U;
 
@@ -1108,18 +1109,20 @@ std::string find_a_factor(std::string toFactorStr, size_t method, size_t nodeCou
   gearFactorizationPrimes.clear();
 
   // Range per parallel node
-  const BigInteger nodeRange = (((backward(SMALLEST_WHEEL)(fullMaxBase) + nodeCount - 1U) / nodeCount) + wheelEntryCount - 1U) / wheelEntryCount;
+  const auto backwardFn = backward(SMALLEST_WHEEL);
+  const BigInteger nodeRange = (((backwardFn(fullMaxBase) + nodeCount - 1U) / nodeCount) + wheelEntryCount - 1U) / wheelEntryCount;
+  const size_t batchStart = ((size_t)backwardFn(trialDivisionLevel)) / wheelEntryCount;
   // This manages the work of all threads.
   Factorizer worker(toFactor * toFactor, toFactor, fullMaxBase,
                     nodeRange, nodeCount, nodeId,
                     wheelEntryCount, (size_t)(batchSizeMultiplier * smoothPrimes.size() * log(smoothPrimes.size())), batchSizeVariance,
-                    ladderMultiple, smoothPrimes, forward(SMALLEST_WHEEL));
+                    ladderMultiple, batchStart, smoothPrimes, forward(SMALLEST_WHEEL));
   // Square of count of smooth primes, for FACTOR_FINDER batch multiplier base unit, was suggested by Lyra (OpenAI GPT)
 
   std::vector<std::future<BigInteger>> futures;
   futures.reserve(CpuCount);
 
-  std::vector<std::mt19937_64> gen;
+  std::vector<boost::random::taus88> gen;
   if (isFactorFinder) {
     std::default_random_engine rng{};
     gen.reserve(CpuCount);

pyproject.toml

@@ -8,7 +8,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "FindAFactor"
-version = "4.7.1"
+version = "4.7.2"
 requires-python = ">=3.8"
 description = "Find any nontrivial factor of a number"
 readme = {file = "README.txt", content-type = "text/markdown"}

setup.py

@@ -40,7 +40,7 @@ def build_extension(self, ext):
 
 setup(
     name='FindAFactor',
-    version='4.7.1',
+    version='4.7.2',
     author='Dan Strano',
     author_email='stranoj@gmail.com',
     description='Find any nontrivial factor of a number',