Use heuristic sieves for prime trial division

PerfectNumbers.Core/HeuristicPrimeSieves.cs

@@ -15,195 +15,111 @@ public static class HeuristicPrimeSieves
     private static readonly ushort[] GroupBResiduesEnding7 = [17, 37, 47, 67, 97, 107, 127, 137, 157, 167, 187, 197];
     private static readonly ushort[] GroupBResiduesEnding9 = [19, 29, 59, 79, 89, 109, 139, 149, 169, 179, 199, 209];
 
-    private static int[]? s_groupADivisors;
-    private static ulong[]? s_groupADivisorSquares;
+    private static readonly int[] GroupADivisorsStorage;
+    private static readonly ulong[] GroupADivisorSquaresStorage;
 
-    private static int[]? s_groupBDivisorsEnding1;
-    private static ulong[]? s_groupBDivisorSquaresEnding1;
+    private static readonly int[] GroupBDivisorsEnding1Storage;
+    private static readonly ulong[] GroupBDivisorSquaresEnding1Storage;
 
-    private static int[]? s_groupBDivisorsEnding3;
-    private static ulong[]? s_groupBDivisorSquaresEnding3;
+    private static readonly int[] GroupBDivisorsEnding3Storage;
+    private static readonly ulong[] GroupBDivisorSquaresEnding3Storage;
 
-    private static int[]? s_groupBDivisorsEnding7;
-    private static ulong[]? s_groupBDivisorSquaresEnding7;
+    private static readonly int[] GroupBDivisorsEnding7Storage;
+    private static readonly ulong[] GroupBDivisorSquaresEnding7Storage;
 
-    private static int[]? s_groupBDivisorsEnding9;
-    private static ulong[]? s_groupBDivisorSquaresEnding9;
+    private static readonly int[] GroupBDivisorsEnding9Storage;
+    private static readonly ulong[] GroupBDivisorSquaresEnding9Storage;
 
-    private static readonly object InitLock = new();
-
-    public static ReadOnlySpan<int> GroupADivisors
+    static HeuristicPrimeSieves()
     {
-        get
-        {
-            EnsureGroupAInitialized();
-            return s_groupADivisors!;
-        }
+        Console.WriteLine("[HeuristicPrimeSieves] Initializing heuristic divisor sieves...");
+
+        (GroupADivisorsStorage, GroupADivisorSquaresStorage) = BuildGroupADivisors();
+        (GroupBDivisorsEnding1Storage, GroupBDivisorSquaresEnding1Storage) = BuildGroupBResidueSieve(GroupBResiduesEnding1);
+        (GroupBDivisorsEnding3Storage, GroupBDivisorSquaresEnding3Storage) = BuildGroupBResidueSieve(GroupBResiduesEnding3);
+        (GroupBDivisorsEnding7Storage, GroupBDivisorSquaresEnding7Storage) = BuildGroupBResidueSieve(GroupBResiduesEnding7);
+        (GroupBDivisorsEnding9Storage, GroupBDivisorSquaresEnding9Storage) = BuildGroupBResidueSieve(GroupBResiduesEnding9);
     }
 
-    public static ReadOnlySpan<ulong> GroupADivisorSquares
+    public static void EnsureInitialized()
     {
-        get
-        {
-            EnsureGroupAInitialized();
-            return s_groupADivisorSquares!;
-        }
+        // Intentionally left blank. Accessing this method forces the static constructor to run.
     }
 
-    public static ReadOnlySpan<int> GroupBDivisorsEnding1
+    public static ReadOnlySpan<int> GroupADivisors => GroupADivisorsStorage;
+
+    public static ReadOnlySpan<ulong> GroupADivisorSquares => GroupADivisorSquaresStorage;
+
+    public static ReadOnlySpan<int> GroupBDivisorsEnding1 => GroupBDivisorsEnding1Storage;
+
+    public static ReadOnlySpan<ulong> GroupBDivisorSquaresEnding1 => GroupBDivisorSquaresEnding1Storage;
+
+    public static ReadOnlySpan<int> GroupBDivisorsEnding3 => GroupBDivisorsEnding3Storage;
+
+    public static ReadOnlySpan<ulong> GroupBDivisorSquaresEnding3 => GroupBDivisorSquaresEnding3Storage;
+
+    public static ReadOnlySpan<int> GroupBDivisorsEnding7 => GroupBDivisorsEnding7Storage;
+
+    public static ReadOnlySpan<ulong> GroupBDivisorSquaresEnding7 => GroupBDivisorSquaresEnding7Storage;
+
+    public static ReadOnlySpan<int> GroupBDivisorsEnding9 => GroupBDivisorsEnding9Storage;
+
+    public static ReadOnlySpan<ulong> GroupBDivisorSquaresEnding9 => GroupBDivisorSquaresEnding9Storage;
+
+    private static (int[] divisors, ulong[] squares) BuildGroupADivisors()
     {
-        get
+        var divisors = new int[SieveLength];
+        var squares = new ulong[SieveLength];
+
+        int index = 0;
+        for (; index < GroupAConstants.Length && index < divisors.Length; index++)
         {
-            EnsureGroupBInitialized(ref s_groupBDivisorsEnding1, ref s_groupBDivisorSquaresEnding1, GroupBResiduesEnding1);
-            return s_groupBDivisorsEnding1!;
+            int value = GroupAConstants[index];
+            divisors[index] = value;
+            squares[index] = (ulong)value * (ulong)value;
         }
+
+        int candidate = 23;
+        int incrementIndex = 0;
+        while (index < divisors.Length)
+        {
+            divisors[index] = candidate;
+            squares[index] = (ulong)candidate * (ulong)candidate;
+
+            candidate += GroupAIncrementPattern[incrementIndex];
+            incrementIndex ^= 1;
+            index++;
+        }
+
+        return (divisors, squares);
     }
 
-    public static ReadOnlySpan<ulong> GroupBDivisorSquaresEnding1
+    private static (int[] divisors, ulong[] squares) BuildGroupBResidueSieve(ReadOnlySpan<ushort> residues)
     {
-        get
-        {
-            EnsureGroupBInitialized(ref s_groupBDivisorsEnding1, ref s_groupBDivisorSquaresEnding1, GroupBResiduesEnding1);
-            return s_groupBDivisorSquaresEnding1!;
-        }
-    }
+        var divisors = new int[SieveLength];
+        var squares = new ulong[SieveLength];
 
-    public static ReadOnlySpan<int> GroupBDivisorsEnding3
-    {
-        get
-        {
-            EnsureGroupBInitialized(ref s_groupBDivisorsEnding3, ref s_groupBDivisorSquaresEnding3, GroupBResiduesEnding3);
-            return s_groupBDivisorsEnding3!;
-        }
-    }
+        int index = 0;
+        ulong baseValue = 0UL;
 
-    public static ReadOnlySpan<ulong> GroupBDivisorSquaresEnding3
-    {
-        get
+        while (index < divisors.Length)
         {
-            EnsureGroupBInitialized(ref s_groupBDivisorsEnding3, ref s_groupBDivisorSquaresEnding3, GroupBResiduesEnding3);
-            return s_groupBDivisorSquaresEnding3!;
-        }
-    }
-
-    public static ReadOnlySpan<int> GroupBDivisorsEnding7
-    {
-        get
-        {
-            EnsureGroupBInitialized(ref s_groupBDivisorsEnding7, ref s_groupBDivisorSquaresEnding7, GroupBResiduesEnding7);
-            return s_groupBDivisorsEnding7!;
-        }
-    }
-
-    public static ReadOnlySpan<ulong> GroupBDivisorSquaresEnding7
-    {
-        get
-        {
-            EnsureGroupBInitialized(ref s_groupBDivisorsEnding7, ref s_groupBDivisorSquaresEnding7, GroupBResiduesEnding7);
-            return s_groupBDivisorSquaresEnding7!;
-        }
-    }
-
-    public static ReadOnlySpan<int> GroupBDivisorsEnding9
-    {
-        get
-        {
-            EnsureGroupBInitialized(ref s_groupBDivisorsEnding9, ref s_groupBDivisorSquaresEnding9, GroupBResiduesEnding9);
-            return s_groupBDivisorsEnding9!;
-        }
-    }
-
-    public static ReadOnlySpan<ulong> GroupBDivisorSquaresEnding9
-    {
-        get
-        {
-            EnsureGroupBInitialized(ref s_groupBDivisorsEnding9, ref s_groupBDivisorSquaresEnding9, GroupBResiduesEnding9);
-            return s_groupBDivisorSquaresEnding9!;
-        }
-    }
-
-    private static void EnsureGroupAInitialized()
-    {
-        if (s_groupADivisors is not null)
-        {
-            return;
-        }
-
-        lock (InitLock)
-        {
-            if (s_groupADivisors is not null)
+            for (int i = 0; i < residues.Length && index < divisors.Length; i++)
             {
-                return;
-            }
+                ulong candidate = baseValue + residues[i];
+                if (candidate <= 13UL)
+                {
+                    continue;
+                }
 
-            var divisors = new int[SieveLength];
-            var squares = new ulong[SieveLength];
-
-            int index = 0;
-            for (; index < GroupAConstants.Length && index < divisors.Length; index++)
-            {
-                int value = GroupAConstants[index];
-                divisors[index] = value;
-                squares[index] = (ulong)value * (ulong)value;
-            }
-
-            int candidate = 23;
-            int incrementIndex = 0;
-            while (index < divisors.Length)
-            {
-                divisors[index] = candidate;
-                squares[index] = (ulong)candidate * (ulong)candidate;
-
-                candidate += GroupAIncrementPattern[incrementIndex];
-                incrementIndex ^= 1;
+                divisors[index] = (int)candidate;
+                squares[index] = candidate * candidate;
                 index++;
             }
 
-            s_groupADivisors = divisors;
-            s_groupADivisorSquares = squares;
-        }
-    }
-
-    private static void EnsureGroupBInitialized(ref int[]? divisorsField, ref ulong[]? squaresField, ReadOnlySpan<ushort> residues)
-    {
-        if (divisorsField is not null)
-        {
-            return;
+            baseValue += (ulong)Wheel210;
         }
 
-        lock (InitLock)
-        {
-            if (divisorsField is not null)
-            {
-                return;
-            }
-
-            var divisors = new int[SieveLength];
-            var squares = new ulong[SieveLength];
-
-            int index = 0;
-            ulong baseValue = 0UL;
-
-            while (index < divisors.Length)
-            {
-                for (int i = 0; i < residues.Length && index < divisors.Length; i++)
-                {
-                    ulong candidate = baseValue + residues[i];
-                    if (candidate <= 13UL)
-                    {
-                        continue;
-                    }
-
-                    divisors[index] = (int)candidate;
-                    squares[index] = candidate * candidate;
-                    index++;
-                }
-
-                baseValue += (ulong)Wheel210;
-            }
-
-            divisorsField = divisors;
-            squaresField = squares;
-        }
+        return (divisors, squares);
     }
 }

PerfectNumbers.Core/PrimeTester.cs

@@ -14,6 +14,11 @@ namespace PerfectNumbers.Core;
 
 public sealed class PrimeTester(bool useInternal = false)
 {
+    static PrimeTester()
+    {
+        HeuristicPrimeSieves.EnsureInitialized();
+    }
+
     // TODO: Don't use these tiny arrays. Hard-code the values / checks instead
     private static readonly ulong[] GroupAConstantDivisors = [3UL, 7UL, 11UL, 13UL];
     private static readonly byte[] GroupAIncrementPattern = [20, 10];
@@ -197,20 +202,13 @@ public sealed class PrimeTester(bool useInternal = false)
 
     private static bool HeuristicTrialDivisionCpu(ulong n, ulong sqrtLimit, byte nMod10, bool includeGroupB)
     {
-        Span<HeuristicGroupBSequenceState> groupBBuffer = stackalloc HeuristicGroupBSequenceState[MaxGroupBSequences];
-        var enumerator = new HeuristicDivisorEnumerator(sqrtLimit, nMod10, groupBBuffer);
-
-        while (enumerator.TryGetNext(out HeuristicDivisorCandidate candidate))
+        ReadOnlySpan<int> groupADivisors = HeuristicPrimeSieves.GroupADivisors;
+        for (int i = 0; i < groupADivisors.Length; i++)
         {
-            if (!includeGroupB && candidate.Group == HeuristicDivisorGroup.GroupB)
+            ulong divisor = (ulong)groupADivisors[i];
+            if (divisor > sqrtLimit)
             {
-                return EvaluateWithOpenNumericFallback(n);
-            }
-
-            ulong divisor = candidate.Value;
-            if (divisor <= 1UL)
-            {
-                continue;
+                break;
             }
 
             if (n % divisor == 0UL)
@@ -219,20 +217,71 @@ public sealed class PrimeTester(bool useInternal = false)
             }
         }
 
-        return includeGroupB ? true : EvaluateWithOpenNumericFallback(n);
+        if (!includeGroupB)
+        {
+            return EvaluateWithOpenNumericFallback(n);
+        }
+
+        ReadOnlySpan<byte> endingOrder = GetGroupBEndingOrder(nMod10);
+        if (endingOrder.IsEmpty)
+        {
+            return EvaluateWithOpenNumericFallback(n);
+        }
+
+        Span<int> indices = endingOrder.Length <= 8
+            ? stackalloc int[endingOrder.Length]
+            : new int[endingOrder.Length];
+
+        while (true)
+        {
+            ulong bestCandidate = ulong.MaxValue;
+            int bestEndingIndex = -1;
+
+            for (int i = 0; i < endingOrder.Length; i++)
+            {
+                ReadOnlySpan<int> divisors = GetGroupBDivisors(endingOrder[i]);
+                int index = indices[i];
+                if ((uint)index >= (uint)divisors.Length)
+                {
+                    continue;
+                }
+
+                ulong candidate = (ulong)divisors[index];
+                if (candidate > sqrtLimit)
+                {
+                    indices[i] = divisors.Length;
+                    continue;
+                }
+
+                if (bestEndingIndex == -1 || candidate < bestCandidate)
+                {
+                    bestCandidate = candidate;
+                    bestEndingIndex = i;
+                }
+            }
+
+            if (bestEndingIndex == -1)
+            {
+                break;
+            }
+
+            if (n % bestCandidate == 0UL)
+            {
+                return false;
+            }
+
+            indices[bestEndingIndex]++;
+        }
+
+        return true;
     }
 
     private bool HeuristicTrialDivisionGpuDetectsDivisor(ulong n, ulong sqrtLimit, byte nMod10)
     {
-        Span<HeuristicGroupBSequenceState> groupBBuffer = stackalloc HeuristicGroupBSequenceState[MaxGroupBSequences];
-        var enumerator = new HeuristicDivisorEnumerator(sqrtLimit, nMod10, groupBBuffer);
-
         int batchCapacity = Math.Max(1, HeuristicGpuDivisorBatchSize);
-        var candidatePool = ArrayPool<HeuristicDivisorCandidate>.Shared;
         var divisorPool = ThreadStaticPools.UlongPool;
         var hitPool = ThreadStaticPools.BytePool;
 
-        HeuristicDivisorCandidate[]? candidateArray = null;
         ulong[]? divisorArray = null;
         byte[]? hitFlags = null;
 
@@ -243,7 +292,6 @@ public sealed class PrimeTester(bool useInternal = false)
 
         try
         {
-            candidateArray = candidatePool.Rent(batchCapacity);
             divisorArray = divisorPool.Rent(batchCapacity);
             hitFlags = hitPool.Rent(batchCapacity);
 
@@ -259,6 +307,11 @@ public sealed class PrimeTester(bool useInternal = false)
 
                     bool ProcessBatch(int length)
                     {
+                        if (length <= 0)
+                        {
+                            return false;
+                        }
+
                         scratch.Input.View.CopyFromCPU(ref divisorArray![0], length);
                         state.HeuristicTrialDivisionKernel(length, scratch.Input.View, n, scratch.Output.View);
                         accelerator.Synchronize();
@@ -271,7 +324,7 @@ public sealed class PrimeTester(bool useInternal = false)
                                 continue;
                             }
 
-                            ulong divisor = candidateArray![i].Value;
+                            ulong divisor = divisorArray![i];
                             if (divisor > 1UL && n % divisor == 0UL)
                             {
                                 return true;
@@ -281,36 +334,87 @@ public sealed class PrimeTester(bool useInternal = false)
                         return false;
                     }
 
-                    while (enumerator.TryGetNext(out HeuristicDivisorCandidate candidate))
+                    bool QueueDivisor(ulong divisor)
                     {
-                        ulong divisor = candidate.Value;
-                        if (divisor <= 1UL)
-                        {
-                            continue;
-                        }
-
-                        candidateArray[count] = candidate;
-                        divisorArray[count] = divisor;
+                        divisorArray![count] = divisor;
                         count++;
 
                         if (count == batchCapacity)
                         {
                             if (ProcessBatch(count))
                             {
-                                compositeDetected = true;
-                                break;
+                                return true;
                             }
 
                             count = 0;
                         }
+
+                        return false;
+                    }
+
+                    ReadOnlySpan<int> groupADivisors = HeuristicPrimeSieves.GroupADivisors;
+                    for (int i = 0; i < groupADivisors.Length && !compositeDetected; i++)
+                    {
+                        ulong divisor = (ulong)groupADivisors[i];
+                        if (divisor > sqrtLimit)
+                        {
+                            break;
+                        }
+
+                        compositeDetected = QueueDivisor(divisor);
+                    }
+
+                    if (!compositeDetected)
+                    {
+                        ReadOnlySpan<byte> endingOrder = GetGroupBEndingOrder(nMod10);
+                        if (!endingOrder.IsEmpty)
+                        {
+                            Span<int> indices = endingOrder.Length <= 8
+                                ? stackalloc int[endingOrder.Length]
+                                : new int[endingOrder.Length];
+
+                            while (!compositeDetected)
+                            {
+                                ulong bestCandidate = ulong.MaxValue;
+                                int bestEndingIndex = -1;
+
+                                for (int i = 0; i < endingOrder.Length; i++)
+                                {
+                                    ReadOnlySpan<int> divisors = GetGroupBDivisors(endingOrder[i]);
+                                    int index = indices[i];
+                                    if ((uint)index >= (uint)divisors.Length)
+                                    {
+                                        continue;
+                                    }
+
+                                    ulong candidate = (ulong)divisors[index];
+                                    if (candidate > sqrtLimit)
+                                    {
+                                        indices[i] = divisors.Length;
+                                        continue;
+                                    }
+
+                                    if (bestEndingIndex == -1 || candidate < bestCandidate)
+                                    {
+                                        bestCandidate = candidate;
+                                        bestEndingIndex = i;
+                                    }
+                                }
+
+                                if (bestEndingIndex == -1)
+                                {
+                                    break;
+                                }
+
+                                compositeDetected = QueueDivisor(bestCandidate);
+                                indices[bestEndingIndex]++;
+                            }
+                        }
                     }
 
                     if (!compositeDetected && count > 0)
                     {
-                        if (ProcessBatch(count))
-                        {
-                            compositeDetected = true;
-                        }
+                        compositeDetected = ProcessBatch(count);
                     }
                 }
                 finally
@@ -331,11 +435,6 @@ public sealed class PrimeTester(bool useInternal = false)
                 divisorPool.Return(divisorArray);
             }
 
-            if (candidateArray is not null)
-            {
-                candidatePool.Return(candidateArray);
-            }
-
             gpu.Dispose();
             limiter.Dispose();
         }
@@ -380,6 +479,15 @@ public sealed class PrimeTester(bool useInternal = false)
         _ => ReadOnlySpan<byte>.Empty,
     };
 
+    private static ReadOnlySpan<int> GetGroupBDivisors(byte ending) => ending switch
+    {
+        1 => HeuristicPrimeSieves.GroupBDivisorsEnding1,
+        3 => HeuristicPrimeSieves.GroupBDivisorsEnding3,
+        7 => HeuristicPrimeSieves.GroupBDivisorsEnding7,
+        9 => HeuristicPrimeSieves.GroupBDivisorsEnding9,
+        _ => ReadOnlySpan<int>.Empty,
+    };
+
     private static ReadOnlySpan<ushort> GetWheel210ResiduesForEnding(byte ending) => ending switch
     {
         1 => Wheel210ResiduesEnding1,

docs/HeuristicIsPrimePlan.md

@@ -116,7 +116,7 @@ The plan incorporates the updated divisor-class heuristics (Groups A/B, wheels,
 1. [done] **Snapshot current implementations** – preserve `IsPrimeGpu` and `IsPrimeInternal` as references for regression comparisons and migration checkpoints (see Section 2).
 2. [done] **Introduce placeholder heuristic entry points** – add `HeuristicIsPrimeCpu` and `HeuristicIsPrimeGpu` by copying the current internal and GPU methods so subsequent commits can iterate without disrupting legacy behavior.
 3. [done] **Design shared heuristic scaffolding**
-   * [done] Introduced `HeuristicDivisorEnumerator` and stackalloc-backed Group B sequence states so CPU/GPU heuristics share the Group A/B wheel metadata and candidate ordering.
+   * [done] Added `HeuristicPrimeSieves` to precompute 4M-entry Group A/B divisor tables so heuristic prime checks reuse cached sequences without regenerating wheel steps at runtime; the enumerator remains available for residue-driven consumers.
    * [done] Surface helpers that accept the precomputed `R = ⌊√n⌋` and `nMod10` values from callers via the new overload of `HeuristicIsPrimeCpu`; PrimeTester still computes them locally until call sites forward the parameters.
    * [done] Simplified the heuristic execution path so it runs without cancellation tokens or callback plumbing while keeping the shared enumerator available to other components.
 4. **Share residue and wheel state with by-divisor sessions**
@@ -127,11 +127,11 @@ The plan incorporates the updated divisor-class heuristics (Groups A/B, wheels,
    * [done] Extend the heuristic scaffolding so it returns `MontgomeryDivisorData` descriptors and cycle-length hints for each candidate divisor via `HeuristicDivisorPreparation` and `PrepareHeuristicDivisor`.
    * Wire these outputs into the existing cycle calculators to avoid duplicate cache lookups inside the CPU/GPU divisor scans.
 6. [done] **Implement `HeuristicIsPrimeCpu`**
-   * Baseline Group A/B enumeration now executes inside `HeuristicIsPrimeCpu` via `HeuristicDivisorEnumerator`, computing `⌊√n⌋` locally and short-circuiting on the first divisor.
+   * Baseline Group A/B enumeration now executes inside `HeuristicIsPrimeCpu` via the cached `HeuristicPrimeSieves` tables, computing `⌊√n⌋` locally and short-circuiting on the first divisor.
    * [done] The CPU heuristic now operates as a pure trial-division sweep without maintaining per-call summaries, relying on `PrepareHeuristicDivisor` only when downstream consumers request Montgomery data.
    * [done] Introduced the temporary `UseHeuristicGroupBTrialDivision` gate so current builds execute Group A locally and then fall back to `Open.Numeric.Primes.Prime.Numbers.IsPrime`, keeping the full Group B implementation available for reactivation.
 7. [done] **Implement `HeuristicIsPrimeGpu`**
-   * [done] Mirror CPU structure while batching divisor checks on the accelerator via `HeuristicTrialDivisionGpu`, reusing caller-provided `sqrtLimit`/`nMod10` inputs and falling back to CPU when GPUs are disabled.
+   * [done] Mirror CPU structure while batching divisor checks on the accelerator via `HeuristicTrialDivisionGpu`, drawing divisors from `HeuristicPrimeSieves`, reusing caller-provided `sqrtLimit`/`nMod10` inputs, and falling back to CPU when GPUs are disabled.
    * [done] Added `PrimeTesterKernels.HeuristicTrialDivisionKernel` so GPU batches respect Group A/B ordering while emitting hit flags for early exits.
    * [done] Kept GPU divisibility in standard modular arithmetic while reserving Montgomery transforms for CPU confirmation logic.
    * [done] When the temporary Group B gate is disabled, delegate GPU calls to the CPU fallback so Group A coverage remains in place while the accelerator path stays available for future re-enablement.

docs/HeuristicIsPrimeSpecification.md

@@ -71,9 +71,9 @@ The by-divisor executors (`MersenneNumberDivisorByDivisorCpuTester` and `Mersenn
 
 * **Residue filters** – The CPU scan maintains rolling residues modulo 10/8/5/3/7/11 to match the filters in `MersenneCpuDivisorScanSession.CheckDivisors`. `HeuristicIsPrime*` must either share the residue update helpers or produce identical stepping deltas so the divisor scanners can reuse the pre-filtered candidate stream without recomputing moduli. The shared `MersenneDivisorResidueStepper` now centralises those modulo calculations so both paths reuse the same state transitions, and the GPU by-divisor tester instantiates the same stepper before filtering candidates.
 * **Stepper integration** – CPU sessions advance residues and unity checks through `ExponentRemainderStepperCpu` together with `CycleRemainderStepper`; GPU kernels mirror this contract via `ExponentRemainderStepperGpu`. The heuristic implementation must call into these steppers (or emit the same remainder deltas) so the shared stepping state remains aligned across CPU/GPU scans and factor validation.
-* **Wheel alignment** – The wheel 30/210 sequences emitted by `PrimeTester.CreateHeuristicDivisorEnumerator` now drive both the CPU and GPU by-divisor testers, keeping the divisor cache and kernel inputs in lockstep without bespoke residue loops.
-* **Divisor streaming API** – `PrimeTester.CreateHeuristicDivisorEnumerator` now feeds `MersenneNumberDivisorByDivisorCpuTester.CheckDivisors` and `MersenneNumberDivisorByDivisorGpuTester.CheckDivisors`, replacing bespoke residue loops while supplying candidates directly to Montgomery preparation and GPU kernels.
-* **GPU batching** – `HeuristicIsPrimeGpu` streams Group A/B candidates through the same enumerator, packs them into batches controlled by `PrimeTester.HeuristicGpuDivisorBatchSize`, and launches `PrimeTesterKernels.HeuristicTrialDivisionKernel` to test each divisor with standard `%` arithmetic. When GPUs are disabled, the method delegates to the CPU heuristic so fallbacks reuse the same candidate ordering without additional plumbing.
+* **Wheel alignment** – `HeuristicPrimeSieves` precomputes the wheel 30/210 sequences once at startup so both the heuristic prime tester and the by-divisor executors consume identical Group A/B streams without regenerating residues on every call.
+* **Divisor streaming API** – `PrimeTester.CreateHeuristicDivisorEnumerator` exposes the same precomputed sieves so by-divisor scans can pull candidates directly into Montgomery preparation and GPU kernels without bespoke residue loops.
+* **GPU batching** – `HeuristicIsPrimeGpu` streams Group A/B candidates from `HeuristicPrimeSieves`, packs them into batches controlled by `PrimeTester.HeuristicGpuDivisorBatchSize`, and launches `PrimeTesterKernels.HeuristicTrialDivisionKernel` to test each divisor with standard `%` arithmetic. When GPUs are disabled, the method delegates to the CPU heuristic so fallbacks reuse the same candidate ordering without additional plumbing.
 * **Cycle preparation** – When the heuristic discovers a divisor candidate, it should populate `MontgomeryDivisorData` and expose the computed cycle length (or the need to recompute it) so that `DivisorCycleCache` and `MersenneDivisorCycles` do not duplicate work for the same `(prime, divisor)` pair during factor verification. `PrimeTester.PrepareHeuristicDivisor` together with `PrimeTester.ResolveHeuristicCycleLength` now provides this payload to both CPU and GPU by-divisor scans, and the trial-division summaries capture the data whenever a divisor is found.
 
 * **Caching contract** – The by-divisor GPU implementation pools `DivisorScanSession` instances and reuses GPU buffers. The heuristic must accept externally allocated spans/buffers for divisor candidates, cycle lengths, and hits so the existing resource pools remain valid.
@@ -97,8 +97,8 @@ The following summaries cover the first 500,000 odd composite numbers after remo
 
 ### Heuristic cheat sheet
 
-* Group A: test `3, 7, 11, 13` and every `d ≡ 3 (mod 10)` up to `⌊√n⌋` (wheel 30).
-* Group B: test only the allowed endings determined by `n mod 10`, scanning in ascending order (wheel 210) and applying the per-class priority order: `n ≡ 1` → `9` then `1`; `n ≡ 3` → `9` then `7`; `n ≡ 7` → `7` then `1`; `n ≡ 9` → `9`, `7`, `1`.
+* Group A: test `3, 7, 11, 13` and every `d ≡ 3 (mod 10)` up to `⌊√n⌋` from the precomputed wheel-30 sieve.
+* Group B: test only the allowed endings determined by `n mod 10`, scanning in ascending order from the wheel-210 sieves and applying the per-class priority order: `n ≡ 1` → `9` then `1`; `n ≡ 3` → `9` then `7`; `n ≡ 7` → `7` then `1`; `n ≡ 9` → `9`, `7`, `1`.
 * Because `min(B)` is overwhelmingly smaller than `A_max(n)` for the “no Group A divisor” cohort, always start Group B from its smallest candidates instead of resuming above `A_max(n)`.
 
 ## Benchmark Reporting