mlemanczyk/even-perfect-numbers-scanner
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add divisor-mode detailed check coverage for factors

Browse Source
This commit is contained in:
Marek Lemańczyk committed 2025-09-17 19:02:21 +02:00
1 parent b1c29ad16e
commit 1b28584168
7 files changed
+465 -224
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

PerfectNumbers.Core/Gpu/MersenneNumberResidueGpuTester.cs
+68 -42
View File
@@ -11,16 +11,28 @@ public class MersenneNumberResidueGpuTester(bool useGpuOrder)
// candidates without launching heavy kernels. Consider a device-visible constant buffer per-accelerator
// via GpuKernelPool to avoid host round-trips.
// GPU residue variant: check 2^p % q == 1 for q = 2*p*k + 1.
public void Scan(ulong exponent, UInt128 twoP, bool lastIsSeven, UInt128 maxK, ref bool isPrime)
{
var gpuLease = GpuKernelPool.GetKernel(_useGpuOrder);
// GPU residue variant: check 2^p % q == 1 for q = 2*p*k + 1.
public void Scan(
ulong exponent,
UInt128 twoP,
bool lastIsSeven,
UInt128 perSetLimit,
ulong setCount,
UInt128 overallLimit,
ref bool isPrime,
ref bool divisorsExhausted)
{
if (setCount == 0UL || perSetLimit == UInt128.Zero || overallLimit == UInt128.Zero)
{
return;
}
var gpuLease = GpuKernelPool.GetKernel(_useGpuOrder);
var accelerator = gpuLease.Accelerator;
// Ensure device has small cycles and primes tables for in-kernel lookup
var smallCyclesView = GpuKernelPool.EnsureSmallCyclesOnDevice(accelerator);
ResiduePrimeViews primeViews = GpuKernelPool.EnsureSmallPrimesOnDevice(accelerator);
int batchSize = GpuConstants.ScanBatchSize;
UInt128 kStart = 1UL;
byte last = lastIsSeven ? (byte)1 : (byte)0;
var kernel = gpuLease.Pow2ModKernel;
ulong step10 = ((exponent % 10UL) << 1) % 10UL;
@@ -31,57 +43,71 @@ public class MersenneNumberResidueGpuTester(bool useGpuOrder)
var orderBuffer = accelerator.Allocate1D<ulong>(batchSize);
ulong[] orderArray = ArrayPool<ulong>.Shared.Rent(batchSize);
UInt128 remaining;
int currentSize;
int i;
UInt128 batchSize128 = (UInt128)batchSize, q;
UInt128 batchSize128 = (UInt128)batchSize;
UInt128 limitInclusive = overallLimit + UInt128.One;
Span<ulong> orders = orderArray.AsSpan(0, batchSize);
try
{
while (kStart < maxK && Volatile.Read(ref isPrime))
for (ulong setIndex = 0; setIndex < setCount && Volatile.Read(ref isPrime); setIndex++)
{
remaining = maxK - kStart;
if (remaining > batchSize128)
{
currentSize = batchSize;
}
else
{
currentSize = (int)remaining;
orders = orderArray.AsSpan(0, currentSize);
}
q = twoP * kStart + UInt128.One;
q.Mod10_8_5_3(out ulong q0m10, out ulong q0m8, out ulong q0m5, out ulong q0m3);
var ra = new ResidueAutomatonArgs(q0m10, step10, q0m8, step8, q0m3, step3, q0m5, step5);
kernel(currentSize, exponent, twoPGpu, (GpuUInt128)kStart, last, 0UL,
ra, orderBuffer.View, smallCyclesView, primeViews.LastOne, primeViews.LastSeven, primeViews.LastOnePow2, primeViews.LastSevenPow2);
accelerator.Synchronize();
orderBuffer.View.CopyToCPU(ref MemoryMarshal.GetReference(orders), currentSize);
if (!Volatile.Read(ref isPrime))
UInt128 setOffset = perSetLimit * (UInt128)setIndex;
UInt128 setStart = setOffset + UInt128.One;
if (setStart >= limitInclusive)
{
break;
}
for (i = 0; i < currentSize; i++)
UInt128 setLimitExclusive = setOffset + perSetLimit + UInt128.One;
if (setLimitExclusive > limitInclusive)
{
if (orders[i] != 0UL)
{
Volatile.Write(ref isPrime, false);
break;
}
setLimitExclusive = limitInclusive;
}
kStart += batchSize128;
UInt128 kStart = setStart;
while (kStart < setLimitExclusive && Volatile.Read(ref isPrime))
{
UInt128 remaining = setLimitExclusive - kStart;
int currentSize = remaining > batchSize128 ? batchSize : (int)remaining;
if (currentSize <= 0)
{
break;
}
orders = orderArray.AsSpan(0, currentSize);
UInt128 q = twoP * kStart + UInt128.One;
q.Mod10_8_5_3(out ulong q0m10, out ulong q0m8, out ulong q0m5, out ulong q0m3);
var ra = new ResidueAutomatonArgs(q0m10, step10, q0m8, step8, q0m3, step3, q0m5, step5);
kernel(currentSize, exponent, twoPGpu, (GpuUInt128)kStart, last, 0UL,
ra, orderBuffer.View, smallCyclesView, primeViews.LastOne, primeViews.LastSeven, primeViews.LastOnePow2, primeViews.LastSevenPow2);
accelerator.Synchronize();
orderBuffer.View.CopyToCPU(ref MemoryMarshal.GetReference(orders), currentSize);
if (!Volatile.Read(ref isPrime))
{
break;
}
for (int i = 0; i < currentSize; i++)
{
if (orders[i] != 0UL)
{
Volatile.Write(ref isPrime, false);
break;
}
}
kStart += batchSize128;
}
}
}
finally
{
ArrayPool<ulong>.Shared.Return(orderArray);
orderBuffer.Dispose();
gpuLease.Dispose();
}
}
orderBuffer.Dispose();
gpuLease.Dispose();
}
divisorsExhausted = true;
}
}