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

Refine GPU residue utilities and add positive divisor tests

Browse Source
This commit is contained in:
Marek Lemańczyk committed 2025-09-13 16:47:38 +02:00
1 parent 515a1bb342
commit 58eb65ca9e
7 files changed
+224 -121
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

PerfectNumbers.Core/Gpu/MersenneNumberResidueGpuTester.cs
+50 -48
View File
@@ -15,26 +15,32 @@ public class MersenneNumberResidueGpuTester(bool useGpuOrder)
public void Scan(ulong exponent, UInt128 twoP, bool lastIsSeven, UInt128 maxK, ref bool isPrime)
{
var gpuLease = GpuKernelPool.GetKernel(_useGpuOrder);
var accelerator = gpuLease.Accelerator;
// Ensure device has small cycles table for in-kernel lookup
var smallCyclesView = GpuKernelPool.EnsureSmallCyclesOnDevice(accelerator);
int batchSize = GpuConstants.ScanBatchSize;
UInt128 kStart = 1UL;
byte last = lastIsSeven ? (byte)1 : (byte)0;
var kernel = gpuLease.Pow2ModKernel;
var accelerator = gpuLease.Accelerator;
// Ensure device has small cycles and primes tables for in-kernel lookup
var smallCyclesView = GpuKernelPool.EnsureSmallCyclesOnDevice(accelerator);
var (smallPrimesView, smallPrimesPow2View) = 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;
ulong step8 = ((exponent & 7UL) << 1) & 7UL;
ulong step3 = ((exponent % 3UL) << 1) % 3UL;
ulong step5 = ((exponent % 5UL) << 1) % 5UL;
GpuUInt128 twoPGpu = (GpuUInt128)twoP;
var orderBuffer = accelerator.Allocate1D<ulong>(batchSize);
ulong[] orderArray = ArrayPool<ulong>.Shared.Rent(batchSize);
UInt128 remaining;
int currentSize;
int i;
UInt128 batchSize128 = (UInt128)batchSize, q;
Span<ulong> orders = orderArray.AsSpan(0, batchSize);
try
{
while (kStart < maxK && Volatile.Read(ref isPrime))
{
remaining = maxK - kStart;
var orderBuffer = accelerator.Allocate1D<ulong>(batchSize);
ulong[] orderArray = ArrayPool<ulong>.Shared.Rent(batchSize);
UInt128 remaining;
int currentSize;
int i;
UInt128 batchSize128 = (UInt128)batchSize, q;
Span<ulong> orders = orderArray.AsSpan(0, batchSize);
try
{
while (kStart < maxK && Volatile.Read(ref isPrime))
{
remaining = maxK - kStart;
if (remaining > batchSize128)
{
currentSize = batchSize;
@@ -45,39 +51,35 @@ public class MersenneNumberResidueGpuTester(bool useGpuOrder)
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);
ulong step10 = ((exponent % 10UL) << 1) % 10UL;
ulong step8 = ((exponent & 7UL) << 1) & 7UL;
ulong step3 = ((exponent % 3UL) << 1) % 3UL;
ulong step5 = ((exponent % 5UL) << 1) % 5UL;
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, (GpuUInt128)twoP, (GpuUInt128)kStart, last, 0UL,
q0m10, q0m8, q0m3, q0m5, orderBuffer.View, smallCyclesView);
kernel(currentSize, exponent, twoPGpu, (GpuUInt128)kStart, last, 0UL,
ra, orderBuffer.View, smallCyclesView, smallPrimesView, smallPrimesPow2View);
accelerator.Synchronize();
orderBuffer.View.CopyToCPU(ref MemoryMarshal.GetReference(orders), currentSize);
accelerator.Synchronize();
if (!Volatile.Read(ref isPrime))
{
break;
}
accelerator.Synchronize();
orderBuffer.View.CopyToCPU(ref MemoryMarshal.GetReference(orders), currentSize);
if (!Volatile.Read(ref isPrime))
{
break;
}
for (i = 0; i < currentSize; i++, q += twoP)
{
if (orders[i] != 0UL && q.IsPrimeCandidate())
{
Volatile.Write(ref isPrime, false);
break;
}
}
for (i = 0; i < currentSize; i++)
{
if (orders[i] != 0UL)
{
Volatile.Write(ref isPrime, false);
break;
}
}
kStart += batchSize128;
}
}
finally
{
ArrayPool<ulong>.Shared.Return(orderArray);
kStart += batchSize128;
}
}
finally
{
ArrayPool<ulong>.Shared.Return(orderArray);
orderBuffer.Dispose();
gpuLease.Dispose();
}