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

Streams added to all kernels and GPU operations to separate threads from each other.

Browse Source
This commit is contained in:
mlemanczyk committed 2025-11-06 18:53:55 +01:00
1 parent 64d615c402
commit de2a78b27b
30 files changed
+1097 -1002
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

PerfectNumbers.Core/Gpu/Kernels/NttMontgomeryKernels.cs
+3 -2
View File
@@ -1,5 +1,6 @@
using System.Runtime.CompilerServices;
using ILGPU;
using ILGPU.Runtime;
namespace PerfectNumbers.Core.Gpu;
@@ -29,7 +30,7 @@ internal static class NttMontgomeryKernels
data[i2] = new GpuUInt128(0UL, diff);
}
public static void ToMont64Kernel(Index1D index, ArrayView<GpuUInt128> data, ulong modulus, ulong nPrime, ulong r2)
public static void ToMont64Kernel(Index1D index, AcceleratorStream stream, ArrayView<GpuUInt128> data, ulong modulus, ulong nPrime, ulong r2)
{
ulong a = data[index].Low;
ulong v = MontMul64(a, r2, modulus, nPrime);
@@ -43,7 +44,7 @@ internal static class NttMontgomeryKernels
data[index] = new GpuUInt128(0UL, v);
}
public static void SquareMont64Kernel(Index1D index, ArrayView<GpuUInt128> data, ulong modulus, ulong nPrime)
public static void SquareMont64Kernel(Index1D index, AcceleratorStream stream, ArrayView<GpuUInt128> data, ulong modulus, ulong nPrime)
{
ulong aR = data[index].Low;
ulong v = MontMul64(aR, aR, modulus, nPrime);