Skip to content

Improve Math.BigMul to fix Decimal compare perf regression #115182

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 2 commits into from
Closed

Improve Math.BigMul to fix Decimal compare perf regression #115182

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2c89d1b
Improve BigMul to fix VarDecCmpSub regression #11243
Daniel-Svensson Mar 6, 2025
18482a8
Fix typo
Daniel-Svensson Apr 30, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 13 additions & 20 deletions src/libraries/System.Private.CoreLib/src/System/Math.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -159,16 +159,6 @@ internal static void ThrowNegateTwosCompOverflow()
[MethodImpl(MethodImplOptions.AggressiveInlining)]
public static unsafe ulong BigMul(uint a, uint b)
{
#if false // TARGET_32BIT
// This generates slower code currently than the simple multiplication
// https://github.com/dotnet/runtime/issues/11782
if (Bmi2.IsSupported)
{
uint low;
uint high = Bmi2.MultiplyNoFlags(a, b, &low);
return ((ulong)high << 32) | low;
}
#endif
return ((ulong)a) * b;
}

Expand All @@ -181,7 +171,7 @@ public static long BigMul(int a, int b)
return ((long)a) * b;
}


#if !TARGET_ARM64
/// <summary>
/// Perform multiplication between 64 and 32 bit numbers, returning lower 64 bits in <paramref name="low"/>
/// </summary>
Expand All @@ -190,21 +180,25 @@ public static long BigMul(int a, int b)
[MethodImpl(MethodImplOptions.AggressiveInlining)]
internal static ulong BigMul(ulong a, uint b, out ulong low)
{
#if TARGET_64BIT
return Math.BigMul((ulong)a, (ulong)b, out low);
#else
if (Bmi2.X64.IsSupported)
{
// Ideally, we should do a single multiply for low and high. Doing this efficiently would mean making BigMul an intrinsic (would work on all x64 targets) or fixing https://github.com/dotnet/runtime/issues/11782
low = a * b;
return Bmi2.X64.MultiplyNoFlags(a, b);
}

ulong prodL = ((ulong)(uint)a) * b;
ulong prodH = (prodL >> 32) + (((ulong)(uint)(a >> 32)) * b);

low = ((prodH << 32) | (uint)prodL);
return (prodH >> 32);
#endif
}

/// <inheritdoc cref="BigMul(ulong, uint, out ulong)"/>
[MethodImpl(MethodImplOptions.AggressiveInlining)]
internal static ulong BigMul(uint a, ulong b, out ulong low)
=> BigMul(b, a, out low);
#endif

/// <summary>Produces the full product of two unsigned 64-bit numbers.</summary>
/// <param name="a">The first number to multiply.</param>
Expand All @@ -213,14 +207,13 @@ internal static ulong BigMul(uint a, ulong b, out ulong low)
/// <returns>The high 64-bit of the product of the specified numbers.</returns>
[CLSCompliant(false)]
[MethodImpl(MethodImplOptions.AggressiveInlining)]
public static unsafe ulong BigMul(ulong a, ulong b, out ulong low)
public static ulong BigMul(ulong a, ulong b, out ulong low)
{
if (Bmi2.X64.IsSupported)
{
ulong tmp;
ulong high = Bmi2.X64.MultiplyNoFlags(a, b, &tmp);
low = tmp;
return high;
// Ideally, we should do a single multiply for low and high. Doing this efficiently would mean making BigMul an intrinsic (would work on all x64 targets) or fixing https://github.com/dotnet/runtime/issues/11782
low = a * b;
return Bmi2.X64.MultiplyNoFlags(a, b);
}
else if (ArmBase.Arm64.IsSupported)
{
Expand Down
Loading