Skip to content

Add AVX and FMA intrinsics in Factorization Machine #3785

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

Add AVX and FMA intrinsics in Factorization Machine #3785

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3695734
Added AVX intrinsics to Factorization Machine
May 28, 2019
3235100
MacOS fix
May 30, 2019
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
1 change: 1 addition & 0 deletions build/Dependencies.props
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,7 @@
<SystemCollectionsImmutableVersion>1.5.0</SystemCollectionsImmutableVersion>
<SystemMemoryVersion>4.5.1</SystemMemoryVersion>
<SystemReflectionEmitLightweightPackageVersion>4.3.0</SystemReflectionEmitLightweightPackageVersion>
<SystemRuntimeCompilerServices>4.5.2</SystemRuntimeCompilerServices>
<SystemThreadingTasksDataflowPackageVersion>4.8.0</SystemThreadingTasksDataflowPackageVersion>
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Any reason we need this line? Could you add a comment to add some details?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi Wei-Sheng! Thanks for reviewing. It’s not needed for now and I will remove it.

</PropertyGroup>

Expand Down
1 change: 1 addition & 0 deletions src/Microsoft.ML.CpuMath/CpuMathUtils.netcoreapp.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@

namespace Microsoft.ML.Internal.CpuMath
{
[BestFriend]
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

BestFriend [](start = 5, length = 10)

Nice!

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why is this change necessary?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi Eric! Access to CpuMathUtils from FactorizationMachineInterface gives an error when using netcoreapp3.0. Issue #3654

internal static partial class CpuMathUtils
{
// The count of bytes in Vector128<T>, corresponding to _cbAlign in AlignedArray
Expand Down
125 changes: 125 additions & 0 deletions ...soft.ML.StandardTrainers/FactorizationMachine/FactorizationMachineInterface.netcoreapp.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,125 @@
using System.Runtime.CompilerServices;
Copy link
Member

@wschin wschin May 30, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Missing license. And why do we need this file?

using System.Runtime.InteropServices;
using System.Runtime.Intrinsics;
using System.Runtime.Intrinsics.X86;
using System.Security;
using Microsoft.ML.Internal.CpuMath;
using Microsoft.ML.Runtime;

namespace Microsoft.ML.Trainers
{
internal static unsafe class FieldAwareFactorizationMachineInterface
{
internal const string NativePath = "FactorizationMachineNative";
public const int CbAlign = 16;

private static bool Compat(AlignedArray a)
{
Contracts.AssertValue(a);
Contracts.Assert(a.Size > 0);
return a.CbAlign == CbAlign;
}

private static unsafe float* Ptr(AlignedArray a, float* p)
{
Contracts.AssertValue(a);
float* q = p + a.GetBase((long)p);
Contracts.Assert(((long)q & (CbAlign - 1)) == 0);
return q;
}

[DllImport(NativePath), SuppressUnmanagedCodeSecurity]
public static extern void CalculateIntermediateVariablesNativeSSE(int fieldCount, int latentDim, int count, int* /*const*/ fieldIndices, int* /*const*/ featureIndices,
float* /*const*/ featureValues, float* /*const*/ linearWeights, float* /*const*/ latentWeights, float* latentSum, float* response);

[DllImport(NativePath), SuppressUnmanagedCodeSecurity]
public static extern void CalculateIntermediateVariablesNativeAVX(int fieldCount, int latentDim, int count, int* /*const*/ fieldIndices, int* /*const*/ featureIndices,
float* /*const*/ featureValues, float* /*const*/ linearWeights, float* /*const*/ latentWeights, float* latentSum, float* response);

[DllImport(NativePath), SuppressUnmanagedCodeSecurity]
public static extern void CalculateIntermediateVariablesNativeFMA(int fieldCount, int latentDim, int count, int* /*const*/ fieldIndices, int* /*const*/ featureIndices,
float* /*const*/ featureValues, float* /*const*/ linearWeights, float* /*const*/ latentWeights, float* latentSum, float* response);

[DllImport(NativePath), SuppressUnmanagedCodeSecurity]
public static extern void CalculateGradientAndUpdateNativeSSE(float lambdaLinear, float lambdaLatent, float learningRate, int fieldCount, int latentDim, float weight,
int count, int* /*const*/ fieldIndices, int* /*const*/ featureIndices, float* /*const*/ featureValues, float* /*const*/ latentSum, float slope,
float* linearWeights, float* latentWeights, float* linearAccumulatedSquaredGrads, float* latentAccumulatedSquaredGrads);

[DllImport(NativePath), SuppressUnmanagedCodeSecurity]
public static extern void CalculateGradientAndUpdateNativeAVX(float lambdaLinear, float lambdaLatent, float learningRate, int fieldCount, int latentDim, float weight,
int count, int* /*const*/ fieldIndices, int* /*const*/ featureIndices, float* /*const*/ featureValues, float* /*const*/ latentSum, float slope,
float* linearWeights, float* latentWeights, float* linearAccumulatedSquaredGrads, float* latentAccumulatedSquaredGrads);

[DllImport(NativePath), SuppressUnmanagedCodeSecurity]
public static extern void CalculateGradientAndUpdateNativeFMA(float lambdaLinear, float lambdaLatent, float learningRate, int fieldCount, int latentDim, float weight,
int count, int* /*const*/ fieldIndices, int* /*const*/ featureIndices, float* /*const*/ featureValues, float* /*const*/ latentSum, float slope,
float* linearWeights, float* latentWeights, float* linearAccumulatedSquaredGrads, float* latentAccumulatedSquaredGrads);

public static void CalculateIntermediateVariables(int fieldCount, int latentDim, int count, int[] fieldIndices, int[] featureIndices, float[] featureValues,
float[] linearWeights, AlignedArray latentWeights, AlignedArray latentSum, ref float response)
{
Contracts.AssertNonEmpty(fieldIndices);
Contracts.AssertNonEmpty(featureValues);
Contracts.AssertNonEmpty(featureIndices);
Contracts.AssertNonEmpty(linearWeights);
Contracts.Assert(Compat(latentWeights));
Contracts.Assert(Compat(latentSum));

unsafe
{
fixed (int* pf = &fieldIndices[0])
fixed (int* pi = &featureIndices[0])
fixed (float* px = &featureValues[0])
fixed (float* pw = &linearWeights[0])
fixed (float* pv = &latentWeights.Items[0])
fixed (float* pq = &latentSum.Items[0])
fixed (float* pr = &response)
{
if (Fma.IsSupported)
CalculateIntermediateVariablesNativeFMA(fieldCount, latentDim, count, pf, pi, px, pw, Ptr(latentWeights, pv), Ptr(latentSum, pq), pr);
else if (Avx.IsSupported)
CalculateIntermediateVariablesNativeAVX(fieldCount, latentDim, count, pf, pi, px, pw, Ptr(latentWeights, pv), Ptr(latentSum, pq), pr);
else
CalculateIntermediateVariablesNativeSSE(fieldCount, latentDim, count, pf, pi, px, pw, Ptr(latentWeights, pv), Ptr(latentSum, pq), pr);
}
}
}

public static void CalculateGradientAndUpdate(float lambdaLinear, float lambdaLatent, float learningRate, int fieldCount, int latentDim,
float weight, int count, int[] fieldIndices, int[] featureIndices, float[] featureValues, AlignedArray latentSum, float slope,
float[] linearWeights, AlignedArray latentWeights, float[] linearAccumulatedSquaredGrads, AlignedArray latentAccumulatedSquaredGrads)
{
Contracts.AssertNonEmpty(fieldIndices);
Contracts.AssertNonEmpty(featureIndices);
Contracts.AssertNonEmpty(featureValues);
Contracts.Assert(Compat(latentSum));
Contracts.AssertNonEmpty(linearWeights);
Contracts.Assert(Compat(latentWeights));
Contracts.AssertNonEmpty(linearAccumulatedSquaredGrads);
Contracts.Assert(Compat(latentAccumulatedSquaredGrads));

unsafe
{
fixed (int* pf = &fieldIndices[0])
fixed (int* pi = &featureIndices[0])
fixed (float* px = &featureValues[0])
fixed (float* pq = &latentSum.Items[0])
fixed (float* pw = &linearWeights[0])
fixed (float* pv = &latentWeights.Items[0])
fixed (float* phw = &linearAccumulatedSquaredGrads[0])
fixed (float* phv = &latentAccumulatedSquaredGrads.Items[0])
{
if (Fma.IsSupported)
CalculateGradientAndUpdateNativeFMA(lambdaLinear, lambdaLatent, learningRate, fieldCount, latentDim, weight, count, pf, pi, px,
Ptr(latentSum, pq), slope, pw, Ptr(latentWeights, pv), phw, Ptr(latentAccumulatedSquaredGrads, phv));
else if (Avx.IsSupported)
CalculateGradientAndUpdateNativeAVX(lambdaLinear, lambdaLatent, learningRate, fieldCount, latentDim, weight, count, pf, pi, px,
Ptr(latentSum, pq), slope, pw, Ptr(latentWeights, pv), phw, Ptr(latentAccumulatedSquaredGrads, phv));
else
CalculateGradientAndUpdateNativeSSE(lambdaLinear, lambdaLatent, learningRate, fieldCount, latentDim, weight, count, pf, pi, px,
Ptr(latentSum, pq), slope, pw, Ptr(latentWeights, pv), phw, Ptr(latentAccumulatedSquaredGrads, phv));
}
}
}
}
}
10 changes: 5 additions & 5 deletions ...nMachine/FactorizationMachineInterface.cs → ...torizationMachineInterface.netstandard.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,11 +30,11 @@ private static bool Compat(AlignedArray a)
}

[DllImport(NativePath), SuppressUnmanagedCodeSecurity]
public static extern void CalculateIntermediateVariablesNative(int fieldCount, int latentDim, int count, int* /*const*/ fieldIndices, int* /*const*/ featureIndices,
public static extern void CalculateIntermediateVariablesNativeSSE(int fieldCount, int latentDim, int count, int* /*const*/ fieldIndices, int* /*const*/ featureIndices,
Copy link
Member

@wschin wschin May 30, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

CalculateIntermediateVariablesNativeSSE ---> CalculateIntermediateVariablesSse

float* /*const*/ featureValues, float* /*const*/ linearWeights, float* /*const*/ latentWeights, float* latentSum, float* response);

[DllImport(NativePath), SuppressUnmanagedCodeSecurity]
public static extern void CalculateGradientAndUpdateNative(float lambdaLinear, float lambdaLatent, float learningRate, int fieldCount, int latentDim, float weight,
public static extern void CalculateGradientAndUpdateNativeSSE(float lambdaLinear, float lambdaLatent, float learningRate, int fieldCount, int latentDim, float weight,
Copy link
Member

@wschin wschin May 30, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

CalculateGradientAndUpdateNativeSSE [](start = 34, length = 35)

CalculateGradientAndUpdateNativeSSE ---> CalculateGradientAndUpdateSse

The same comment are applicable to other similar places.

int count, int* /*const*/ fieldIndices, int* /*const*/ featureIndices, float* /*const*/ featureValues, float* /*const*/ latentSum, float slope,
float* linearWeights, float* latentWeights, float* linearAccumulatedSquaredGrads, float* latentAccumulatedSquaredGrads);

Expand All @@ -57,7 +57,7 @@ public static void CalculateIntermediateVariables(int fieldCount, int latentDim,
fixed (float* pv = &latentWeights.Items[0])
fixed (float* pq = &latentSum.Items[0])
fixed (float* pr = &response)
CalculateIntermediateVariablesNative(fieldCount, latentDim, count, pf, pi, px, pw, Ptr(latentWeights, pv), Ptr(latentSum, pq), pr);
CalculateIntermediateVariablesNativeSSE(fieldCount, latentDim, count, pf, pi, px, pw, Ptr(latentWeights, pv), Ptr(latentSum, pq), pr);
}
}

Expand All @@ -84,10 +84,10 @@ public static void CalculateGradientAndUpdate(float lambdaLinear, float lambdaLa
fixed (float* pv = &latentWeights.Items[0])
fixed (float* phw = &linearAccumulatedSquaredGrads[0])
fixed (float* phv = &latentAccumulatedSquaredGrads.Items[0])
CalculateGradientAndUpdateNative(lambdaLinear, lambdaLatent, learningRate, fieldCount, latentDim, weight, count, pf, pi, px,
CalculateGradientAndUpdateNativeSSE(lambdaLinear, lambdaLatent, learningRate, fieldCount, latentDim, weight, count, pf, pi, px,
Ptr(latentSum, pq), slope, pw, Ptr(latentWeights, pv), phw, Ptr(latentAccumulatedSquaredGrads, phv));
}

}
}
}
}
14 changes: 13 additions & 1 deletion src/Microsoft.ML.StandardTrainers/Microsoft.ML.StandardTrainers.csproj
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,8 @@
<Project Sdk="Microsoft.NET.Sdk">

<PropertyGroup>
<TargetFramework>netstandard2.0</TargetFramework>
<TargetFramework Condition="'$(UseIntrinsics)' != 'true'">netstandard2.0</TargetFramework>
Copy link
Member

@wschin wschin May 30, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@eerhardt, could you please take a look? I am not quite familiar with the build settings. Thank you!

<TargetFrameworks Condition="'$(UseIntrinsics)' == 'true'">netstandard2.0;netcoreapp3.0</TargetFrameworks>
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The NuGet package is not going to work the way this is currently written. See #534 for more information. Basically, once you introduce a single assembly in lib\netcoreapp3.0 of the NuGet package, ALL the assemblies need to be in the lib\netcoreapp3.0 folder of the package.

It would probably be easiest if the new AVX/FMA were put in the CpuMath assembly/package instead.

<IncludeInPackage>Microsoft.ML</IncludeInPackage>
<AllowUnsafeBlocks>true</AllowUnsafeBlocks>
</PropertyGroup>
Expand All @@ -10,6 +11,17 @@
<ProjectReference Include="..\Microsoft.ML.Core\Microsoft.ML.Core.csproj" />
<ProjectReference Include="..\Microsoft.ML.CpuMath\Microsoft.ML.CpuMath.csproj" />
<ProjectReference Include="..\Microsoft.ML.Data\Microsoft.ML.Data.csproj" />
<!-- Workaround https://github.com/dotnet/project-system/issues/935 -->
<None Include="**/*.cs" />
</ItemGroup>

<ItemGroup Condition="'$(TargetFramework)' == 'netcoreapp3.0'">
<Compile Remove="FactorizationMachine/FactorizationMachineInterface.netstandard.cs" />
<PackageReference Include="System.Runtime.CompilerServices.Unsafe" Version="$(SystemRuntimeCompilerServices)" />
</ItemGroup>

<ItemGroup Condition="'$(TargetFramework)' == 'netstandard2.0'">
<Compile Remove="FactorizationMachine/FactorizationMachineInterface.netcoreapp.cs" />
</ItemGroup>

</Project>
6 changes: 5 additions & 1 deletion src/Native/FactorizationMachineNative/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,12 +1,16 @@
project (FactorizationMachineNative)

set(SOURCES
FactorizationMachineCore.cpp
FactorizationMachineCoreSSE.cpp
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Are we compiling all of source files? Are all of them cross-platform?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, we are compiling all the source files. Not all of them are cross-platform. Avx.isSupported (and similar) dispatches the code as needed.

FactorizationMachineCoreAVX.cpp
FactorizationMachineCoreFMA.cpp
)

if(WIN32)
else()
list(APPEND SOURCES ${VERSION_FILE_PATH})
set_property(SOURCE FactorizationMachineCoreAVX.cpp APPEND_STRING PROPERTY COMPILE_FLAGS " -mavx")
set_property(SOURCE FactorizationMachineCoreFMA.cpp APPEND_STRING PROPERTY COMPILE_FLAGS " -mfma")
endif()

add_library(FactorizationMachineNative SHARED ${SOURCES} ${RESOURCES})
Expand Down
Loading