The NeonBackendModelContext is used to pass in Neon specific backend ModelOptions. More...

#include <NeonBackendModelContext.hpp>

Inheritance diagram for NeonBackendModelContext:

Collaboration diagram for NeonBackendModelContext:

Public Member Functions
	NeonBackendModelContext (const ModelOptions &modelOptions)

bool	IsFastMathEnabled () const

unsigned int	GetNumberOfThreads () const

Public Member Functions inherited from IBackendModelContext
virtual	~IBackendModelContext ()

Detailed Description

The NeonBackendModelContext is used to pass in Neon specific backend ModelOptions.

The supported backend ModelOptions are:

"FastMathEnabled"
Using the fast_math flag can lead to performance improvements in fp32 and fp16 layers but may result in
results with reduced or different precision. The fast_math flag will not have any effect on int8 performance.
"NumberOfThreads"
Specify the number of threads used by the CpuAcc backend.

Definition at line 19 of file NeonBackendModelContext.hpp.

Constructor & Destructor Documentation

◆ NeonBackendModelContext()

NeonBackendModelContext ( const ModelOptions & modelOptions )

Definition at line 34 of file NeonBackendModelContext.cpp.

     : m_IsFastMathEnabled(false), m_NumberOfThreads(0)
 {
    if (!modelOptions.empty())
    {
        ParseOptions(modelOptions, "CpuAcc", [&](std::string name, const BackendOptions::Var& value)
        {
            if (name == "FastMathEnabled")
            {
                m_IsFastMathEnabled |= ParseBool(value, false);
            }
            if (name == "NumberOfThreads")
            {
                m_NumberOfThreads |= ParseUnsignedInt(value, 0);
            }
        });
    }
 }

References armnn::ParseOptions().

Member Function Documentation

◆ GetNumberOfThreads()

unsigned int GetNumberOfThreads ( ) const

Definition at line 58 of file NeonBackendModelContext.cpp.

 {
     return m_NumberOfThreads;
 }

◆ IsFastMathEnabled()

bool IsFastMathEnabled ( ) const

Definition at line 53 of file NeonBackendModelContext.cpp.

 {
     return m_IsFastMathEnabled;
 }

Referenced by NeonWorkloadFactory::CreateWorkload(), NeonLayerSupport::IsBatchMatMulSupported(), NeonLayerSupport::IsConvolution2dSupported(), and NeonLayerSupport::IsConvolution3dSupported().

The documentation for this class was generated from the following files: