Partager via


Intrinsic Functions for Renesas Microprocessors (Windows Embedded CE 6.0)

1/5/2010

The Renesas SH-4 microprocessor includes a number of functions that you can use intrinsically for faster programs. These functions aid mathematical computations.

Programs that use intrinsic functions do not have the overhead of function calls, but they can be larger because of the additional code created.

To replace functions with their intrinsic (inline) forms, use the /Oi (Generate Intrinsic Functions) option or the #pragma intrinsic.

The following table shows SH-4 microprocessor assembly language functions implemented as intrinsic functions.

SH-4 instruction Description

_Convolve

Computes the summation of two vectors.

_Dot3dVW0

Computes the inner product of a pair of three- or four-dimensional vectors with the "w" coordinate forced to 0.

_Dot3dVW1

Computes the inner product of a pair of three- or four-dimensional vectors with the "w" coordinate forced to 1.

_Dot4dV

Computes the inner product of a pair of four-dimensional vectors.

__fmac

Multiplies two float values and adds to a third float value.

_LoadMatrix

Loads a 4x4 matrix into the extended floating-point register bank.

__movca

Moves with cache block allocation.

_Multiply4dM

Multiplies a 4x4 matrix by a 4x4 matrix.

_SaveMatrix

Stores the extended floating-point register bank into a 4x4 matrix.

_XDMultMatrix

Multiplies a 4x4 matrix by a 4x4 matrix.

_XDXform3dV

Multiplies a three-dimensional vector by a 3x3 matrix.

_XDXform4dV

Multiplies a four-dimensional vector by a 4x4 matrix.

_Xform3dV

Multiplies a three-dimensional vector by a 3x3 matrix.

_Xform4dV

Multiplies a four-dimensional vector by a 4x4 matrix.

See Also

Other Resources

Renesas Family Processors