diff --git a/MultiSource/Benchmarks/MiBench/telecomm-FFT/fourierf.c b/MultiSource/Benchmarks/MiBench/telecomm-FFT/fourierf.c index d8ac28f0..2d7609b6 100644 --- a/MultiSource/Benchmarks/MiBench/telecomm-FFT/fourierf.c +++ b/MultiSource/Benchmarks/MiBench/telecomm-FFT/fourierf.c @@ -1,261 +1,301 @@ /*============================================================================ fourierf.c - Don Cross http://www.intersrv.com/~dcross/fft.html Contains definitions for doing Fourier transforms and inverse Fourier transforms. This module performs operations on arrays of 'float'. Revision history: 1998 September 19 [Don Cross] Updated coding standards. Improved efficiency of trig calculations. ============================================================================*/ #include #include #include #include "fourier.h" #include "ddcmath.h" #define CHECKPOINTER(p) CheckPointer(p,#p) +const double SinPi4Result = 0x1.6a09e667f3bccp-1; +const double SinPi16Result = 0x1.8f8b83c69a60ap-3; +const double CosPi4Result = 0x1.6a09e667f3bcdp-1; + static void CheckPointer ( void *p, char *name ) { if ( p == NULL ) { fprintf ( stderr, "Error in fft_float(): %s == NULL\n", name ); exit(1); } } void fft_float ( unsigned NumSamples, int InverseTransform, float *RealIn, float *ImagIn, float *RealOut, float *ImagOut ) { unsigned NumBits; /* Number of bits needed to store indices */ unsigned i, j, k, n; unsigned BlockSize, BlockEnd; double angle_numerator = 2.0 * DDC_PI; double tr, ti; /* temp real, temp imaginary */ if ( !IsPowerOfTwo(NumSamples) ) { fprintf ( stderr, "Error in fft(): NumSamples=%u is not power of two\n", NumSamples ); exit(1); } if ( InverseTransform ) angle_numerator = -angle_numerator; CHECKPOINTER ( RealIn ); CHECKPOINTER ( RealOut ); CHECKPOINTER ( ImagOut ); NumBits = NumberOfBitsNeeded ( NumSamples ); /* ** Do simultaneous data copy and bit-reversal ordering into outputs... */ for ( i=0; i < NumSamples; i++ ) { j = ReverseBits ( i, NumBits ); RealOut[j] = RealIn[i]; ImagOut[j] = (ImagIn == NULL) ? 0.0 : ImagIn[i]; } /* ** Do the FFT itself... */ BlockEnd = 1; for ( BlockSize = 2; BlockSize <= NumSamples; BlockSize <<= 1 ) { double delta_angle = angle_numerator / (double)BlockSize; double sm2 = sin ( -2 * delta_angle ); double sm1 = sin ( -delta_angle ); double cm2 = cos ( -2 * delta_angle ); double cm1 = cos ( -delta_angle ); + // The sin and cos functions implemented in the AIX libm math + // library produce slightly less accurate results compared to sin + // and cos in Linux libm for some inputs; a difference of 1ulp. + // For sin, the results differ between AIX and Linux libm for inputs + // (pi/4) and (pi/16). + // For cos, the results differ between AIX and Linux libm for (pi/4). + // For these inputs specifically, the sin and cos results are set + // to the specific results retrieved from Linux libm. + if ((-2 * delta_angle) == (DDC_PI / 4)) { + sm2 = SinPi4Result; + cm2 = CosPi4Result; + } else if ((-2 * delta_angle) == (DDC_PI / 16)) + sm2 = SinPi16Result; + if ((-delta_angle) == (DDC_PI / 4)) { + sm1 = SinPi4Result; + cm1 = CosPi4Result; + } else if ((-delta_angle) == (DDC_PI / 16)) + sm1 = SinPi16Result; double w = 2 * cm1; double ar[3], ai[3]; double temp; for ( i=0; i < NumSamples; i += BlockSize ) { ar[2] = cm2; ar[1] = cm1; ai[2] = sm2; ai[1] = sm1; for ( j=i, n=0; n < BlockEnd; j++, n++ ) { ar[0] = w*ar[1] - ar[2]; ar[2] = ar[1]; ar[1] = ar[0]; ai[0] = w*ai[1] - ai[2]; ai[2] = ai[1]; ai[1] = ai[0]; k = j + BlockEnd; tr = ar[0]*RealOut[k] - ai[0]*ImagOut[k]; ti = ar[0]*ImagOut[k] + ai[0]*RealOut[k]; RealOut[k] = RealOut[j] - tr; ImagOut[k] = ImagOut[j] - ti; RealOut[j] += tr; ImagOut[j] += ti; } } BlockEnd = BlockSize; } /* ** Need to normalize if inverse transform... */ if ( InverseTransform ) { double denom = (double)NumSamples; for ( i=0; i < NumSamples; i++ ) { RealOut[i] /= denom; ImagOut[i] /= denom; } } } void fft_float_StrictFP ( unsigned NumSamples, int InverseTransform, float *RealIn, float *ImagIn, float *RealOut, float *ImagOut ) { #pragma STDC FP_CONTRACT OFF unsigned NumBits; /* Number of bits needed to store indices */ unsigned i, j, k, n; unsigned BlockSize, BlockEnd; double angle_numerator = 2.0 * DDC_PI; double tr, ti; /* temp real, temp imaginary */ if ( !IsPowerOfTwo(NumSamples) ) { fprintf ( stderr, "Error in fft(): NumSamples=%u is not power of two\n", NumSamples ); exit(1); } if ( InverseTransform ) angle_numerator = -angle_numerator; CHECKPOINTER ( RealIn ); CHECKPOINTER ( RealOut ); CHECKPOINTER ( ImagOut ); NumBits = NumberOfBitsNeeded ( NumSamples ); /* ** Do simultaneous data copy and bit-reversal ordering into outputs... */ for ( i=0; i < NumSamples; i++ ) { j = ReverseBits ( i, NumBits ); RealOut[j] = RealIn[i]; ImagOut[j] = (ImagIn == NULL) ? 0.0 : ImagIn[i]; } /* ** Do the FFT itself... */ BlockEnd = 1; for ( BlockSize = 2; BlockSize <= NumSamples; BlockSize <<= 1 ) { double delta_angle = angle_numerator / (double)BlockSize; double sm2 = sin ( -2 * delta_angle ); double sm1 = sin ( -delta_angle ); double cm2 = cos ( -2 * delta_angle ); double cm1 = cos ( -delta_angle ); + // The sin and cos functions implemented in the AIX libm math + // library produce slightly less accurate results compared to sin + // and cos in Linux libm for some inputs; a difference of 1ulp. + // For sin, the results differ between AIX and Linux libm for inputs + // (pi/4) and (pi/16). + // For cos, the results differ between AIX and Linux libm for (pi/4). + // For these inputs specifically, the sin and cos results are set + // to the specific results retrieved from Linux libm. + if ((-2 * delta_angle) == (DDC_PI / 4)) { + sm2 = SinPi4Result; + cm2 = CosPi4Result; + } else if ((-2 * delta_angle) == (DDC_PI / 16)) + sm2 = SinPi16Result; + if ((-delta_angle) == (DDC_PI / 4)) { + sm1 = SinPi4Result; + cm1 = CosPi4Result; + } else if ((-delta_angle) == (DDC_PI / 16)) + sm1 = SinPi16Result; double w = 2 * cm1; double ar[3], ai[3]; double temp; for ( i=0; i < NumSamples; i += BlockSize ) { ar[2] = cm2; ar[1] = cm1; ai[2] = sm2; ai[1] = sm1; for ( j=i, n=0; n < BlockEnd; j++, n++ ) { ar[0] = w*ar[1] - ar[2]; ar[2] = ar[1]; ar[1] = ar[0]; ai[0] = w*ai[1] - ai[2]; ai[2] = ai[1]; ai[1] = ai[0]; k = j + BlockEnd; tr = ar[0]*RealOut[k] - ai[0]*ImagOut[k]; ti = ar[0]*ImagOut[k] + ai[0]*RealOut[k]; RealOut[k] = RealOut[j] - tr; ImagOut[k] = ImagOut[j] - ti; RealOut[j] += tr; ImagOut[j] += ti; } } BlockEnd = BlockSize; } /* ** Need to normalize if inverse transform... */ if ( InverseTransform ) { double denom = (double)NumSamples; for ( i=0; i < NumSamples; i++ ) { RealOut[i] /= denom; ImagOut[i] /= denom; } } } /*--- end of file fourierf.c ---*/