aboutsummaryrefslogtreecommitdiff
path: root/source/testfloat_ver.c
diff options
context:
space:
mode:
Diffstat (limited to 'source/testfloat_ver.c')
-rw-r--r--source/testfloat_ver.c622
1 files changed, 622 insertions, 0 deletions
diff --git a/source/testfloat_ver.c b/source/testfloat_ver.c
new file mode 100644
index 0000000..97edcbf
--- /dev/null
+++ b/source/testfloat_ver.c
@@ -0,0 +1,622 @@
+
+/*============================================================================
+
+This C source file is part of TestFloat, Release 3, a package of programs for
+testing the correctness of floating-point arithmetic complying with the IEEE
+Standard for Floating-Point, by John R. Hauser.
+
+Copyright 2011, 2012, 2013, 2014 The Regents of the University of California
+(Regents). All Rights Reserved. Redistribution and use in source and binary
+forms, with or without modification, are permitted provided that the following
+conditions are met:
+
+Redistributions of source code must retain the above copyright notice,
+this list of conditions, and the following two paragraphs of disclaimer.
+Redistributions in binary form must reproduce the above copyright notice,
+this list of conditions, and the following two paragraphs of disclaimer in the
+documentation and/or other materials provided with the distribution. Neither
+the name of the Regents nor the names of its contributors may be used to
+endorse or promote products derived from this software without specific prior
+written permission.
+
+IN NO EVENT SHALL REGENTS BE LIABLE TO ANY PARTY FOR DIRECT, INDIRECT,
+SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES, INCLUDING LOST PROFITS, ARISING
+OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF REGENTS HAS
+BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+REGENTS SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING, BUT NOT LIMITED
+TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE. THE SOFTWARE AND ACCOMPANYING DOCUMENTATION, IF ANY, PROVIDED
+HEREUNDER IS PROVIDED "AS IS". REGENTS HAS NO OBLIGATION TO PROVIDE
+MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
+
+=============================================================================*/
+
+#include <stdbool.h>
+#include <stdint.h>
+#include <string.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <signal.h>
+#include "platform.h"
+#include "fail.h"
+#include "softfloat.h"
+#include "functions.h"
+#include "verCases.h"
+#include "writeCase.h"
+#include "verLoops.h"
+
+static void catchSIGINT( int signalCode )
+{
+
+ if ( verCases_stop ) exit( EXIT_FAILURE );
+ verCases_stop = true;
+
+}
+
+int main( int argc, char *argv[] )
+{
+ bool exact;
+ int functionCode, roundingCode, tininessCode;
+ const char *argPtr;
+ long i;
+ int functionAttribs;
+ uint_fast8_t roundingMode;
+ float32_t (*trueFunction_abz_f32)( float32_t, float32_t );
+ bool (*trueFunction_ab_f32_z_bool)( float32_t, float32_t );
+ float64_t (*trueFunction_abz_f64)( float64_t, float64_t );
+ bool (*trueFunction_ab_f64_z_bool)( float64_t, float64_t );
+#ifdef EXTFLOAT80
+ void
+ (*trueFunction_abz_extF80)(
+ const extFloat80_t *, const extFloat80_t *, extFloat80_t * );
+ bool
+ (*trueFunction_ab_extF80_z_bool)(
+ const extFloat80_t *, const extFloat80_t * );
+#endif
+#ifdef FLOAT128
+ void
+ (*trueFunction_abz_f128)(
+ const float128_t *, const float128_t *, float128_t * );
+ bool
+ (*trueFunction_ab_f128_z_bool)( const float128_t *, const float128_t * );
+#endif
+
+ /*------------------------------------------------------------------------
+ *------------------------------------------------------------------------*/
+ fail_programName = "testfloat_ver";
+ if ( argc <= 1 ) goto writeHelpMessage;
+ softfloat_detectTininess = softfloat_tininess_beforeRounding;
+#ifdef EXTFLOAT80
+ extF80_roundingPrecision = 80;
+#endif
+ roundingMode = softfloat_round_near_even;
+ exact = false;
+ verCases_maxErrorCount = 20;
+ verLoops_trueFlagsPtr = &softfloat_exceptionFlags;
+ functionCode = 0;
+ roundingCode = ROUND_NEAR_EVEN;
+ tininessCode = TININESS_BEFORE_ROUNDING;
+ for (;;) {
+ --argc;
+ if ( ! argc ) break;
+ argPtr = *++argv;
+ if ( ! argPtr ) break;
+ if ( argPtr[0] == '-' ) ++argPtr;
+ if (
+ ! strcmp( argPtr, "help" ) || ! strcmp( argPtr, "-help" )
+ || ! strcmp( argPtr, "h" )
+ ) {
+ writeHelpMessage:
+ fputs(
+"testfloat_ver [<option>...] <function>\n"
+" <option>: (* is default)\n"
+" -help --Write this message and exit.\n"
+" -errors <num> --Stop after <num> errors.\n"
+" * -errors 20\n"
+" -checkNaNs --Check for bitwise correctness of NaN results.\n"
+#ifdef EXTFLOAT80
+" -precision32 --For extF80, rounding precision is 32 bits.\n"
+" -precision64 --For extF80, rounding precision is 64 bits.\n"
+" * -precision80 --For extF80, rounding precision is 80 bits.\n"
+#endif
+" * -rnear_even --Round to nearest/even.\n"
+" -rminMag --Round to minimum magnitude (toward zero).\n"
+" -rmin --Round to minimum (down).\n"
+" -rmax --Round to maximum (up).\n"
+" -rnear_maxMag --Round to nearest/maximum magnitude (nearest/away).\n"
+" * -tininessbefore --Detect underflow tininess before rounding.\n"
+" -tininessafter --Detect underflow tininess after rounding.\n"
+" * -notexact --Rounding to integer is not exact (no inexact\n"
+" exceptions).\n"
+" -exact --Rounding to integer is exact (raising inexact\n"
+" exceptions).\n"
+" <function>:\n"
+" <int>_to_<float> <float>_add <float>_eq\n"
+" <float>_to_<int> <float>_sub <float>_le\n"
+" <float>_to_<float> <float>_mul <float>_lt\n"
+" <float>_roundToInt <float>_mulAdd <float>_eq_signaling\n"
+" <float>_div <float>_le_quiet\n"
+" <float>_rem <float>_lt_quiet\n"
+" <float>_sqrt\n"
+" <int>:\n"
+" ui32 --Unsigned 32-bit integer.\n"
+" ui64 --Unsigned 64-bit integer.\n"
+" i32 --Signed 32-bit integer.\n"
+" i64 --Signed 64-bit integer.\n"
+" <float>:\n"
+" f32 --Binary 32-bit floating-point (single-precision).\n"
+" f64 --Binary 64-bit floating-point (double-precision).\n"
+#ifdef EXTFLOAT80
+" extF80 --Binary 80-bit extended floating-point.\n"
+#endif
+#ifdef FLOAT128
+" f128 --Binary 128-bit floating-point (quadruple-precision).\n"
+#endif
+ ,
+ stdout
+ );
+ return EXIT_SUCCESS;
+ } else if ( ! strcmp( argPtr, "errors" ) ) {
+ if ( argc < 2 ) goto optionError;
+ i = strtol( argv[1], (char **) &argPtr, 10 );
+ if ( *argPtr ) goto optionError;
+ verCases_maxErrorCount = i;
+ --argc;
+ ++argv;
+ } else if (
+ ! strcmp( argPtr, "checkNaNs" ) || ! strcmp( argPtr, "checknans" )
+ ) {
+ verCases_checkNaNs = true;
+#ifdef EXTFLOAT80
+ } else if ( ! strcmp( argPtr, "precision32" ) ) {
+ extF80_roundingPrecision = 32;
+ } else if ( ! strcmp( argPtr, "precision64" ) ) {
+ extF80_roundingPrecision = 64;
+ } else if ( ! strcmp( argPtr, "precision80" ) ) {
+ extF80_roundingPrecision = 80;
+#endif
+ } else if (
+ ! strcmp( argPtr, "rnear_even" )
+ || ! strcmp( argPtr, "rneareven" )
+ || ! strcmp( argPtr, "rnearest_even" )
+ ) {
+ roundingCode = ROUND_NEAR_EVEN;
+ } else if (
+ ! strcmp( argPtr, "rminmag" ) || ! strcmp( argPtr, "rminMag" )
+ ) {
+ roundingCode = ROUND_MINMAG;
+ } else if ( ! strcmp( argPtr, "rmin" ) ) {
+ roundingCode = ROUND_MIN;
+ } else if ( ! strcmp( argPtr, "rmax" ) ) {
+ roundingCode = ROUND_MAX;
+ } else if (
+ ! strcmp( argPtr, "rnear_maxmag" )
+ || ! strcmp( argPtr, "rnear_maxMag" )
+ || ! strcmp( argPtr, "rnearmaxmag" )
+ || ! strcmp( argPtr, "rnearest_maxmag" )
+ || ! strcmp( argPtr, "rnearest_maxMag" )
+ ) {
+ roundingCode = ROUND_NEAR_MAXMAG;
+ } else if ( ! strcmp( argPtr, "tininessbefore" ) ) {
+ tininessCode = TININESS_BEFORE_ROUNDING;
+ } else if ( ! strcmp( argPtr, "tininessafter" ) ) {
+ tininessCode = TININESS_AFTER_ROUNDING;
+ } else if ( ! strcmp( argPtr, "notexact" ) ) {
+ exact = false;
+ } else if ( ! strcmp( argPtr, "exact" ) ) {
+ exact = true;
+ } else {
+ functionCode = 1;
+ while ( strcmp( argPtr, functionInfos[functionCode].namePtr ) ) {
+ ++functionCode;
+ if ( functionCode == NUM_FUNCTIONS ) goto invalidArg;
+ }
+ functionAttribs = functionInfos[functionCode].attribs;
+ if (
+ (functionAttribs & FUNC_ARG_EXACT)
+ && ! (functionAttribs & FUNC_ARG_ROUNDINGMODE)
+ ) {
+ goto invalidArg;
+ }
+ }
+ }
+ if ( ! functionCode ) fail( "Function argument required" );
+ /*------------------------------------------------------------------------
+ *------------------------------------------------------------------------*/
+ verCases_functionNamePtr = functionInfos[functionCode].namePtr;
+#ifdef EXTFLOAT80
+ verCases_roundingPrecision =
+ functionAttribs & FUNC_EFF_ROUNDINGPRECISION ? extF80_roundingPrecision
+ : 0;
+#endif
+ verCases_roundingCode =
+ functionAttribs & (FUNC_ARG_ROUNDINGMODE | FUNC_EFF_ROUNDINGMODE)
+ ? roundingCode
+ : 0;
+ verCases_tininessCode =
+ functionAttribs
+ & (extF80_roundingPrecision && (extF80_roundingPrecision < 80)
+ ? FUNC_EFF_TININESSMODE_REDUCEDPREC
+ : FUNC_EFF_TININESSMODE)
+ ? tininessCode
+ : 0;
+ verCases_usesExact = ((functionAttribs & FUNC_ARG_EXACT) != 0);
+ verCases_exact = exact;
+ roundingMode = roundingModes[roundingCode];
+ softfloat_roundingMode = roundingMode;
+ softfloat_detectTininess = tininessModes[tininessCode];
+ signal( SIGINT, catchSIGINT );
+ signal( SIGTERM, catchSIGINT );
+ fputs( "Testing ", stderr );
+ verCases_writeFunctionName( stderr );
+ fputs( ".\n", stderr );
+ switch ( functionCode ) {
+ /*--------------------------------------------------------------------
+ *--------------------------------------------------------------------*/
+ case UI32_TO_F32:
+ ver_a_ui32_z_f32( ui32_to_f32 );
+ break;
+ case UI32_TO_F64:
+ ver_a_ui32_z_f64( ui32_to_f64 );
+ break;
+#ifdef EXTFLOAT80
+ case UI32_TO_EXTF80:
+ ver_a_ui32_z_extF80( ui32_to_extF80M );
+ break;
+#endif
+#ifdef FLOAT128
+ case UI32_TO_F128:
+ ver_a_ui32_z_f128( ui32_to_f128M );
+ break;
+#endif
+ case UI64_TO_F32:
+ ver_a_ui64_z_f32( ui64_to_f32 );
+ break;
+ case UI64_TO_F64:
+ ver_a_ui64_z_f64( ui64_to_f64 );
+ break;
+#ifdef EXTFLOAT80
+ case UI64_TO_EXTF80:
+ ver_a_ui64_z_extF80( ui64_to_extF80M );
+ break;
+#endif
+#ifdef FLOAT128
+ case UI64_TO_F128:
+ ver_a_ui64_z_f128( ui64_to_f128M );
+ break;
+#endif
+ case I32_TO_F32:
+ ver_a_i32_z_f32( i32_to_f32 );
+ break;
+ case I32_TO_F64:
+ ver_a_i32_z_f64( i32_to_f64 );
+ break;
+#ifdef EXTFLOAT80
+ case I32_TO_EXTF80:
+ ver_a_i32_z_extF80( i32_to_extF80M );
+ break;
+#endif
+#ifdef FLOAT128
+ case I32_TO_F128:
+ ver_a_i32_z_f128( i32_to_f128M );
+ break;
+#endif
+ case I64_TO_F32:
+ ver_a_i64_z_f32( i64_to_f32 );
+ break;
+ case I64_TO_F64:
+ ver_a_i64_z_f64( i64_to_f64 );
+ break;
+#ifdef EXTFLOAT80
+ case I64_TO_EXTF80:
+ ver_a_i64_z_extF80( i64_to_extF80M );
+ break;
+#endif
+#ifdef FLOAT128
+ case I64_TO_F128:
+ ver_a_i64_z_f128( i64_to_f128M );
+ break;
+#endif
+ /*--------------------------------------------------------------------
+ *--------------------------------------------------------------------*/
+ case F32_TO_UI32:
+ ver_a_f32_z_ui32_rx( f32_to_ui32, roundingMode, exact );
+ break;
+ case F32_TO_UI64:
+ ver_a_f32_z_ui64_rx( f32_to_ui64, roundingMode, exact );
+ break;
+ case F32_TO_I32:
+ ver_a_f32_z_i32_rx( f32_to_i32, roundingMode, exact );
+ break;
+ case F32_TO_I64:
+ ver_a_f32_z_i64_rx( f32_to_i64, roundingMode, exact );
+ break;
+ case F32_TO_F64:
+ ver_a_f32_z_f64( f32_to_f64 );
+ break;
+#ifdef EXTFLOAT80
+ case F32_TO_EXTF80:
+ ver_a_f32_z_extF80( f32_to_extF80M );
+ break;
+#endif
+#ifdef FLOAT128
+ case F32_TO_F128:
+ ver_a_f32_z_f128( f32_to_f128M );
+ break;
+#endif
+ case F32_ROUNDTOINT:
+ ver_az_f32_rx( f32_roundToInt, roundingMode, exact );
+ break;
+ case F32_ADD:
+ trueFunction_abz_f32 = f32_add;
+ goto ver_abz_f32;
+ case F32_SUB:
+ trueFunction_abz_f32 = f32_sub;
+ goto ver_abz_f32;
+ case F32_MUL:
+ trueFunction_abz_f32 = f32_mul;
+ goto ver_abz_f32;
+ case F32_DIV:
+ trueFunction_abz_f32 = f32_div;
+ goto ver_abz_f32;
+ case F32_REM:
+ trueFunction_abz_f32 = f32_rem;
+ ver_abz_f32:
+ ver_abz_f32( trueFunction_abz_f32 );
+ break;
+ case F32_MULADD:
+ ver_abcz_f32( f32_mulAdd );
+ break;
+ case F32_SQRT:
+ ver_az_f32( f32_sqrt );
+ break;
+ case F32_EQ:
+ trueFunction_ab_f32_z_bool = f32_eq;
+ goto ver_ab_f32_z_bool;
+ case F32_LE:
+ trueFunction_ab_f32_z_bool = f32_le;
+ goto ver_ab_f32_z_bool;
+ case F32_LT:
+ trueFunction_ab_f32_z_bool = f32_lt;
+ goto ver_ab_f32_z_bool;
+ case F32_EQ_SIGNALING:
+ trueFunction_ab_f32_z_bool = f32_eq_signaling;
+ goto ver_ab_f32_z_bool;
+ case F32_LE_QUIET:
+ trueFunction_ab_f32_z_bool = f32_le_quiet;
+ goto ver_ab_f32_z_bool;
+ case F32_LT_QUIET:
+ trueFunction_ab_f32_z_bool = f32_lt_quiet;
+ ver_ab_f32_z_bool:
+ ver_ab_f32_z_bool( trueFunction_ab_f32_z_bool );
+ break;
+ /*--------------------------------------------------------------------
+ *--------------------------------------------------------------------*/
+ case F64_TO_UI32:
+ ver_a_f64_z_ui32_rx( f64_to_ui32, roundingMode, exact );
+ break;
+ case F64_TO_UI64:
+ ver_a_f64_z_ui64_rx( f64_to_ui64, roundingMode, exact );
+ break;
+ case F64_TO_I32:
+ ver_a_f64_z_i32_rx( f64_to_i32, roundingMode, exact );
+ break;
+ case F64_TO_I64:
+ ver_a_f64_z_i64_rx( f64_to_i64, roundingMode, exact );
+ break;
+ case F64_TO_F32:
+ ver_a_f64_z_f32( f64_to_f32 );
+ break;
+#ifdef EXTFLOAT80
+ case F64_TO_EXTF80:
+ ver_a_f64_z_extF80( f64_to_extF80M );
+ break;
+#endif
+#ifdef FLOAT128
+ case F64_TO_F128:
+ ver_a_f64_z_f128( f64_to_f128M );
+ break;
+#endif
+ case F64_ROUNDTOINT:
+ ver_az_f64_rx( f64_roundToInt, roundingMode, exact );
+ break;
+ case F64_ADD:
+ trueFunction_abz_f64 = f64_add;
+ goto ver_abz_f64;
+ case F64_SUB:
+ trueFunction_abz_f64 = f64_sub;
+ goto ver_abz_f64;
+ case F64_MUL:
+ trueFunction_abz_f64 = f64_mul;
+ goto ver_abz_f64;
+ case F64_DIV:
+ trueFunction_abz_f64 = f64_div;
+ goto ver_abz_f64;
+ case F64_REM:
+ trueFunction_abz_f64 = f64_rem;
+ ver_abz_f64:
+ ver_abz_f64( trueFunction_abz_f64 );
+ break;
+ case F64_MULADD:
+ ver_abcz_f64( f64_mulAdd );
+ break;
+ case F64_SQRT:
+ ver_az_f64( f64_sqrt );
+ break;
+ case F64_EQ:
+ trueFunction_ab_f64_z_bool = f64_eq;
+ goto ver_ab_f64_z_bool;
+ case F64_LE:
+ trueFunction_ab_f64_z_bool = f64_le;
+ goto ver_ab_f64_z_bool;
+ case F64_LT:
+ trueFunction_ab_f64_z_bool = f64_lt;
+ goto ver_ab_f64_z_bool;
+ case F64_EQ_SIGNALING:
+ trueFunction_ab_f64_z_bool = f64_eq_signaling;
+ goto ver_ab_f64_z_bool;
+ case F64_LE_QUIET:
+ trueFunction_ab_f64_z_bool = f64_le_quiet;
+ goto ver_ab_f64_z_bool;
+ case F64_LT_QUIET:
+ trueFunction_ab_f64_z_bool = f64_lt_quiet;
+ ver_ab_f64_z_bool:
+ ver_ab_f64_z_bool( trueFunction_ab_f64_z_bool );
+ break;
+ /*--------------------------------------------------------------------
+ *--------------------------------------------------------------------*/
+#ifdef EXTFLOAT80
+ case EXTF80_TO_UI32:
+ ver_a_extF80_z_ui32_rx( extF80M_to_ui32, roundingMode, exact );
+ break;
+ case EXTF80_TO_UI64:
+ ver_a_extF80_z_ui64_rx( extF80M_to_ui64, roundingMode, exact );
+ break;
+ case EXTF80_TO_I32:
+ ver_a_extF80_z_i32_rx( extF80M_to_i32, roundingMode, exact );
+ break;
+ case EXTF80_TO_I64:
+ ver_a_extF80_z_i64_rx( extF80M_to_i64, roundingMode, exact );
+ break;
+ case EXTF80_TO_F32:
+ ver_a_extF80_z_f32( extF80M_to_f32 );
+ break;
+ case EXTF80_TO_F64:
+ ver_a_extF80_z_f64( extF80M_to_f64 );
+ break;
+#ifdef FLOAT128
+ case EXTF80_TO_F128:
+ ver_a_extF80_z_f128( extF80M_to_f128M );
+ break;
+#endif
+ case EXTF80_ROUNDTOINT:
+ ver_az_extF80_rx( extF80M_roundToInt, roundingMode, exact );
+ break;
+ case EXTF80_ADD:
+ trueFunction_abz_extF80 = extF80M_add;
+ goto ver_abz_extF80;
+ case EXTF80_SUB:
+ trueFunction_abz_extF80 = extF80M_sub;
+ goto ver_abz_extF80;
+ case EXTF80_MUL:
+ trueFunction_abz_extF80 = extF80M_mul;
+ goto ver_abz_extF80;
+ case EXTF80_DIV:
+ trueFunction_abz_extF80 = extF80M_div;
+ goto ver_abz_extF80;
+ case EXTF80_REM:
+ trueFunction_abz_extF80 = extF80M_rem;
+ ver_abz_extF80:
+ ver_abz_extF80( trueFunction_abz_extF80 );
+ break;
+ case EXTF80_SQRT:
+ ver_az_extF80( extF80M_sqrt );
+ break;
+ case EXTF80_EQ:
+ trueFunction_ab_extF80_z_bool = extF80M_eq;
+ goto ver_ab_extF80_z_bool;
+ case EXTF80_LE:
+ trueFunction_ab_extF80_z_bool = extF80M_le;
+ goto ver_ab_extF80_z_bool;
+ case EXTF80_LT:
+ trueFunction_ab_extF80_z_bool = extF80M_lt;
+ goto ver_ab_extF80_z_bool;
+ case EXTF80_EQ_SIGNALING:
+ trueFunction_ab_extF80_z_bool = extF80M_eq_signaling;
+ goto ver_ab_extF80_z_bool;
+ case EXTF80_LE_QUIET:
+ trueFunction_ab_extF80_z_bool = extF80M_le_quiet;
+ goto ver_ab_extF80_z_bool;
+ case EXTF80_LT_QUIET:
+ trueFunction_ab_extF80_z_bool = extF80M_lt_quiet;
+ ver_ab_extF80_z_bool:
+ ver_ab_extF80_z_bool( trueFunction_ab_extF80_z_bool );
+ break;
+#endif
+ /*--------------------------------------------------------------------
+ *--------------------------------------------------------------------*/
+#ifdef FLOAT128
+ case F128_TO_UI32:
+ ver_a_f128_z_ui32_rx( f128M_to_ui32, roundingMode, exact );
+ break;
+ case F128_TO_UI64:
+ ver_a_f128_z_ui64_rx( f128M_to_ui64, roundingMode, exact );
+ break;
+ case F128_TO_I32:
+ ver_a_f128_z_i32_rx( f128M_to_i32, roundingMode, exact );
+ break;
+ case F128_TO_I64:
+ ver_a_f128_z_i64_rx( f128M_to_i64, roundingMode, exact );
+ break;
+ case F128_TO_F32:
+ ver_a_f128_z_f32( f128M_to_f32 );
+ break;
+ case F128_TO_F64:
+ ver_a_f128_z_f64( f128M_to_f64 );
+ break;
+#ifdef EXTFLOAT80
+ case F128_TO_EXTF80:
+ ver_a_f128_z_extF80( f128M_to_extF80M );
+ break;
+#endif
+ case F128_ROUNDTOINT:
+ ver_az_f128_rx( f128M_roundToInt, roundingMode, exact );
+ break;
+ case F128_ADD:
+ trueFunction_abz_f128 = f128M_add;
+ goto ver_abz_f128;
+ case F128_SUB:
+ trueFunction_abz_f128 = f128M_sub;
+ goto ver_abz_f128;
+ case F128_MUL:
+ trueFunction_abz_f128 = f128M_mul;
+ goto ver_abz_f128;
+ case F128_DIV:
+ trueFunction_abz_f128 = f128M_div;
+ goto ver_abz_f128;
+ case F128_REM:
+ trueFunction_abz_f128 = f128M_rem;
+ ver_abz_f128:
+ ver_abz_f128( trueFunction_abz_f128 );
+ break;
+ case F128_MULADD:
+ ver_abcz_f128( f128M_mulAdd );
+ break;
+ case F128_SQRT:
+ ver_az_f128( f128M_sqrt );
+ break;
+ case F128_EQ:
+ trueFunction_ab_f128_z_bool = f128M_eq;
+ goto ver_ab_f128_z_bool;
+ case F128_LE:
+ trueFunction_ab_f128_z_bool = f128M_le;
+ goto ver_ab_f128_z_bool;
+ case F128_LT:
+ trueFunction_ab_f128_z_bool = f128M_lt;
+ goto ver_ab_f128_z_bool;
+ case F128_EQ_SIGNALING:
+ trueFunction_ab_f128_z_bool = f128M_eq_signaling;
+ goto ver_ab_f128_z_bool;
+ case F128_LE_QUIET:
+ trueFunction_ab_f128_z_bool = f128M_le_quiet;
+ goto ver_ab_f128_z_bool;
+ case F128_LT_QUIET:
+ trueFunction_ab_f128_z_bool = f128M_lt_quiet;
+ ver_ab_f128_z_bool:
+ ver_ab_f128_z_bool( trueFunction_ab_f128_z_bool );
+ break;
+#endif
+ }
+ verCases_exitWithStatus();
+ /*------------------------------------------------------------------------
+ *------------------------------------------------------------------------*/
+ optionError:
+ fail( "`%s' option requires numeric argument", *argv );
+ /*------------------------------------------------------------------------
+ *------------------------------------------------------------------------*/
+ invalidArg:
+ fail( "Invalid argument `%s'", *argv );
+
+}
+