Skip to content

Commit

Permalink
fix bugs on exception flags
Browse files Browse the repository at this point in the history
  • Loading branch information
wangkui0508 committed Aug 21, 2018
1 parent 924a150 commit 3896b56
Show file tree
Hide file tree
Showing 19 changed files with 698 additions and 113 deletions.
14 changes: 8 additions & 6 deletions SCR/tf_mac.sh
Original file line number Diff line number Diff line change
Expand Up @@ -5,10 +5,12 @@
# 4 -rnear_maxMag --Test only rounding to nearest/maximum magnitude
# (nearest/away).

#TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 1 -tininessafter -rnear_even |grep "DUMP:" > tests.dat
#cat tests.dat | RND=0 ./obj_dir/Vtb_fp_mac

TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 2 -tininessafter -rnear_even |grep "DUMP:" | RND=0 ./obj_dir/Vtb_fp_mac
TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 2 -tininessafter -rminMag |grep "DUMP:" | RND=1 ./obj_dir/Vtb_fp_mac
TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 2 -tininessafter -rmin |grep "DUMP:" | RND=2 ./obj_dir/Vtb_fp_mac
TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 2 -tininessafter -rmax |grep "DUMP:" | RND=3 ./obj_dir/Vtb_fp_mac
TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 2 -tininessafter -rnear_maxMag |grep "DUMP:" | RND=4 ./obj_dir/Vtb_fp_mac

TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 1 -tininessafter -rnear_even |grep "DUMP:" | RND=0 ./obj_dir/Vtb_fp_mac
TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 1 -tininessafter -rminMag |grep "DUMP:" | RND=1 ./obj_dir/Vtb_fp_mac
TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 1 -tininessafter -rmin |grep "DUMP:" | RND=2 ./obj_dir/Vtb_fp_mac
TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 1 -tininessafter -rmax |grep "DUMP:" | RND=3 ./obj_dir/Vtb_fp_mac
TF_DUMP=1 ../../TestFloat-3e/build/Linux-x86_64-GCC/testsoftfloat f32_mulAdd -seed 314 -level 1 -tininessafter -rnear_maxMag |grep "DUMP:" | RND=4 ./obj_dir/Vtb_fp_mac
#
159 changes: 159 additions & 0 deletions patching_testfloat/test_abcz_f32.c
Original file line number Diff line number Diff line change
@@ -0,0 +1,159 @@

/*============================================================================
This C source file is part of TestFloat, Release 3e, 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, 2017 The Regents of the University of
California. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
1. Redistributions of source code must retain the above copyright notice,
this list of conditions, and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright notice,
this list of conditions, and the following disclaimer in the documentation
and/or other materials provided with the distribution.
3. Neither the name of the University nor the names of its contributors may
be used to endorse or promote products derived from this software without
specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS", AND ANY
EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, ARE
DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
=============================================================================*/

#include <stdint.h>
#include "platform.h"
#include "softfloat.h"
#include "genCases.h"
#include "verCases.h"
#include "writeCase.h"
#include "testLoops.h"
#include <stdlib.h>
#include <fenv.h>
#include <math.h>

#pragma STDC FENV_ACCESS ON

void
test_abcz_f32(
float32_t trueFunction( float32_t, float32_t, float32_t ),
float32_t subjFunction( float32_t, float32_t, float32_t )
)
{
int count;
float32_t trueZ;
uint_fast8_t trueFlags;
float32_t subjZ;
uint_fast8_t subjFlags;

genCases_f32_abc_init();
genCases_writeTestsTotal( testLoops_forever );
verCases_errorCount = 0;
verCases_tenThousandsCount = 0;
count = 10000;
int dumpit=(getenv("TF_DUMP")!=NULL);
if(dumpit) {
int rnd=(getenv("RND")==NULL)? 0 : atoi(getenv("RND"));
if(rnd==0) fesetround(FE_TONEAREST);
if(rnd==1) fesetround(FE_TOWARDZERO);
if(rnd==2) fesetround(FE_DOWNWARD);
if(rnd==3) fesetround(FE_UPWARD);
}
while ( ! genCases_done || testLoops_forever ) {
genCases_f32_abc_next();
*testLoops_trueFlagsPtr = 0;
trueZ = trueFunction( genCases_f32_a, genCases_f32_b, genCases_f32_c );
if(dumpit) {
uint32_t *x,*y,*z,*o,f;
float *xf,*yf,*zf;
float oRef;
uint32_t oRefI, fRef;
x=(uint32_t*)(&genCases_f32_a);
y=(uint32_t*)(&genCases_f32_b);
z=(uint32_t*)(&genCases_f32_c);
o=(uint32_t*)(&trueZ);
f=*(uint32_t*)testLoops_trueFlagsPtr;
printf("DUMP: %08x %08x %08x %08x %08x\n", *x, *y, *z, *o, f);

//const uint32_t EXP_FIELD=0xFF<<23;
//const uint32_t MANT_FIELD=(1<<23)-1;
//feclearexcept(FE_ALL_EXCEPT);
//xf=(float*)(&genCases_f32_a);
//yf=(float*)(&genCases_f32_b);
//zf=(float*)(&genCases_f32_c);
//oRef=fmaf(*xf, *yf, *zf);
//oRefI=*((uint32_t*)(&oRef));
//fRef=0;
//int resIsNaN=(*o&MANT_FIELD)!=0 && (*o&EXP_FIELD)==EXP_FIELD;
//int refIsNaN=(oRefI&MANT_FIELD)!=0 && (oRefI&EXP_FIELD)==EXP_FIELD;
//if(fetestexcept(FE_INEXACT)) fRef|=softfloat_flag_inexact;
//if(fetestexcept(FE_INVALID)) fRef|=softfloat_flag_invalid;
//if(fetestexcept(FE_OVERFLOW)) fRef|=softfloat_flag_overflow;
//if(fetestexcept(FE_UNDERFLOW))fRef|=softfloat_flag_underflow;
//if(refIsNaN) fRef=(fRef&~softfloat_flag_invalid);//do not care invalid
//if(resIsNaN) f=( f&~softfloat_flag_invalid);//do not care invalid
//f=f&~softfloat_flag_infinite;
//int pass;
//pass=(oRefI==*o || (resIsNaN&&refIsNaN));
//if(fRef!=f) pass=0;
//if(!pass) {
// printf("ERR: %08x %08x %08x %08x %08x | %08x %08x %d %d\n",
// *x, *y, *z, *o, f, oRefI, fRef, refIsNaN, resIsNaN);
//}
//else {
// //printf("DUMP: %08x %08x %08x %08x %08x\n", *x, *y, *z, *o, f);
//}
}
trueFlags = *testLoops_trueFlagsPtr;
testLoops_subjFlagsFunction();
subjZ = subjFunction( genCases_f32_a, genCases_f32_b, genCases_f32_c );
subjFlags = testLoops_subjFlagsFunction();
--count;
if ( ! count ) {
verCases_perTenThousand();
count = 10000;
}
if ( ! f32_same( trueZ, subjZ ) || (trueFlags != subjFlags) ) {
if (
! verCases_checkNaNs
&& ( f32_isSignalingNaN( genCases_f32_a )
|| f32_isSignalingNaN( genCases_f32_b )
|| f32_isSignalingNaN( genCases_f32_c )
)
) {
trueFlags |= softfloat_flag_invalid;
}
if (
verCases_checkNaNs
|| ! f32_isNaN( trueZ )
|| ! f32_isNaN( subjZ )
|| f32_isSignalingNaN( subjZ )
|| (trueFlags != subjFlags)
) {
++verCases_errorCount;
verCases_writeErrorFound( 10000 - count );
writeCase_abc_f32(
genCases_f32_a, genCases_f32_b, genCases_f32_c );
writeCase_z_f32( trueZ, trueFlags, subjZ, subjFlags );
if ( verCases_errorCount == verCases_maxErrorCount ) break;
}
}
}
verCases_writeTestsPerformed( 10000 - count );

}

163 changes: 163 additions & 0 deletions patching_testfloat/test_abz_f32.c
Original file line number Diff line number Diff line change
@@ -0,0 +1,163 @@

/*============================================================================
This C source file is part of TestFloat, Release 3e, 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, 2017 The Regents of the University of
California. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
1. Redistributions of source code must retain the above copyright notice,
this list of conditions, and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright notice,
this list of conditions, and the following disclaimer in the documentation
and/or other materials provided with the distribution.
3. Neither the name of the University nor the names of its contributors may
be used to endorse or promote products derived from this software without
specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS", AND ANY
EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, ARE
DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
=============================================================================*/

#include <stdint.h>
#include "platform.h"
#include "softfloat.h"
#include "genCases.h"
#include "verCases.h"
#include "writeCase.h"
#include "testLoops.h"
#include <stdlib.h>
#include <fenv.h>

#pragma STDC FENV_ACCESS ON

void
test_abz_f32(
float32_t trueFunction( float32_t, float32_t ),
float32_t subjFunction( float32_t, float32_t )
)
{
int count;
float32_t trueZ;
uint_fast8_t trueFlags;
float32_t subjZ;
uint_fast8_t subjFlags;

genCases_f32_ab_init();
genCases_writeTestsTotal( testLoops_forever );
verCases_errorCount = 0;
verCases_tenThousandsCount = 0;
count = 10000;
int dumpit=(getenv("TF_DUMP")!=NULL);
int OP;
if(dumpit) {
int rnd=(getenv("RND")==NULL)? 0 : atoi(getenv("RND"));
OP=(getenv("OP")==NULL)? 0 : atoi(getenv("OP"));
if(rnd==0) fesetround(FE_TONEAREST);
if(rnd==1) fesetround(FE_TOWARDZERO);
if(rnd==2) fesetround(FE_DOWNWARD);
if(rnd==3) fesetround(FE_UPWARD);
}
while ( ! genCases_done || testLoops_forever ) {
genCases_f32_ab_next();
*testLoops_trueFlagsPtr = 0;
trueZ = trueFunction( genCases_f32_a, genCases_f32_b );
if(dumpit) {
uint32_t *x,*y,*o,f;
float *xf,*yf,*zf;
float oRef;
uint32_t oRefI, fRef;
x=(uint32_t*)(&genCases_f32_a);
y=(uint32_t*)(&genCases_f32_b);
o=(uint32_t*)(&trueZ);
f=*(uint32_t*)testLoops_trueFlagsPtr;
printf("DUMP: %08x %08x %08x %08x\n", *x, *y, *o,f);

//const uint32_t EXP_FIELD=0xFF<<23;
//const uint32_t MANT_FIELD=(1<<23)-1;
//feclearexcept(FE_ALL_EXCEPT);
//xf=(float*)(&genCases_f32_a);
//yf=(float*)(&genCases_f32_b);
//if(OP==0) {
// oRef=(*xf)+(*yf);
//}
//else if(OP==1) {
// oRef=(*xf)*(*yf);
//}
//else if(OP==2) {
// oRef=(*xf)/(*yf);
//}
//oRefI=*((uint32_t*)(&oRef));
//fRef=0;
//int resIsNaN=(*o&MANT_FIELD)!=0 && (*o&EXP_FIELD)==EXP_FIELD;
//int refIsNaN=(oRefI&MANT_FIELD)!=0 && (oRefI&EXP_FIELD)==EXP_FIELD;
//if(fetestexcept(FE_INEXACT)) fRef|=softfloat_flag_inexact;
//if(fetestexcept(FE_INVALID)) fRef|=softfloat_flag_invalid;
//if(fetestexcept(FE_OVERFLOW)) fRef|=softfloat_flag_overflow;
//if(fetestexcept(FE_UNDERFLOW))fRef|=softfloat_flag_underflow;
//if(refIsNaN) fRef=(fRef&~softfloat_flag_invalid);//do not care invalid
//if(resIsNaN) f=( f&~softfloat_flag_invalid);//do not care invalid
//f=f&~softfloat_flag_infinite;
//int pass;
//pass=(oRefI==*o || (resIsNaN&&refIsNaN));
//if(fRef!=f) pass=0;
//if(!pass) {
// printf("ERR: %08x %08x %08x %08x | %08x %08x %d %d %s\n",
// *x, *y, *o, f, oRefI, fRef, refIsNaN, resIsNaN, getenv("OP"));
//}
//else {
// //printf("DUMP: %08x %08x %08x %08x\n", *x, *y, *o, f);
//}
}
trueFlags = *testLoops_trueFlagsPtr;
testLoops_subjFlagsFunction();
subjZ = subjFunction( genCases_f32_a, genCases_f32_b );
subjFlags = testLoops_subjFlagsFunction();
--count;
if ( ! count ) {
verCases_perTenThousand();
count = 10000;
}
if ( ! f32_same( trueZ, subjZ ) || (trueFlags != subjFlags) ) {
if (
! verCases_checkNaNs
&& (f32_isSignalingNaN( genCases_f32_a )
|| f32_isSignalingNaN( genCases_f32_b ))
) {
trueFlags |= softfloat_flag_invalid;
}
if (
verCases_checkNaNs
|| ! f32_isNaN( trueZ )
|| ! f32_isNaN( subjZ )
|| f32_isSignalingNaN( subjZ )
|| (trueFlags != subjFlags)
) {
++verCases_errorCount;
verCases_writeErrorFound( 10000 - count );
writeCase_ab_f32( genCases_f32_a, genCases_f32_b );
writeCase_z_f32( trueZ, trueFlags, subjZ, subjFlags );
if ( verCases_errorCount == verCases_maxErrorCount ) break;
}
}
}
verCases_writeTestsPerformed( 10000 - count );

}

Loading

0 comments on commit 3896b56

Please sign in to comment.