Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Credit goes to Steven Munroe <munroesj@linux.vnet.ibm.com>
for the float{uns,}titd routines. ti -> {dd,sd} is still
a work in progress.
These still leaves plenty of room for optimization, but these
are a functional first pass.
This does a bit of build system shuffling as TImode is not
supported on all platforms, and thus a separate ABI list
needs to be kept for these cases.
Signed-off-by: Paul E. Murphy <murphyp@linux.vnet.ibm.com>
Signed-off-by: Steven Munroe <munroesj@linux.vnet.ibm.com>- Loading branch information
Paul E. Murphy
committed
Dec 16, 2015
1 parent
d86a852
commit a45d232
Showing
31 changed files
with
2,447 additions
and
7 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -9,5 +9,6 @@ libdfp { | |
| LIBDFP_1.0.8 | ||
| LIBDFP_1.0.9 | ||
| LIBDFP_1.0.10 | ||
| LIBDFP_1.0.13 | ||
| LIBDFP_PRIVATE | ||
| } | ||
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,31 @@ | ||
| /* fixddti implementation. | ||
| Copyright (C) 2015 Free Software Foundation, Inc. | ||
| This file is part of the Decimal Floating Point C Library. | ||
| Author(s): Paul E. Murphy <murphyp@linux.vnet.ibm.com> | ||
| The Decimal Floating Point C Library is free software; you can | ||
| redistribute it and/or modify it under the terms of the GNU Lesser | ||
| General Public License version 2.1. | ||
| The Decimal Floating Point C Library is distributed in the hope that | ||
| it will be useful, but WITHOUT ANY WARRANTY; without even the implied | ||
| warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See | ||
| the GNU Lesser General Public License version 2.1 for more details. | ||
| You should have received a copy of the GNU Lesser General Public | ||
| License version 2.1 along with the Decimal Floating Point C Library; | ||
| if not, write to the Free Software Foundation, Inc., 59 Temple Place, | ||
| Suite 330, Boston, MA 02111-1307 USA. | ||
| Please see libdfp/COPYING.txt for more information. */ | ||
|
|
||
| #define SRC_SIZE 64 | ||
| #define FUNC fixddti | ||
| #define SHIFT_EXP 16 | ||
| #define DSUFF DD | ||
| #define SIGNED 1 | ||
|
|
||
| #include "fixsdti.c" |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,150 @@ | ||
| /* Common fix*ti and fixsdti implementation. | ||
| Copyright (C) 2015 Free Software Foundation, Inc. | ||
| This file is part of the Decimal Floating Point C Library. | ||
| Author(s): Paul E. Murphy <murphyp@linux.vnet.ibm.com> | ||
| The Decimal Floating Point C Library is free software; you can | ||
| redistribute it and/or modify it under the terms of the GNU Lesser | ||
| General Public License version 2.1. | ||
| The Decimal Floating Point C Library is distributed in the hope that | ||
| it will be useful, but WITHOUT ANY WARRANTY; without even the implied | ||
| warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See | ||
| the GNU Lesser General Public License version 2.1 for more details. | ||
| You should have received a copy of the GNU Lesser General Public | ||
| License version 2.1 along with the Decimal Floating Point C Library; | ||
| if not, write to the Free Software Foundation, Inc., 59 Temple Place, | ||
| Suite 330, Boston, MA 02111-1307 USA. | ||
| Please see libdfp/COPYING.txt for more information. */ | ||
|
|
||
| #ifndef SRC_SIZE | ||
| # define SRC_SIZE 32 | ||
| # define FUNC fixsdti | ||
| # define SHIFT_EXP 7 | ||
| # define DSUFF DF | ||
| # define SIGNED 1 | ||
| #endif | ||
|
|
||
| #define _DECIMAL_SIZE SRC_SIZE | ||
|
|
||
| #include <dfpacc.h> | ||
| #include <stdint.h> | ||
| #include <dfp/math.h> | ||
| #include <dfpmacro.h> | ||
| #include <ieee754r_private.h> | ||
|
|
||
| #if SIGNED == 1 | ||
| # if __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__ | ||
| # define MAX_RET_VAL {.di = {0x7FFFFFFFFFFFFFFFLL, 0xFFFFFFFFFFFFFFFFLL}} | ||
| # define MIN_RET_VAL {.di = {0x8000000000000000LL, 0LL}} | ||
| # else | ||
| # define MAX_RET_VAL {.di = {0xFFFFFFFFFFFFFFFFLL, 0x7FFFFFFFFFFFFFFFLL}} | ||
| # define MIN_RET_VAL {.di = { 0LL, 0x8000000000000000LL}} | ||
| # endif | ||
| # if SRC_SIZE == 32 | ||
| # define MIN_VAL (-170141100000000000000000000000000000000.DF) | ||
| # define MAX_VAL ( 170141100000000000000000000000000000000.DF) | ||
| # elif SRC_SIZE == 64 | ||
| # define MIN_VAL (-170141183460469200000000000000000000000.DD) | ||
| # define MAX_VAL ( 170141183460469200000000000000000000000.DD) | ||
| # else | ||
| # define MIN_VAL (-170141183460469231731687303715884100000.DL) | ||
| # define MAX_VAL ( 170141183460469231731687303715884100000.DL) | ||
| # endif | ||
| # define RET_TYPE __int128 | ||
| # define QUICK_CONVERT(_x) ({ long long int _a_di = _x; _a_di;}) | ||
| #else | ||
| # define MAX_RET_VAL {.di = {0xFFFFFFFFFFFFFFFFLL, 0xFFFFFFFFFFFFFFFFLL}} | ||
| # define MIN_RET_VAL {.di = {0LL, 0LL}} | ||
| # define MIN_VAL PASTE(0., DSUFF) | ||
| # if SRC_SIZE == 32 | ||
| # define MAX_VAL (340282300000000000000000000000000000000.DF) | ||
| # elif SRC_SIZE == 64 | ||
| # define MAX_VAL (340282366920938400000000000000000000000.DD) | ||
| # else | ||
| # define MAX_VAL (340282366920938463463374607431768200000.DL) | ||
| # endif | ||
| # define RET_TYPE unsigned __int128 | ||
| # define QUICK_CONVERT(_x) ({ unsigned long long int _a_di = _x; _a_di;}) | ||
| #endif | ||
|
|
||
| #define TRUNC(_x) (PASTE(__truncd, SRC_SIZE)(_x)) | ||
| #define FREXP(_x,_y) (PASTE(__frexpd,SRC_SIZE)(_x,_y)) | ||
| #define CLASSIFY(_x) (PASTE(__fpclassifyd,SRC_SIZE)(_x)) | ||
| #define SHIFT_EXP_VAL PASTE(PASTE(1e,SHIFT_EXP),PASTE(DSUFF,)) | ||
| #define SIGNBIT(_x) (PASTE(__signbitd,SRC_SIZE)(_x)) | ||
|
|
||
| #if SIGNED == 1 | ||
| # define FABS(_x) (PASTE(__fabsd,SRC_SIZE)(_x)) | ||
| #else | ||
| # define FABS(_x) _x | ||
| #endif | ||
|
|
||
| extern const __int128 exp10_ti[39]; | ||
|
|
||
| RET_TYPE __BACKEND_(FUNC) (DEC_TYPE a) | ||
| { | ||
| union ret_type | ||
| { | ||
| RET_TYPE ti; | ||
| uint64_t di[2]; | ||
| }; | ||
| union ret_type min = MIN_RET_VAL; | ||
| union ret_type max = MAX_RET_VAL; | ||
|
|
||
| switch (CLASSIFY (a)) | ||
| { | ||
| case FP_ZERO: | ||
| return 0; | ||
|
|
||
| case FP_INFINITE: | ||
| DFP_EXCEPT (FE_INVALID); | ||
| return SIGNBIT(a) ? min.ti : max.ti; | ||
|
|
||
| case FP_NAN: | ||
| DFP_EXCEPT (FE_INVALID); | ||
| return min.ti; | ||
| } | ||
|
|
||
| if(a < MIN_VAL || a > MAX_VAL) | ||
| { | ||
| DFP_EXCEPT (FE_INVALID); | ||
| return (a < MIN_VAL) ? min.ti : max.ti; | ||
| } | ||
|
|
||
| /* A 64 bit mantissa is sufficient for dd/sd conversions. */ | ||
| #if SRC_SIZE <= 64 | ||
| RET_TYPE result; | ||
| int e; | ||
|
|
||
| if (FABS(a) < SHIFT_EXP_VAL) | ||
| { | ||
| return (RET_TYPE) QUICK_CONVERT (a); | ||
| } | ||
|
|
||
| DEC_TYPE a_int = TRUNC (a); | ||
| DEC_TYPE a_norm = FREXP (a_int, &e); | ||
|
|
||
|
|
||
| a_norm *= SHIFT_EXP_VAL; | ||
| e -= SHIFT_EXP; | ||
| /* Force a 64b conversion of the mantissa */ | ||
| result = ((int64_t) a_norm); | ||
| if (e > 0) | ||
| return result * (RET_TYPE) exp10_ti[e]; | ||
| else if (e < 0) | ||
| return result / (RET_TYPE) exp10_ti[-e]; | ||
| else | ||
| return result; | ||
| #else | ||
| /* This requires specific knowledge of the encoding. Punt it. */ | ||
| return D128_CONVERSION(a); | ||
| #endif | ||
| } | ||
|
|
||
| hidden_def (__BACKEND_(FUNC)) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,31 @@ | ||
| /* fixunsddti implementation. | ||
| Copyright (C) 2015 Free Software Foundation, Inc. | ||
| This file is part of the Decimal Floating Point C Library. | ||
| Author(s): Paul E. Murphy <murphyp@linux.vnet.ibm.com> | ||
| The Decimal Floating Point C Library is free software; you can | ||
| redistribute it and/or modify it under the terms of the GNU Lesser | ||
| General Public License version 2.1. | ||
| The Decimal Floating Point C Library is distributed in the hope that | ||
| it will be useful, but WITHOUT ANY WARRANTY; without even the implied | ||
| warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See | ||
| the GNU Lesser General Public License version 2.1 for more details. | ||
| You should have received a copy of the GNU Lesser General Public | ||
| License version 2.1 along with the Decimal Floating Point C Library; | ||
| if not, write to the Free Software Foundation, Inc., 59 Temple Place, | ||
| Suite 330, Boston, MA 02111-1307 USA. | ||
| Please see libdfp/COPYING.txt for more information. */ | ||
|
|
||
| #define SRC_SIZE 64 | ||
| #define FUNC fixunsddti | ||
| #define SHIFT_EXP 16 | ||
| #define DSUFF DD | ||
| #define SIGNED 0 | ||
|
|
||
| #include "fixsdti.c" |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,31 @@ | ||
| /* fixunssdti implementation. | ||
| Copyright (C) 2015 Free Software Foundation, Inc. | ||
| This file is part of the Decimal Floating Point C Library. | ||
| Author(s): Paul E. Murphy <murphyp@linux.vnet.ibm.com> | ||
| The Decimal Floating Point C Library is free software; you can | ||
| redistribute it and/or modify it under the terms of the GNU Lesser | ||
| General Public License version 2.1. | ||
| The Decimal Floating Point C Library is distributed in the hope that | ||
| it will be useful, but WITHOUT ANY WARRANTY; without even the implied | ||
| warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See | ||
| the GNU Lesser General Public License version 2.1 for more details. | ||
| You should have received a copy of the GNU Lesser General Public | ||
| License version 2.1 along with the Decimal Floating Point C Library; | ||
| if not, write to the Free Software Foundation, Inc., 59 Temple Place, | ||
| Suite 330, Boston, MA 02111-1307 USA. | ||
| Please see libdfp/COPYING.txt for more information. */ | ||
|
|
||
| #define SRC_SIZE 32 | ||
| #define FUNC fixunssdti | ||
| #define SHIFT_EXP 7 | ||
| #define DSUFF DF | ||
| #define SIGNED 0 | ||
|
|
||
| #include "fixsdti.c" |
Oops, something went wrong.