OpenCores
URL https://opencores.org/ocsvn/openrisc/openrisc/trunk

Subversion Repositories openrisc

[/] [openrisc/] [trunk/] [gnu-dev/] [or1k-gcc/] [libgcc/] [config/] [libbid/] [bid64_string.c] - Rev 801

Go to most recent revision | Compare with Previous | Blame | View Log

/* Copyright (C) 2007, 2009  Free Software Foundation, Inc.
 
This file is part of GCC.
 
GCC is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation; either version 3, or (at your option) any later
version.
 
GCC 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 General Public License
for more details.
 
Under Section 7 of GPL version 3, you are granted additional
permissions described in the GCC Runtime Library Exception, version
3.1, as published by the Free Software Foundation.
 
You should have received a copy of the GNU General Public License and
a copy of the GCC Runtime Library Exception along with this program;
see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
<http://www.gnu.org/licenses/>.  */
 
#include <ctype.h>
#include "bid_internal.h"
#include "bid128_2_str.h"
#include "bid128_2_str_macros.h"
 
#define MAX_FORMAT_DIGITS     16
#define DECIMAL_EXPONENT_BIAS 398
#define MAX_DECIMAL_EXPONENT  767
 
#if DECIMAL_CALL_BY_REFERENCE
 
void
bid64_to_string (char *ps, UINT64 * px
		 _EXC_FLAGS_PARAM _EXC_MASKS_PARAM _EXC_INFO_PARAM) {
  UINT64 x;
#else
 
void
bid64_to_string (char *ps, UINT64 x
		 _EXC_FLAGS_PARAM _EXC_MASKS_PARAM _EXC_INFO_PARAM) {
#endif
// the destination string (pointed to by ps) must be pre-allocated
  UINT64 sign_x, coefficient_x, D, ER10;
  int istart, exponent_x, j, digits_x, bin_expon_cx;
  int_float tempx;
  UINT32 MiDi[12], *ptr;
  UINT64 HI_18Dig, LO_18Dig, Tmp;
  char *c_ptr_start, *c_ptr;
  int midi_ind, k_lcv, len;
  unsigned int save_fpsf;
 
#if DECIMAL_CALL_BY_REFERENCE
  x = *px;
#endif
 
  save_fpsf = *pfpsf; // place holder only
  // unpack arguments, check for NaN or Infinity
  if (!unpack_BID64 (&sign_x, &exponent_x, &coefficient_x, x)) {
    // x is Inf. or NaN or 0
 
    // Inf or NaN?
    if ((x & 0x7800000000000000ull) == 0x7800000000000000ull) {
      if ((x & 0x7c00000000000000ull) == 0x7c00000000000000ull) {
    ps[0] = (sign_x) ? '-' : '+';
    ps[1] = ((x & MASK_SNAN) == MASK_SNAN)? 'S':'Q';
	ps[2] = 'N';
	ps[3] = 'a';
	ps[4] = 'N';
	ps[5] = 0;
	return;
      }
      // x is Inf
      ps[0] = (sign_x) ? '-' : '+';
      ps[1] = 'I';
      ps[2] = 'n';
      ps[3] = 'f';
      ps[4] = 0;
      return;
    }
    // 0
    istart = 0;
    if (sign_x) {
      ps[istart++] = '-';
    }
 
    ps[istart++] = '0';
    ps[istart++] = 'E';
 
    exponent_x -= 398;
    if (exponent_x < 0) {
      ps[istart++] = '-';
      exponent_x = -exponent_x;
    } else
      ps[istart++] = '+';
 
    if (exponent_x) {
      // get decimal digits in coefficient_x
      tempx.d = (float) exponent_x;
      bin_expon_cx = ((tempx.i >> 23) & 0xff) - 0x7f;
      digits_x = estimate_decimal_digits[bin_expon_cx];
      if ((UINT64)exponent_x >= power10_table_128[digits_x].w[0])
	digits_x++;
 
      j = istart + digits_x - 1;
      istart = j + 1;
 
      // 2^32/10
      ER10 = 0x1999999a;
 
      while (exponent_x > 9) {
	D = (UINT64) exponent_x *ER10;
	D >>= 32;
	exponent_x = exponent_x - (D << 1) - (D << 3);
 
	ps[j--] = '0' + (char) exponent_x;
	exponent_x = D;
      }
      ps[j] = '0' + (char) exponent_x;
    } else {
      ps[istart++] = '0';
    }
 
    ps[istart] = 0;
 
    return;
  }
  // convert expon, coeff to ASCII
  exponent_x -= DECIMAL_EXPONENT_BIAS;
 
  ER10 = 0x1999999a;
 
  istart = 0;
  if (sign_x) {
    ps[0] = '-';
    istart = 1;
  }
  // if zero or non-canonical, set coefficient to '0'
  if ((coefficient_x > 9999999999999999ull) ||	// non-canonical
      ((coefficient_x == 0))	// significand is zero
    ) {
    ps[istart++] = '0';
  } else {
    /* ****************************************************
       This takes a bid coefficient in C1.w[1],C1.w[0] 
       and put the converted character sequence at location 
       starting at &(str[k]). The function returns the number
       of MiDi returned. Note that the character sequence 
       does not have leading zeros EXCEPT when the input is of
       zero value. It will then output 1 character '0'
       The algorithm essentailly tries first to get a sequence of
       Millenial Digits "MiDi" and then uses table lookup to get the
       character strings of these MiDis.
       **************************************************** */
    /* Algorithm first decompose possibly 34 digits in hi and lo
       18 digits. (The high can have at most 16 digits). It then
       uses macro that handle 18 digit portions.
       The first step is to get hi and lo such that
       2^(64) C1.w[1] + C1.w[0] = hi * 10^18  + lo,   0 <= lo < 10^18.
       We use a table lookup method to obtain the hi and lo 18 digits.
       [C1.w[1],C1.w[0]] = c_8 2^(107) + c_7 2^(101) + ... + c_0 2^(59) + d
       where 0 <= d < 2^59 and each c_j has 6 bits. Because d fits in
       18 digits,  we set hi = 0, and lo = d to begin with.
       We then retrieve from a table, for j = 0, 1, ..., 8
       that gives us A and B where c_j 2^(59+6j) = A * 10^18 + B.
       hi += A ; lo += B; After each accumulation into lo, we normalize 
       immediately. So at the end, we have the decomposition as we need. */
 
    Tmp = coefficient_x >> 59;
    LO_18Dig = (coefficient_x << 5) >> 5;
    HI_18Dig = 0;
    k_lcv = 0;
 
    while (Tmp) {
      midi_ind = (int) (Tmp & 0x000000000000003FLL);
      midi_ind <<= 1;
      Tmp >>= 6;
      HI_18Dig += mod10_18_tbl[k_lcv][midi_ind++];
      LO_18Dig += mod10_18_tbl[k_lcv++][midi_ind];
      __L0_Normalize_10to18 (HI_18Dig, LO_18Dig);
    }
 
    ptr = MiDi;
    __L1_Split_MiDi_6_Lead (LO_18Dig, ptr);
    len = ptr - MiDi;
    c_ptr_start = &(ps[istart]);
    c_ptr = c_ptr_start;
 
    /* now convert the MiDi into character strings */
    __L0_MiDi2Str_Lead (MiDi[0], c_ptr);
    for (k_lcv = 1; k_lcv < len; k_lcv++) {
      __L0_MiDi2Str (MiDi[k_lcv], c_ptr);
    }
    istart = istart + (c_ptr - c_ptr_start);
  }
 
  ps[istart++] = 'E';
 
  if (exponent_x < 0) {
    ps[istart++] = '-';
    exponent_x = -exponent_x;
  } else
    ps[istart++] = '+';
 
  if (exponent_x) {
    // get decimal digits in coefficient_x
    tempx.d = (float) exponent_x;
    bin_expon_cx = ((tempx.i >> 23) & 0xff) - 0x7f;
    digits_x = estimate_decimal_digits[bin_expon_cx];
    if ((UINT64)exponent_x >= power10_table_128[digits_x].w[0])
      digits_x++;
 
    j = istart + digits_x - 1;
    istart = j + 1;
 
    // 2^32/10
    ER10 = 0x1999999a;
 
    while (exponent_x > 9) {
      D = (UINT64) exponent_x *ER10;
      D >>= 32;
      exponent_x = exponent_x - (D << 1) - (D << 3);
 
      ps[j--] = '0' + (char) exponent_x;
      exponent_x = D;
    }
    ps[j] = '0' + (char) exponent_x;
  } else {
    ps[istart++] = '0';
  }
 
  ps[istart] = 0;
 
  return;
 
}
 
 
#if DECIMAL_CALL_BY_REFERENCE
void
bid64_from_string (UINT64 * pres, char *ps
		   _RND_MODE_PARAM _EXC_FLAGS_PARAM 
                   _EXC_MASKS_PARAM _EXC_INFO_PARAM) {
#else
UINT64
bid64_from_string (char *ps
		   _RND_MODE_PARAM _EXC_FLAGS_PARAM 
                   _EXC_MASKS_PARAM _EXC_INFO_PARAM) {
#endif
  UINT64 sign_x, coefficient_x = 0, rounded = 0, res;
  int expon_x = 0, sgn_expon, ndigits, add_expon = 0, midpoint =
    0, rounded_up = 0;
  int dec_expon_scale = 0, right_radix_leading_zeros = 0, rdx_pt_enc =
    0;
  unsigned fpsc;
  char c;
  unsigned int save_fpsf;
 
#if DECIMAL_CALL_BY_REFERENCE
#if !DECIMAL_GLOBAL_ROUNDING
  _IDEC_round rnd_mode = *prnd_mode;
#endif
#endif
 
  save_fpsf = *pfpsf; // place holder only
  // eliminate leading whitespace
  while (((*ps == ' ') || (*ps == '\t')) && (*ps))
    ps++;
 
  // get first non-whitespace character
  c = *ps;
 
  // detect special cases (INF or NaN)
  if (!c || (c != '.' && c != '-' && c != '+' && (c < '0' || c > '9'))) {
    // Infinity?
    if ((tolower_macro (ps[0]) == 'i' && tolower_macro (ps[1]) == 'n' && 
        tolower_macro (ps[2]) == 'f') && (!ps[3] || 
        (tolower_macro (ps[3]) == 'i' && 
        tolower_macro (ps[4]) == 'n' && tolower_macro (ps[5]) == 'i' && 
        tolower_macro (ps[6]) == 't' && tolower_macro (ps[7]) == 'y' && 
        !ps[8]))) {
      res = 0x7800000000000000ull;
      BID_RETURN (res);
    }
    // return sNaN
    if (tolower_macro (ps[0]) == 's' && tolower_macro (ps[1]) == 'n' && 
        tolower_macro (ps[2]) == 'a' && tolower_macro (ps[3]) == 'n') { 
        // case insensitive check for snan
      res = 0x7e00000000000000ull;
      BID_RETURN (res);
    } else {
      // return qNaN
      res = 0x7c00000000000000ull;
      BID_RETURN (res);
    }
  }
  // detect +INF or -INF
  if ((tolower_macro (ps[1]) == 'i' && tolower_macro (ps[2]) == 'n' && 
      tolower_macro (ps[3]) == 'f') && (!ps[4] || 
      (tolower_macro (ps[4]) == 'i' && tolower_macro (ps[5]) == 'n' && 
      tolower_macro (ps[6]) == 'i' && tolower_macro (ps[7]) == 't' && 
      tolower_macro (ps[8]) == 'y' && !ps[9]))) {
    if (c == '+')
      res = 0x7800000000000000ull;
    else if (c == '-')
      res = 0xf800000000000000ull;
    else
      res = 0x7c00000000000000ull;
    BID_RETURN (res);
  }
  // if +sNaN, +SNaN, -sNaN, or -SNaN
  if (tolower_macro (ps[1]) == 's' && tolower_macro (ps[2]) == 'n'
      && tolower_macro (ps[3]) == 'a' && tolower_macro (ps[4]) == 'n') {
    if (c == '-')
      res = 0xfe00000000000000ull;
    else
      res = 0x7e00000000000000ull;
    BID_RETURN (res);
  }
  // determine sign
  if (c == '-')
    sign_x = 0x8000000000000000ull;
  else
    sign_x = 0;
 
  // get next character if leading +/- sign
  if (c == '-' || c == '+') {
    ps++;
    c = *ps;
  }
  // if c isn't a decimal point or a decimal digit, return NaN
  if (c != '.' && (c < '0' || c > '9')) {
    // return NaN
    res = 0x7c00000000000000ull | sign_x;
    BID_RETURN (res);
  }
 
  rdx_pt_enc = 0;
 
  // detect zero (and eliminate/ignore leading zeros)
  if (*(ps) == '0' || *(ps) == '.') {
 
    if (*(ps) == '.') {
      rdx_pt_enc = 1;
      ps++;
    }
    // if all numbers are zeros (with possibly 1 radix point, the number is zero
    // should catch cases such as: 000.0
    while (*ps == '0') {
      ps++;
      // for numbers such as 0.0000000000000000000000000000000000001001, 
      // we want to count the leading zeros
      if (rdx_pt_enc) {
	right_radix_leading_zeros++;
      }
      // if this character is a radix point, make sure we haven't already 
      // encountered one
      if (*(ps) == '.') {
	if (rdx_pt_enc == 0) {
	  rdx_pt_enc = 1;
	  // if this is the first radix point, and the next character is NULL, 
          // we have a zero
	  if (!*(ps + 1)) {
	    res =
	      ((UINT64) (398 - right_radix_leading_zeros) << 53) |
	      sign_x;
	    BID_RETURN (res);
	  }
	  ps = ps + 1;
	} else {
	  // if 2 radix points, return NaN
	  res = 0x7c00000000000000ull | sign_x;
	  BID_RETURN (res);
	}
      } else if (!*(ps)) {
	//pres->w[1] = 0x3040000000000000ull | sign_x;
	res =
	  ((UINT64) (398 - right_radix_leading_zeros) << 53) | sign_x;
	BID_RETURN (res);
      }
    }
  }
 
  c = *ps;
 
  ndigits = 0;
  while ((c >= '0' && c <= '9') || c == '.') {
    if (c == '.') {
      if (rdx_pt_enc) {
	// return NaN
	res = 0x7c00000000000000ull | sign_x;
	BID_RETURN (res);
      }
      rdx_pt_enc = 1;
      ps++;
      c = *ps;
      continue;
    }
    dec_expon_scale += rdx_pt_enc;
 
    ndigits++;
    if (ndigits <= 16) {
      coefficient_x = (coefficient_x << 1) + (coefficient_x << 3);
      coefficient_x += (UINT64) (c - '0');
    } else if (ndigits == 17) {
      // coefficient rounding
		switch(rnd_mode){
	case ROUNDING_TO_NEAREST:
      midpoint = (c == '5' && !(coefficient_x & 1)) ? 1 : 0; 
          // if coefficient is even and c is 5, prepare to round up if 
          // subsequent digit is nonzero
      // if str[MAXDIG+1] > 5, we MUST round up
      // if str[MAXDIG+1] == 5 and coefficient is ODD, ROUND UP!
      if (c > '5' || (c == '5' && (coefficient_x & 1))) {
	coefficient_x++;
	rounded_up = 1;
	break;
 
	case ROUNDING_DOWN:
		if(sign_x) { coefficient_x++; rounded_up=1; }
		break;
	case ROUNDING_UP:
		if(!sign_x) { coefficient_x++; rounded_up=1; }
		break;
	case ROUNDING_TIES_AWAY:
		if(c>='5') { coefficient_x++; rounded_up=1; }
		break;
	  }
	if (coefficient_x == 10000000000000000ull) {
	  coefficient_x = 1000000000000000ull;
	  add_expon = 1;
	}
      }
      if (c > '0')
	rounded = 1;
      add_expon += 1;
    } else { // ndigits > 17
      add_expon++;
      if (midpoint && c > '0') {
	coefficient_x++;
	midpoint = 0;
	rounded_up = 1;
      }
      if (c > '0')
	rounded = 1;
    }
    ps++;
    c = *ps;
  }
 
  add_expon -= (dec_expon_scale + right_radix_leading_zeros);
 
  if (!c) {
    res =
      fast_get_BID64_check_OF (sign_x,
			       add_expon + DECIMAL_EXPONENT_BIAS,
			       coefficient_x, 0, &fpsc);
    BID_RETURN (res);
  }
 
  if (c != 'E' && c != 'e') {
    // return NaN
    res = 0x7c00000000000000ull | sign_x;
    BID_RETURN (res);
  }
  ps++;
  c = *ps;
  sgn_expon = (c == '-') ? 1 : 0;
  if (c == '-' || c == '+') {
    ps++;
    c = *ps;
  }
  if (!c || c < '0' || c > '9') {
    // return NaN
    res = 0x7c00000000000000ull | sign_x;
    BID_RETURN (res);
  }
 
  while (c >= '0' && c <= '9') {
    expon_x = (expon_x << 1) + (expon_x << 3);
    expon_x += (int) (c - '0');
 
    ps++;
    c = *ps;
  }
 
  if (c) {
    // return NaN
    res = 0x7c00000000000000ull | sign_x;
    BID_RETURN (res);
  }
 
  if (sgn_expon)
    expon_x = -expon_x;
 
  expon_x += add_expon + DECIMAL_EXPONENT_BIAS;
 
  if (expon_x < 0) {
    if (rounded_up)
      coefficient_x--;
    rnd_mode = 0;
    res =
      get_BID64_UF (sign_x, expon_x, coefficient_x, rounded, rnd_mode,
		    &fpsc);
    BID_RETURN (res);
  }
  res = get_BID64 (sign_x, expon_x, coefficient_x, rnd_mode, &fpsc);
  BID_RETURN (res);
}
 

Go to most recent revision | Compare with Previous | Blame | View Log

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.