URL
https://opencores.org/ocsvn/openrisc/openrisc/trunk
Subversion Repositories openrisc
[/] [openrisc/] [trunk/] [gnu-stable/] [gcc-4.5.1/] [libdecnumber/] [bid/] [bid2dpd_dpd2bid.c] - Rev 841
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/>. */ #undef IN_LIBGCC2 #include "bid-dpd.h" /* get full 64x64bit product */ #define __mul_64x64_to_128(P, CX, CY) \ { \ UINT64 CXH, CXL, CYH,CYL,PL,PH,PM,PM2; \ CXH = (CX) >> 32; \ CXL = (UINT32)(CX); \ CYH = (CY) >> 32; \ CYL = (UINT32)(CY); \ \ PM = CXH*CYL; \ PH = CXH*CYH; \ PL = CXL*CYL; \ PM2 = CXL*CYH; \ PH += (PM>>32); \ PM = (UINT64)((UINT32)PM)+PM2+(PL>>32); \ \ (P).w[1] = PH + (PM>>32); \ (P).w[0] = (PM<<32)+(UINT32)PL; \ } /* add 64-bit value to 128-bit */ #define __add_128_64(R128, A128, B64) \ { \ UINT64 R64H; \ R64H = (A128).w[1]; \ (R128).w[0] = (B64) + (A128).w[0]; \ if((R128).w[0] < (B64)) R64H ++; \ (R128).w[1] = R64H; \ } /* add 128-bit value to 128-bit (assume no carry-out) */ #define __add_128_128(R128, A128, B128) \ { \ UINT128 Q128; \ Q128.w[1] = (A128).w[1]+(B128).w[1]; \ Q128.w[0] = (B128).w[0] + (A128).w[0]; \ if(Q128.w[0] < (B128).w[0]) Q128.w[1] ++; \ (R128).w[1] = Q128.w[1]; \ (R128).w[0] = Q128.w[0]; \ } #define __mul_128x128_high(Q, A, B) \ { \ UINT128 ALBL, ALBH, AHBL, AHBH, QM, QM2; \ \ __mul_64x64_to_128(ALBH, (A).w[0], (B).w[1]); \ __mul_64x64_to_128(AHBL, (B).w[0], (A).w[1]); \ __mul_64x64_to_128(ALBL, (A).w[0], (B).w[0]); \ __mul_64x64_to_128(AHBH, (A).w[1],(B).w[1]); \ \ __add_128_128(QM, ALBH, AHBL); \ __add_128_64(QM2, QM, ALBL.w[1]); \ __add_128_64((Q), AHBH, QM2.w[1]); \ } #include "bid2dpd_dpd2bid.h" static const unsigned int dm103[] = { 0, 1000, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, 10000, 11000 }; void _bid_to_dpd32 (_Decimal32 *, _Decimal32 *); void _bid_to_dpd32 (_Decimal32 *pres, _Decimal32 *px) { unsigned int sign, coefficient_x, exp, dcoeff; unsigned int b2, b1, b0, b01, res; _Decimal32 x = *px; sign = (x & 0x80000000); if ((x & 0x60000000ul) == 0x60000000ul) { /* special encodings */ if ((x & 0x78000000ul) == 0x78000000ul) { *pres = x; /* NaN or Infinity */ return; } /* coefficient */ coefficient_x = (x & 0x001ffffful) | 0x00800000ul; if (coefficient_x >= 10000000) coefficient_x = 0; /* get exponent */ exp = (x >> 21) & 0xff; } else { exp = (x >> 23) & 0xff; coefficient_x = (x & 0x007ffffful); } b01 = coefficient_x / 1000; b2 = coefficient_x - 1000 * b01; b0 = b01 / 1000; b1 = b01 - 1000 * b0; dcoeff = b2d[b2] | b2d2[b1]; if (b0 >= 8) { /* is b0 8 or 9? */ res = sign | ((0x600 | ((exp >> 6) << 7) | ((b0 & 1) << 6) | (exp & 0x3f)) << 20) | dcoeff; } else { /* else b0 is 0..7 */ res = sign | ((((exp >> 6) << 9) | (b0 << 6) | (exp & 0x3f)) << 20) | dcoeff; } *pres = res; } void _dpd_to_bid32 (_Decimal32 *, _Decimal32 *); void _dpd_to_bid32 (_Decimal32 *pres, _Decimal32 *px) { unsigned int r; unsigned int sign, exp, bcoeff; UINT64 trailing; unsigned int d0, d1, d2; _Decimal32 x = *px; sign = (x & 0x80000000); trailing = (x & 0x000fffff); if ((x & 0x78000000) == 0x78000000) { *pres = x; return; } else { /* normal number */ if ((x & 0x60000000) == 0x60000000) { /* G0..G1 = 11 -> d0 = 8 + G4 */ d0 = d2b3[((x >> 26) & 1) | 8]; /* d0 = (comb & 0x0100 ? 9 : 8); */ exp = (x >> 27) & 3; /* exp leading bits are G2..G3 */ } else { d0 = d2b3[(x >> 26) & 0x7]; exp = (x >> 29) & 3; /* exp loading bits are G0..G1 */ } d1 = d2b2[(trailing >> 10) & 0x3ff]; d2 = d2b[(trailing) & 0x3ff]; bcoeff = d2 + d1 + d0; exp = (exp << 6) + ((x >> 20) & 0x3f); if (bcoeff < (1 << 23)) { r = exp; r <<= 23; r |= (bcoeff | sign); } else { r = exp; r <<= 21; r |= (sign | 0x60000000ul); /* add coeff, without leading bits */ r |= (((unsigned int) bcoeff) & 0x1fffff); } } *pres = r; } void _bid_to_dpd64 (_Decimal64 *, _Decimal64 *); void _bid_to_dpd64 (_Decimal64 *pres, _Decimal64 *px) { UINT64 res; UINT64 sign, comb, exp, B34, B01; UINT64 d103, D61; UINT64 b0, b2, b3, b5; unsigned int b1, b4; UINT64 bcoeff; UINT64 dcoeff; unsigned int yhi, ylo; _Decimal64 x = *px; sign = (x & 0x8000000000000000ull); comb = (x & 0x7ffc000000000000ull) >> 51; if ((comb & 0xf00) == 0xf00) { *pres = x; return; } else { /* Normal number */ if ((comb & 0xc00) == 0xc00) { /* G0..G1 = 11 -> exp is G2..G11 */ exp = (comb) & 0x3ff; bcoeff = (x & 0x0007ffffffffffffull) | 0x0020000000000000ull; } else { exp = (comb >> 2) & 0x3ff; bcoeff = (x & 0x001fffffffffffffull); } D61 = 2305843009ull; /* Floor(2^61 / 10^9) */ /* Multiply the binary coefficient by ceil(2^64 / 1000), and take the upper 64-bits in order to compute a division by 1000. */ yhi = (D61 * (UINT64)(bcoeff >> (UINT64)27)) >> (UINT64)34; ylo = bcoeff - 1000000000ull * yhi; if (ylo >= 1000000000) { ylo = ylo - 1000000000; yhi = yhi + 1; } d103 = 0x4189374c; B34 = ((UINT64) ylo * d103) >> (32 + 8); B01 = ((UINT64) yhi * d103) >> (32 + 8); b5 = ylo - B34 * 1000; b2 = yhi - B01 * 1000; b3 = ((UINT64) B34 * d103) >> (32 + 8); b0 = ((UINT64) B01 * d103) >> (32 + 8); b4 = (unsigned int) B34 - (unsigned int) b3 *1000; b1 = (unsigned int) B01 - (unsigned int) dm103[b0]; dcoeff = b2d[b5] | b2d2[b4] | b2d3[b3] | b2d4[b2] | b2d5[b1]; if (b0 >= 8) /* is b0 8 or 9? */ res = sign | ((0x1800 | ((exp >> 8) << 9) | ((b0 & 1) << 8) | (exp & 0xff)) << 50) | dcoeff; else /* else b0 is 0..7 */ res = sign | ((((exp >> 8) << 11) | (b0 << 8) | (exp & 0xff)) << 50) | dcoeff; } *pres = res; } void _dpd_to_bid64 (_Decimal64 *, _Decimal64 *); void _dpd_to_bid64 (_Decimal64 *pres, _Decimal64 *px) { UINT64 res; UINT64 sign, comb, exp; UINT64 trailing; UINT64 d0, d1, d2; unsigned int d3, d4, d5; UINT64 bcoeff, mask; _Decimal64 x = *px; sign = (x & 0x8000000000000000ull); comb = (x & 0x7ffc000000000000ull) >> 50; trailing = (x & 0x0003ffffffffffffull); if ((comb & 0x1e00) == 0x1e00) { if ((comb & 0x1f00) == 0x1f00) { /* G0..G4 = 11111 -> NaN */ if (comb & 0x0100) { /* G5 = 1 -> sNaN */ *pres = x; } else { /* G5 = 0 -> qNaN */ *pres = x; } } else { /*if ((comb & 0x1e00) == 0x1e00); G0..G4 = 11110 -> INF */ *pres = x; } return; } else { /* normal number */ if ((comb & 0x1800) == 0x1800) { /* G0..G1 = 11 -> d0 = 8 + G4 */ d0 = d2b6[((comb >> 8) & 1) | 8]; /* d0 = (comb & 0x0100 ? 9 : 8); */ exp = (comb & 0x600) >> 1; /* exp = (comb & 0x0400 ? 1 : 0) * 0x200 + (comb & 0x0200 ? 1 : 0) * 0x100; exp leading bits are G2..G3 */ } else { d0 = d2b6[(comb >> 8) & 0x7]; exp = (comb & 0x1800) >> 3; /* exp = (comb & 0x1000 ? 1 : 0) * 0x200 + (comb & 0x0800 ? 1 : 0) * 0x100; exp loading bits are G0..G1 */ } d1 = d2b5[(trailing >> 40) & 0x3ff]; d2 = d2b4[(trailing >> 30) & 0x3ff]; d3 = d2b3[(trailing >> 20) & 0x3ff]; d4 = d2b2[(trailing >> 10) & 0x3ff]; d5 = d2b[(trailing) & 0x3ff]; bcoeff = (d5 + d4 + d3) + d2 + d1 + d0; exp += (comb & 0xff); mask = 1; mask <<= 53; if (bcoeff < mask) { /* check whether coefficient fits in 10*5+3 bits */ res = exp; res <<= 53; res |= (bcoeff | sign); *pres = res; return; } /* special format */ res = (exp << 51) | (sign | 0x6000000000000000ull); /* add coeff, without leading bits */ mask = (mask >> 2) - 1; bcoeff &= mask; res |= bcoeff; } *pres = res; } void _bid_to_dpd128 (_Decimal128 *, _Decimal128 *); void _bid_to_dpd128 (_Decimal128 *pres, _Decimal128 *px) { UINT128 res; UINT128 sign; unsigned int comb; UINT128 bcoeff; UINT128 dcoeff; UINT128 BH, d1018, BT2, BT1; UINT64 exp, BL, d109; UINT64 d106, d103; UINT64 k1, k2, k4, k5, k7, k8, k10, k11; unsigned int BHH32, BLL32, BHL32, BLH32, k0, k3, k6, k9, amount; _Decimal128 x = *px; sign.w[1] = (x.w[1] & 0x8000000000000000ull); sign.w[0] = 0; comb = (x.w[1] /*& 0x7fffc00000000000ull */ ) >> 46; exp = 0; if ((comb & 0x1e000) == 0x1e000) { if ((comb & 0x1f000) == 0x1f000) { /* G0..G4 = 11111 -> NaN */ if (comb & 0x01000) { /* G5 = 1 -> sNaN */ res = x; } else { /* G5 = 0 -> qNaN */ res = x; } } else { /* G0..G4 = 11110 -> INF */ res = x; } } else { /* normal number */ exp = ((x.w[1] & 0x7fff000000000000ull) >> 49) & 0x3fff; bcoeff.w[1] = (x.w[1] & 0x0001ffffffffffffull); bcoeff.w[0] = x.w[0]; d1018 = reciprocals10_128[18]; __mul_128x128_high (BH, bcoeff, d1018); amount = recip_scale[18]; BH.w[0] = (BH.w[0] >> amount) | (BH.w[1] << (64 - amount)); BL = bcoeff.w[0] - BH.w[0] * 1000000000000000000ull; d109 = reciprocals10_64[9]; __mul_64x64_to_128 (BT1, BH.w[0], d109); BHH32 = (unsigned int) (BT1.w[1] >> short_recip_scale[9]); BHL32 = (unsigned int) BH.w[0] - BHH32 * 1000000000; __mul_64x64_to_128 (BT2, BL, d109); BLH32 = (unsigned int) (BT2.w[1] >> short_recip_scale[9]); BLL32 = (unsigned int) BL - BLH32 * 1000000000; d106 = 0x431BDE83; d103 = 0x4189374c; k0 = ((UINT64) BHH32 * d106) >> (32 + 18); BHH32 -= (unsigned int) k0 *1000000; k1 = ((UINT64) BHH32 * d103) >> (32 + 8); k2 = BHH32 - (unsigned int) k1 *1000; k3 = ((UINT64) BHL32 * d106) >> (32 + 18); BHL32 -= (unsigned int) k3 *1000000; k4 = ((UINT64) BHL32 * d103) >> (32 + 8); k5 = BHL32 - (unsigned int) k4 *1000; k6 = ((UINT64) BLH32 * d106) >> (32 + 18); BLH32 -= (unsigned int) k6 *1000000; k7 = ((UINT64) BLH32 * d103) >> (32 + 8); k8 = BLH32 - (unsigned int) k7 *1000; k9 = ((UINT64) BLL32 * d106) >> (32 + 18); BLL32 -= (unsigned int) k9 *1000000; k10 = ((UINT64) BLL32 * d103) >> (32 + 8); k11 = BLL32 - (unsigned int) k10 *1000; dcoeff.w[1] = (b2d[k5] >> 4) | (b2d[k4] << 6) | (b2d[k3] << 16) | (b2d[k2] << 26) | (b2d[k1] << 36); dcoeff.w[0] = b2d[k11] | (b2d[k10] << 10) | (b2d[k9] << 20) | (b2d[k8] << 30) | (b2d[k7] << 40) | (b2d[k6] << 50) | (b2d[k5] << 60); res.w[0] = dcoeff.w[0]; if (k0 >= 8) { res.w[1] = sign.w[1] | ((0x18000 | ((exp >> 12) << 13) | ((k0 & 1) << 12) | (exp & 0xfff)) << 46) | dcoeff.w[1]; } else { res.w[1] = sign.w[1] | ((((exp >> 12) << 15) | (k0 << 12) | (exp & 0xfff)) << 46) | dcoeff.w[1]; } } *pres = res; } void _dpd_to_bid128 (_Decimal128 *, _Decimal128 *); void _dpd_to_bid128 (_Decimal128 *pres, _Decimal128 *px) { UINT128 res; UINT128 sign; UINT64 exp, comb; UINT128 trailing; UINT64 d0, d1, d2, d3, d4, d5, d6, d7, d8, d9, d10, d11; UINT128 bcoeff; UINT64 tl, th; _Decimal128 x = *px; sign.w[1] = (x.w[1] & 0x8000000000000000ull); sign.w[0] = 0; comb = (x.w[1] & 0x7fffc00000000000ull) >> 46; trailing.w[1] = x.w[1]; trailing.w[0] = x.w[0]; if ((comb & 0x1e000) == 0x1e000) { if ((comb & 0x1f000) == 0x1f000) { /* G0..G4 = 11111 -> NaN */ if (comb & 0x01000) { /* G5 = 1 -> sNaN */ *pres = x; } else { /* G5 = 0 -> qNaN */ *pres = x; } } else { /* G0..G4 = 11110 -> INF */ *pres = x; } return; } else { /* Normal number */ if ((comb & 0x18000) == 0x18000) { /* G0..G1 = 11 -> d0 = 8 + G4 */ d0 = d2b6[8 + ((comb & 0x01000) >> 12)]; exp = (comb & 0x06000) >> 1; /* exp leading bits are G2..G3 */ } else { d0 = d2b6[((comb & 0x07000) >> 12)]; exp = (comb & 0x18000) >> 3; /* exp loading bits are G0..G1 */ } d11 = d2b[(trailing.w[0]) & 0x3ff]; d10 = d2b2[(trailing.w[0] >> 10) & 0x3ff]; d9 = d2b3[(trailing.w[0] >> 20) & 0x3ff]; d8 = d2b4[(trailing.w[0] >> 30) & 0x3ff]; d7 = d2b5[(trailing.w[0] >> 40) & 0x3ff]; d6 = d2b6[(trailing.w[0] >> 50) & 0x3ff]; d5 = d2b[(trailing.w[0] >> 60) | ((trailing.w[1] & 0x3f) << 4)]; d4 = d2b2[(trailing.w[1] >> 6) & 0x3ff]; d3 = d2b3[(trailing.w[1] >> 16) & 0x3ff]; d2 = d2b4[(trailing.w[1] >> 26) & 0x3ff]; d1 = d2b5[(trailing.w[1] >> 36) & 0x3ff]; tl = d11 + d10 + d9 + d8 + d7 + d6; th = d5 + d4 + d3 + d2 + d1 + d0; __mul_64x64_to_128 (bcoeff, th, 1000000000000000000ull); __add_128_64 (bcoeff, bcoeff, tl); exp += (comb & 0xfff); res.w[0] = bcoeff.w[0]; res.w[1] = (exp << 49) | sign.w[1] | bcoeff.w[1]; } *pres = res; }
Go to most recent revision | Compare with Previous | Blame | View Log