Annotation of OpenXM_contrib2/asir2018/asm/ddM.c, Revision 1.1
1.1 ! noro 1: /*
! 2: * Copyright (c) 1994-2000 FUJITSU LABORATORIES LIMITED
! 3: * All rights reserved.
! 4: *
! 5: * FUJITSU LABORATORIES LIMITED ("FLL") hereby grants you a limited,
! 6: * non-exclusive and royalty-free license to use, copy, modify and
! 7: * redistribute, solely for non-commercial and non-profit purposes, the
! 8: * computer program, "Risa/Asir" ("SOFTWARE"), subject to the terms and
! 9: * conditions of this Agreement. For the avoidance of doubt, you acquire
! 10: * only a limited right to use the SOFTWARE hereunder, and FLL or any
! 11: * third party developer retains all rights, including but not limited to
! 12: * copyrights, in and to the SOFTWARE.
! 13: *
! 14: * (1) FLL does not grant you a license in any way for commercial
! 15: * purposes. You may use the SOFTWARE only for non-commercial and
! 16: * non-profit purposes only, such as academic, research and internal
! 17: * business use.
! 18: * (2) The SOFTWARE is protected by the Copyright Law of Japan and
! 19: * international copyright treaties. If you make copies of the SOFTWARE,
! 20: * with or without modification, as permitted hereunder, you shall affix
! 21: * to all such copies of the SOFTWARE the above copyright notice.
! 22: * (3) An explicit reference to this SOFTWARE and its copyright owner
! 23: * shall be made on your publication or presentation in any form of the
! 24: * results obtained by use of the SOFTWARE.
! 25: * (4) In the event that you modify the SOFTWARE, you shall notify FLL by
! 26: * e-mail at risa-admin@sec.flab.fujitsu.co.jp of the detailed specification
! 27: * for such modification or the source code of the modified part of the
! 28: * SOFTWARE.
! 29: *
! 30: * THE SOFTWARE IS PROVIDED AS IS WITHOUT ANY WARRANTY OF ANY KIND. FLL
! 31: * MAKES ABSOLUTELY NO WARRANTIES, EXPRESSED, IMPLIED OR STATUTORY, AND
! 32: * EXPRESSLY DISCLAIMS ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS
! 33: * FOR A PARTICULAR PURPOSE OR NONINFRINGEMENT OF THIRD PARTIES'
! 34: * RIGHTS. NO FLL DEALER, AGENT, EMPLOYEES IS AUTHORIZED TO MAKE ANY
! 35: * MODIFICATIONS, EXTENSIONS, OR ADDITIONS TO THIS WARRANTY.
! 36: * UNDER NO CIRCUMSTANCES AND UNDER NO LEGAL THEORY, TORT, CONTRACT,
! 37: * OR OTHERWISE, SHALL FLL BE LIABLE TO YOU OR ANY OTHER PERSON FOR ANY
! 38: * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE OR CONSEQUENTIAL
! 39: * DAMAGES OF ANY CHARACTER, INCLUDING, WITHOUT LIMITATION, DAMAGES
! 40: * ARISING OUT OF OR RELATING TO THE SOFTWARE OR THIS AGREEMENT, DAMAGES
! 41: * FOR LOSS OF GOODWILL, WORK STOPPAGE, OR LOSS OF DATA, OR FOR ANY
! 42: * DAMAGES, EVEN IF FLL SHALL HAVE BEEN INFORMED OF THE POSSIBILITY OF
! 43: * SUCH DAMAGES, OR FOR ANY CLAIM BY ANY OTHER PARTY. EVEN IF A PART
! 44: * OF THE SOFTWARE HAS BEEN DEVELOPED BY A THIRD PARTY, THE THIRD PARTY
! 45: * DEVELOPER SHALL HAVE NO LIABILITY IN CONNECTION WITH THE USE,
! 46: * PERFORMANCE OR NON-PERFORMANCE OF THE SOFTWARE.
! 47: *
! 48: * $OpenXM$
! 49: */
! 50: #include "ca.h"
! 51: #include "base.h"
! 52: #include "inline.h"
! 53:
! 54: /*
! 55: * mod is declared as 'int', because several xxxum functions contains signed
! 56: * integer addition/subtraction. So mod should be less than 2^31.
! 57: */
! 58:
! 59: void mulum(int mod,UM p1,UM p2,UM pr)
! 60: {
! 61: int *pc1,*pcr;
! 62: int *c1,*c2,*cr;
! 63: unsigned int mul;
! 64: int i,j,d1,d2;
! 65:
! 66: if ( ( (d1 = DEG(p1)) < 0) || ( (d2 = DEG(p2)) < 0 ) ) {
! 67: DEG(pr) = -1;
! 68: return;
! 69: }
! 70: c1 = COEF(p1); c2 = COEF(p2); cr = COEF(pr);
! 71: bzero((char *)cr,(int)((d1+d2+1)*sizeof(int)));
! 72: for ( i = 0; i <= d2; i++, cr++ )
! 73: if ( mul = *c2++ )
! 74: for ( j = 0, pc1 = c1, pcr = cr; j <= d1; j++, pc1++, pcr++ ) {
! 75: DMAR(*pc1,mul,*pcr,mod,*pcr)
! 76: }
! 77: DEG(pr) = d1 + d2;
! 78: }
! 79:
! 80: void mulsum(int mod,UM p,int n,UM pr)
! 81: {
! 82: int *sp,*dp;
! 83: int i;
! 84:
! 85: for ( i = DEG(pr) = DEG(p), sp = COEF(p)+i, dp = COEF(pr)+i;
! 86: i >= 0; i--, dp--, sp-- ) {
! 87: DMAR(*sp,n,0,mod,*dp)
! 88: }
! 89: }
! 90:
! 91: int divum(int mod,UM p1,UM p2,UM pq)
! 92: {
! 93: int *pc1,*pct;
! 94: int *c1,*c2,*ct;
! 95: unsigned int inv,hd,tmp;
! 96: int i,j, d1,d2,dd;
! 97:
! 98: if ( (d1 = DEG(p1)) < (d2 = DEG(p2)) ) {
! 99: DEG(pq) = -1;
! 100: return d1;
! 101: }
! 102: c1 = COEF(p1); c2 = COEF(p2); dd = d1-d2;
! 103: if ( ( hd = c2[d2] ) != 1 ) {
! 104: inv = invm(hd,mod);
! 105: for ( pc1 = c2 + d2; pc1 >= c2; pc1-- ) {
! 106: DMAR(*pc1,inv,0,mod,*pc1)
! 107: }
! 108: } else
! 109: inv = 1;
! 110: for ( i = dd, ct = c1+d1; i >= 0; i-- )
! 111: if ( tmp = *ct-- ) {
! 112: tmp = mod - tmp;
! 113: for ( j = d2-1, pct = ct, pc1 = c2+j; j >= 0; j--, pct--, pc1-- ) {
! 114: DMAR(*pc1,tmp,*pct,mod,*pct)
! 115: }
! 116: }
! 117: if ( inv != 1 ) {
! 118: for ( pc1 = c1+d2, pct = c1+d1; pc1 <= pct; pc1++ ) {
! 119: DMAR(*pc1,inv,0,mod,*pc1)
! 120: }
! 121: for ( pc1 = c2, pct = c2+d2, inv = hd; pc1 <= pct; pc1++ ) {
! 122: DMAR(*pc1,inv,0,mod,*pc1)
! 123: }
! 124: }
! 125: for ( i = d2-1, pc1 = c1+i; i >= 0 && !(*pc1); pc1--, i-- );
! 126: for ( DEG(pq) = j = dd, pc1 = c1+d1, pct = COEF(pq)+j; j >= 0; j-- )
! 127: *pct-- = *pc1--;
! 128: return i;
! 129: }
! 130:
! 131: void diffum(int mod,UM f,UM fd)
! 132: {
! 133: int *dp,*sp;
! 134: int i;
! 135:
! 136: for ( i = DEG(f), dp = COEF(fd)+i-1, sp = COEF(f)+i;
! 137: i >= 1; i--, dp--, sp-- ) {
! 138: DMAR(*sp,i,0,mod,*dp)
! 139: }
! 140: degum(fd,DEG(f) - 1);
! 141: }
! 142:
! 143: unsigned int pwrm(int mod,int a,int n)
! 144: {
! 145: unsigned int s,t;
! 146:
! 147: if ( !n )
! 148: return 1;
! 149: else if ( n == 1 )
! 150: return a;
! 151: else {
! 152: t = pwrm(mod,a,n/2);
! 153: DMAR(t,t,0,mod,s)
! 154: if ( n % 2 ) {
! 155: DMAR(s,a,0,mod,t)
! 156: return t;
! 157: } else
! 158: return s;
! 159: }
! 160: }
! 161:
! 162: unsigned int invm(unsigned int s,int mod)
! 163: {
! 164: unsigned int r,a2,q;
! 165: unsigned int f1,f2,a1;
! 166:
! 167: for ( f1 = s, f2 = mod, a1 = 1, a2 = 0; ; ) {
! 168: q = f1/f2; r = f1 - f2*q; f1 = f2;
! 169: if ( !(f2 = r) )
! 170: break;
! 171: DMAR(a2,q,0,mod,r)
! 172: /* r = ( a1 - r + mod ) % mod; */
! 173: if ( a1 >= r )
! 174: r = a1 - r;
! 175: else {
! 176: r = (mod - r) + a1;
! 177: }
! 178: a1 = a2; a2 = r;
! 179: }
! 180: /* return( ( a2 >= 0 ? a2 : a2 + mod ) ); */
! 181: return a2;
! 182: }
! 183:
! 184: #if !defined(sparc) || defined(__sparcv9)
! 185: void addpadic(int mod,int n,unsigned int *n1,unsigned int *n2)
! 186: {
! 187: unsigned int carry,tmp;
! 188: int i;
! 189:
! 190: for ( i = 0, carry = 0; i < n; i++ ) {
! 191: tmp = *n1++ + *n2 + carry;
! 192: DQR(tmp,mod,carry,*n2++)
! 193: /*
! 194: carry = tmp / mod;
! 195: *n2++ = tmp - ( carry * mod );
! 196: */
! 197: }
! 198: }
! 199: #endif
! 200:
! 201: void mulpadic(int mod,int n,unsigned int *n1,unsigned int *n2,unsigned int *nr)
! 202: {
! 203: unsigned int *pn1,*pnr;
! 204: unsigned int carry,mul;
! 205: int i,j;
! 206:
! 207: bzero((char *)nr,(int)(n*sizeof(int)));
! 208: for ( j = 0; j < n; j++, n2++, nr++ )
! 209: if ( mul = *n2 )
! 210: for ( i = j, carry = 0, pn1 = n1, pnr = nr;
! 211: i < n; i++, pn1++, pnr++ ) {
! 212: carry += *pnr;
! 213: DMAB(mod,*pn1,mul,carry,carry,*pnr)
! 214: }
! 215: }
! 216:
! 217: extern int up_kara_mag;
! 218:
! 219: void kmulum(int mod,UM n1,UM n2,UM nr)
! 220: {
! 221: UM n,t,s,m,carry;
! 222: int d,d1,d2,len,i,l;
! 223: unsigned int *r,*r0;
! 224:
! 225: if ( !n1 || !n2 ) {
! 226: nr->d = -1; return;
! 227: }
! 228: d1 = DEG(n1)+1; d2 = DEG(n2)+1;
! 229: if ( (d1 < up_kara_mag) || (d2 < up_kara_mag) ) {
! 230: mulum(mod,n1,n2,nr); return;
! 231: }
! 232: if ( d1 < d2 ) {
! 233: n = n1; n1 = n2; n2 = n;
! 234: d = d1; d1 = d2; d2 = d;
! 235: }
! 236: if ( d2 > (d1+1)/2 ) {
! 237: kmulummain(mod,n1,n2,nr); return;
! 238: }
! 239: d = (d1/d2)+((d1%d2)!=0);
! 240: len = (d+1)*d2;
! 241: r0 = (unsigned int *)ALLOCA(len*sizeof(int));
! 242: bzero((char *)r0,len*sizeof(int));
! 243: m = W_UMALLOC(d2+1);
! 244: carry = W_UMALLOC(d2+1);
! 245: t = W_UMALLOC(d1+d2+1);
! 246: s = W_UMALLOC(d1+d2+1);
! 247: for ( DEG(carry) = -1, i = 0, r = r0; i < d; i++, r += d2 ) {
! 248: extractum(n1,i*d2,d2,m);
! 249: if ( m ) {
! 250: kmulum(mod,m,n2,t);
! 251: addum(mod,t,carry,s);
! 252: c_copyum(s,d2,r);
! 253: extractum(s,d2,d2,carry);
! 254: } else {
! 255: c_copyum(carry,d2,r);
! 256: carry = 0;
! 257: }
! 258: }
! 259: c_copyum(carry,d2,r);
! 260: for ( l = len - 1; !r0[l]; l-- );
! 261: l++;
! 262: DEG(nr) = l-1;
! 263: bcopy((char *)r0,(char *)COEF(nr),l*sizeof(int));
! 264: }
! 265:
! 266: void ksquareum(int mod,UM n1,UM nr)
! 267: {
! 268: int d1;
! 269:
! 270: if ( !n1 ) {
! 271: nr->d = -1; return;
! 272: }
! 273: d1 = DEG(n1)+1;
! 274: if ( (d1 < up_kara_mag) ) {
! 275: pwrum(mod,n1,2,nr); return;
! 276: }
! 277: ksquareummain(mod,n1,nr);
! 278: }
! 279:
! 280: void extractum(UM n,int index,int len,UM nr)
! 281: {
! 282: int *m;
! 283: int l;
! 284:
! 285: if ( !n ) {
! 286: nr->d = -1; return;
! 287: }
! 288: m = COEF(n)+index;
! 289: if ( (l = (DEG(n)+1)-index) >= len ) {
! 290: for ( l = len - 1; (l >= 0) && !m[l]; l-- );
! 291: l++;
! 292: }
! 293: if ( l <= 0 ) {
! 294: nr->d = -1; return;
! 295: } else {
! 296: DEG(nr) = l-1;
! 297: bcopy((char *)m,(char *)COEF(nr),l*sizeof(Q));
! 298: }
! 299: }
! 300:
! 301: void copyum(UM n1,UM n2)
! 302: {
! 303: n2->d = n1->d;
! 304: bcopy((char *)n1->c,(char *)n2->c,(n1->d+1)*sizeof(int));
! 305: }
! 306:
! 307: void c_copyum(UM n,int len,int *p)
! 308: {
! 309: if ( n )
! 310: bcopy((char *)COEF(n),(char *)p,MIN((DEG(n)+1),len)*sizeof(int));
! 311: }
! 312:
! 313: void kmulummain(int mod,UM n1,UM n2,UM nr)
! 314: {
! 315: int d1,d2,h,len;
! 316: UM n1lo,n1hi,n2lo,n2hi,hi,lo,mid1,mid2,mid,s1,s2,t1,t2;
! 317:
! 318: d1 = DEG(n1)+1; d2 = DEG(n2)+1; h = (d1+1)/2;
! 319: n1lo = W_UMALLOC(d1+1); n1hi = W_UMALLOC(d1+1);
! 320: n2lo = W_UMALLOC(d2+1); n2hi = W_UMALLOC(d2+1);
! 321: lo = W_UMALLOC(d1+d2+1); hi = W_UMALLOC(d1+d2+1);
! 322: mid1 = W_UMALLOC(d1+d2+1); mid2 = W_UMALLOC(d1+d2+1);
! 323: mid = W_UMALLOC(d1+d2+1);
! 324: s1 = W_UMALLOC(d1+d2+1); s2 = W_UMALLOC(d1+d2+1);
! 325: extractum(n1,0,h,n1lo); extractum(n1,h,d1-h,n1hi);
! 326: extractum(n2,0,h,n2lo); extractum(n2,h,d2-h,n2hi);
! 327: kmulum(mod,n1hi,n2hi,hi); kmulum(mod,n1lo,n2lo,lo);
! 328: len = DEG(hi)+1+2*h; t1 = W_UMALLOC(len-1); DEG(t1) = len-1;
! 329: bzero((char *)COEF(t1),len*sizeof(int));
! 330: if ( lo )
! 331: bcopy((char *)COEF(lo),(char *)COEF(t1),(DEG(lo)+1)*sizeof(int));
! 332: if ( hi )
! 333: bcopy((char *)COEF(hi),(char *)(COEF(t1)+2*h),(DEG(hi)+1)*sizeof(int));
! 334:
! 335: addum(mod,hi,lo,mid1);
! 336: subum(mod,n1hi,n1lo,s1); subum(mod,n2lo,n2hi,s2); kmulum(mod,s1,s2,mid2);
! 337: addum(mod,mid1,mid2,mid);
! 338: if ( mid ) {
! 339: len = DEG(mid)+1+h; t2 = W_UMALLOC(len-1); DEG(t2) = len-1;
! 340: bzero((char *)COEF(t2),len*sizeof(int));
! 341: bcopy((char *)COEF(mid),(char *)(COEF(t2)+h),(DEG(mid)+1)*sizeof(int));
! 342: addum(mod,t1,t2,nr);
! 343: } else
! 344: copyum(t1,nr);
! 345: }
! 346:
! 347: void ksquareummain(int mod,UM n1,UM nr)
! 348: {
! 349: int d1,h,len;
! 350: UM n1lo,n1hi,hi,lo,mid1,mid2,mid,s1,t1,t2;
! 351:
! 352: d1 = DEG(n1)+1; h = (d1+1)/2;
! 353: n1lo = W_UMALLOC(d1+1); n1hi = W_UMALLOC(d1+1);
! 354: lo = W_UMALLOC(2*d1+1); hi = W_UMALLOC(2*d1+1);
! 355: mid1 = W_UMALLOC(2*d1+1); mid2 = W_UMALLOC(2*d1+1);
! 356: mid = W_UMALLOC(2*d1+1);
! 357: s1 = W_UMALLOC(2*d1+1);
! 358: extractum(n1,0,h,n1lo); extractum(n1,h,d1-h,n1hi);
! 359: ksquareum(mod,n1hi,hi); ksquareum(mod,n1lo,lo);
! 360: len = DEG(hi)+1+2*h; t1 = W_UMALLOC(len-1); DEG(t1) = len-1;
! 361: bzero((char *)COEF(t1),len*sizeof(int));
! 362: if ( lo )
! 363: bcopy((char *)COEF(lo),(char *)COEF(t1),(DEG(lo)+1)*sizeof(int));
! 364: if ( hi )
! 365: bcopy((char *)COEF(hi),(char *)(COEF(t1)+2*h),(DEG(hi)+1)*sizeof(int));
! 366:
! 367: addum(mod,hi,lo,mid1);
! 368: subum(mod,n1hi,n1lo,s1); ksquareum(mod,s1,mid2);
! 369: subum(mod,mid1,mid2,mid);
! 370: if ( mid ) {
! 371: len = DEG(mid)+1+h; t2 = W_UMALLOC(len-1); DEG(t2) = len-1;
! 372: bzero((char *)COEF(t2),len*sizeof(int));
! 373: bcopy((char *)COEF(mid),(char *)(COEF(t2)+h),(DEG(mid)+1)*sizeof(int));
! 374: addum(mod,t1,t2,nr);
! 375: } else
! 376: copyum(t1,nr);
! 377: }
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>