added libtommath-0.11

This commit is contained in:
Tom St Denis 2003-02-28 16:07:58 +00:00 committed by Steffen Jaeckel
parent fb93a30a25
commit 33c5019985
9 changed files with 885 additions and 828 deletions

5
b.bat
View File

@ -1,3 +1,2 @@
nasm -f coff timer.asm
gcc -Wall -W -O3 -fomit-frame-pointer -funroll-loops -DTIMER_X86 demo.c bn.c timer.o -o ltmdemo
rem gcc -I./mtest/ -DU_MPI -Wall -W -O3 -fomit-frame-pointer -funroll-loops -DTIMER_X86 demo.c mtest/mpi.c timer.o -o mpidemo
nasm -f elf timer.asm
gcc -Wall -W -O3 -fomit-frame-pointer -funroll-loops -DTIMER_X86 demo.c bn.c timer.o -o ltmdemo

152
bn.c
View File

@ -99,7 +99,8 @@ void dump_timings(void)
memset(&functime, 0, sizeof(functime));
total = 0;
for (x = 0; x < _itims; x++) {
total += timings[x].tot;
if (strcmp(timings[x].func, "_verify"))
total += timings[x].tot;
/* try to find this entry */
for (y = 0; functime[y].func != NULL; y++) {
@ -1053,7 +1054,7 @@ static int fast_s_mp_mul_digs(mp_int *a, mp_int *b, mp_int *c, int digs)
c->dp[digs-1] = (mp_digit)(W[digs-1] & ((mp_word)MP_MASK));
/* clear unused */
for (ix = c->used; ix < olduse; ix++) {
for (; ix < olduse; ix++) {
c->dp[ix] = 0;
}
@ -1194,13 +1195,13 @@ static int fast_s_mp_mul_high_digs(mp_int *a, mp_int *b, mp_int *c, int digs)
c->used = newused;
/* now convert the array W downto what we need */
for (ix = digs+1; ix < (pa+pb+1); ix++) {
for (ix = digs+1; ix < newused; ix++) {
W[ix] += (W[ix-1] >> ((mp_word)DIGIT_BIT));
c->dp[ix-1] = (mp_digit)(W[ix-1] & ((mp_word)MP_MASK));
}
c->dp[(pa+pb+1)-1] = (mp_digit)(W[(pa+pb+1)-1] & ((mp_word)MP_MASK));
for (ix = c->used; ix < oldused; ix++) {
for (; ix < oldused; ix++) {
c->dp[ix] = 0;
}
mp_clamp(c);
@ -1339,17 +1340,17 @@ static int fast_s_mp_sqr(mp_int *a, mp_int *b)
b->used = newused;
/* now compute digits */
for (ix = 1; ix < (pa+pa+1); ix++) {
for (ix = 1; ix < newused; ix++) {
/* double/add next digit */
W[ix] += W[ix] + W2[ix];
W[ix] = W[ix] + (W[ix-1] >> ((mp_word)DIGIT_BIT));
b->dp[ix-1] = (mp_digit)(W[ix-1] & ((mp_word)MP_MASK));
}
b->dp[(pa+pa+1)-1] = (mp_digit)(W[(pa+pa+1)-1] & ((mp_word)MP_MASK));
b->dp[(newused)-1] = (mp_digit)(W[(newused)-1] & ((mp_word)MP_MASK));
/* clear high */
for (ix = b->used; ix < olduse; ix++) {
for (; ix < olduse; ix++) {
b->dp[ix] = 0;
}
@ -1580,9 +1581,7 @@ static int mp_karatsuba_mul(mp_int *a, mp_int *b, mp_int *c)
}
mp_clamp(&x0);
mp_clamp(&x1);
mp_clamp(&y0);
mp_clamp(&y1);
/* now calc the products x0y0 and x1y1 */
if (mp_mul(&x0, &y0, &x0y0) != MP_OKAY) goto X1Y1; /* x0y0 = x0*y0 */
@ -1679,15 +1678,14 @@ static int mp_karatsuba_sqr(mp_int *a, mp_int *b)
x1.used = a->used - B;
mp_clamp(&x0);
mp_clamp(&x1);
/* now calc the products x0*x0 and x1*x1 */
if (mp_sqr(&x0, &x0x0) != MP_OKAY) goto X1X1; /* x0x0 = x0*x0 */
if (mp_sqr(&x1, &x1x1) != MP_OKAY) goto X1X1; /* x1x1 = x1*x1 */
if (mp_sqr(&x0, &x0x0) != MP_OKAY) goto X1X1; /* x0x0 = x0*x0 */
if (mp_sqr(&x1, &x1x1) != MP_OKAY) goto X1X1; /* x1x1 = x1*x1 */
/* now calc x1-x0 and y1-y0 */
if (mp_sub(&x1, &x0, &t1) != MP_OKAY) goto X1X1; /* t1 = x1 - x0 */
if (mp_sqr(&t1, &t1) != MP_OKAY) goto X1X1; /* t1 = (x1 - x0) * (y1 - y0) */
if (mp_sqr(&t1, &t1) != MP_OKAY) goto X1X1; /* t1 = (x1 - x0) * (y1 - y0) */
/* add x0y0 */
if (mp_add(&x0x0, &x1x1, &t2) != MP_OKAY) goto X1X1; /* t2 = x0y0 + x1y1 */
@ -2760,8 +2758,7 @@ int mp_reduce_setup(mp_int *a, mp_int *b)
VERIFY(a);
VERIFY(b);
mp_set(a, 1);
if ((res = mp_lshd(a, b->used * 2)) != MP_OKAY) {
if ((res = mp_2expt(a, b->used * 2 * DIGIT_BIT)) != MP_OKAY) {
DECFUNC();
return res;
}
@ -2876,7 +2873,6 @@ __T: mp_clear(&t);
return res;
}
/* computes xR^-1 == x (mod N) via Montgomery Reduction (comba) */
static int fast_mp_montgomery_reduce(mp_int *a, mp_int *m, mp_digit mp)
{
@ -2884,29 +2880,53 @@ static int fast_mp_montgomery_reduce(mp_int *a, mp_int *m, mp_digit mp)
mp_digit ui;
mp_word W[512];
REGFUNC("fast_mp_montgomery_reduce");
VERIFY(a);
VERIFY(m);
/* get old used count */
olduse = a->used;
/* grow a as required */
if (a->alloc < m->used*2+1) {
if ((res = mp_grow(a, m->used*2+1)) != MP_OKAY) {
if (a->alloc < m->used+1) {
if ((res = mp_grow(a, m->used+1)) != MP_OKAY) {
DECFUNC();
return res;
}
}
/* copy and clear */
/* copy the digits of a */
for (ix = 0; ix < a->used; ix++) {
W[ix] = a->dp[ix];
}
/* zero the high words */
for (; ix < m->used * 2 + 1; ix++) {
W[ix] = 0;
}
for (ix = 0; ix < m->used; ix++) {
/* ui = ai * m' mod b */
/* ui = ai * m' mod b
*
* We avoid a double precision multiplication (which isn't required)
* by casting the value down to a mp_digit. Note this requires that W[ix-1] have
* the carry cleared (see after the inner loop)
*/
ui = (((mp_digit)(W[ix] & MP_MASK)) * mp) & MP_MASK;
/* a = a + ui * m * b^i */
/* a = a + ui * m * b^i
*
* This is computed in place and on the fly. The multiplication
* by b^i is handled by offseting which columns the results
* are added to.
*
* Note the comba method normally doesn't handle carries in the inner loop
* In this case we fix the carry from the previous column since the Montgomery
* reduction requires digits of the result (so far) [see above] to work. This is
* handled by fixing up one carry after the inner loop. The carry fixups are done
* in order so after these loops the first m->used words of W[] have the carries
* fixed
*/
{
register int iy;
register mp_digit *tmpx;
@ -2916,32 +2936,36 @@ static int fast_mp_montgomery_reduce(mp_int *a, mp_int *m, mp_digit mp)
tmpx = m->dp;
_W = W + ix;
/* inner loop */
for (iy = 0; iy < m->used; iy++) {
*_W++ += ((mp_word)ui) * ((mp_word)*tmpx++);
}
/* now fix carry for W[ix+1] */
W[ix+1] += W[ix] >> ((mp_word)DIGIT_BIT);
W[ix] &= ((mp_word)MP_MASK);
}
/* now fix carry for next digit, W[ix+1] */
W[ix+1] += W[ix] >> ((mp_word)DIGIT_BIT);
}
/* nox fix rest of carries */
for (; ix <= m->used * 2 + 1; ix++) {
for (++ix; ix <= m->used * 2 + 1; ix++) {
W[ix] += (W[ix-1] >> ((mp_word)DIGIT_BIT));
W[ix-1] &= ((mp_word)MP_MASK);
}
/* copy out */
/* A = A/b^n */
/* copy out, A = A/b^n
*
* The result is A/b^n but instead of converting from an array of mp_word
* to mp_digit than calling mp_rshd we just copy them in the right
* order
*/
for (ix = 0; ix < m->used + 1; ix++) {
a->dp[ix] = W[ix+m->used];
a->dp[ix] = W[ix+m->used] & ((mp_word)MP_MASK);
}
/* set the max used */
a->used = m->used + 1;
/* zero oldused digits */
/* zero oldused digits, if the input a was larger than
* m->used+1 we'll have to clear the digits */
for (; ix < olduse; ix++) {
a->dp[ix] = 0;
}
@ -2951,10 +2975,12 @@ static int fast_mp_montgomery_reduce(mp_int *a, mp_int *m, mp_digit mp)
/* if A >= m then A = A - m */
if (mp_cmp_mag(a, m) != MP_LT) {
if ((res = s_mp_sub(a, m, a)) != MP_OKAY) {
DECFUNC();
return res;
}
}
}
DECFUNC();
return MP_OKAY;
}
@ -3036,7 +3062,7 @@ int mp_montgomery_reduce(mp_int *a, mp_int *m, mp_digit mp)
*/
static int mp_exptmod_fast(mp_int *G, mp_int *X, mp_int *P, mp_int *Y)
{
mp_int M[64], res;
mp_int M[256], res;
mp_digit buf, mp;
int err, bitbuf, bitcpy, bitcnt, mode, digidx, x, y, winsize;
@ -3048,12 +3074,14 @@ static int mp_exptmod_fast(mp_int *G, mp_int *X, mp_int *P, mp_int *Y)
/* find window size */
x = mp_count_bits(X);
if (x <= 18) { winsize = 2; }
else if (x <= 84) { winsize = 3; }
else if (x <= 300) { winsize = 4; }
else if (x <= 930) { winsize = 5; }
else { winsize = 6; }
if (x <= 7) { winsize = 2; }
else if (x <= 36) { winsize = 3; }
else if (x <= 140) { winsize = 4; }
else if (x <= 450) { winsize = 5; }
else if (x <= 1303) { winsize = 6; }
else if (x <= 3529) { winsize = 7; }
else { winsize = 8; }
/* init G array */
for (x = 0; x < (1<<winsize); x++) {
if ((err = mp_init_size(&M[x], 1)) != MP_OKAY) {
@ -3072,15 +3100,14 @@ static int mp_exptmod_fast(mp_int *G, mp_int *X, mp_int *P, mp_int *Y)
/* setup result */
if ((err = mp_init(&res)) != MP_OKAY) {
goto __M;
goto __RES;
}
/* now we need R mod m */
mp_set(&res, 1);
if ((err = mp_lshd(&res, P->used)) != MP_OKAY) {
if ((err = mp_2expt(&res, P->used * DIGIT_BIT)) != MP_OKAY) {
goto __RES;
}
/* res = R mod m */
if ((err = mp_mod(&res, P, &res)) != MP_OKAY) {
goto __RES;
@ -3092,7 +3119,6 @@ static int mp_exptmod_fast(mp_int *G, mp_int *X, mp_int *P, mp_int *Y)
*
* The first half of the table is not computed though accept for M[0] and M[1]
*/
mp_set(&M[0], 1);
if ((err = mp_mod(G, P, &M[1])) != MP_OKAY) {
goto __RES;
}
@ -3101,7 +3127,7 @@ static int mp_exptmod_fast(mp_int *G, mp_int *X, mp_int *P, mp_int *Y)
if ((err = mp_mulmod(&M[1], &res, P, &M[1])) != MP_OKAY) {
goto __RES;
}
/* compute the value at M[1<<(winsize-1)] by squaring M[1] (winsize-1) times */
if ((err = mp_copy(&M[1], &M[1<<(winsize-1)])) != MP_OKAY) {
goto __RES;
@ -3236,10 +3262,9 @@ __M :
return err;
}
int mp_exptmod(mp_int *G, mp_int *X, mp_int *P, mp_int *Y)
{
mp_int M[64], res, mu;
mp_int M[256], res, mu;
mp_digit buf;
int err, bitbuf, bitcpy, bitcnt, mode, digidx, x, y, winsize;
@ -3258,11 +3283,13 @@ int mp_exptmod(mp_int *G, mp_int *X, mp_int *P, mp_int *Y)
/* find window size */
x = mp_count_bits(X);
if (x <= 18) { winsize = 2; }
else if (x <= 84) { winsize = 3; }
else if (x <= 300) { winsize = 4; }
else if (x <= 930) { winsize = 5; }
else { winsize = 6; }
if (x <= 7) { winsize = 2; }
else if (x <= 36) { winsize = 3; }
else if (x <= 140) { winsize = 4; }
else if (x <= 450) { winsize = 5; }
else if (x <= 1303) { winsize = 6; }
else if (x <= 3529) { winsize = 7; }
else { winsize = 8; }
/* init G array */
for (x = 0; x < (1<<winsize); x++) {
@ -3289,7 +3316,6 @@ int mp_exptmod(mp_int *G, mp_int *X, mp_int *P, mp_int *Y)
*
* The first half of the table is not computed though accept for M[0] and M[1]
*/
mp_set(&M[0], 1);
if ((err = mp_mod(G, P, &M[1])) != MP_OKAY) {
goto __MU;
}
@ -3430,6 +3456,22 @@ __M :
return err;
}
/* computes a = 2^b */
int mp_2expt(mp_int *a, int b)
{
int res;
mp_zero(a);
if ((res = mp_grow(a, b/DIGIT_BIT + 1)) != MP_OKAY) {
return res;
}
a->used = b/DIGIT_BIT + 1;
a->dp[b/DIGIT_BIT] = 1 << (b % DIGIT_BIT);
return MP_OKAY;
}
/* find the n'th root of an integer
*
* Result found such that (c)^b <= a and (c+1)^b > a

3
bn.h
View File

@ -158,6 +158,9 @@ int mp_mul_2(mp_int *a, mp_int *b);
/* c = a mod 2^d */
int mp_mod_2d(mp_int *a, int b, mp_int *c);
/* computes a = 2^b */
int mp_2expt(mp_int *a, int b);
/* ---> Basic arithmetic <--- */
/* b = -a */

BIN
bn.pdf

Binary file not shown.

1255
bn.tex

File diff suppressed because it is too large Load Diff

View File

@ -1,70 +1,76 @@
Jan 9th, 2003
v0.10 -- Pekka Riikonen suggested fixes to the radix conversion code.
-- Added baseline montgomery and comba montgomery reductions, sped up exptmods
[to a point, see bn.h for MONTGOMERY_EXPT_CUTOFF]
Jan 6th, 2003
v0.09 -- Updated the manual to reflect recent changes. :-)
-- Added Jacobi function (mp_jacobi) to supplement the number theory side of the lib
-- Added a Mersenne prime finder demo in ./etc/mersenne.c
Jan 2nd, 2003
v0.08 -- Sped up the multipliers by moving the inner loop variables into a smaller scope
-- Corrected a bunch of small "warnings"
-- Added more comments
-- Made "mtest" be able to use /dev/random, /dev/urandom or stdin for RNG data
-- Corrected some bugs where error messages were potentially ignored
-- add etc/pprime.c program which makes numbers which are provably prime.
Jan 1st, 2003
v0.07 -- Removed alot of heap operations from core functions to speed them up
-- Added a root finding function [and mp_sqrt macro like from MPI]
-- Added more to manual
Dec 31st, 2002
v0.06 -- Sped up the s_mp_add, s_mp_sub which inturn sped up mp_invmod, mp_exptmod, etc...
-- Cleaned up the header a bit more
Dec 30th, 2002
v0.05 -- Builds with MSVC out of the box
-- Fixed a bug in mp_invmod w.r.t. even moduli
-- Made mp_toradix and mp_read_radix use char instead of unsigned char arrays
-- Fixed up exptmod to use fewer multiplications
-- Fixed up mp_init_size to use only one heap operation
-- Note there is a slight "off-by-one" bug in the library somewhere
without the padding (see the source for comment) the library
crashes in libtomcrypt. Anyways a reasonable workaround is to pad the
numbers which will always correct it since as the numbers grow the padding
will still be beyond the end of the number
-- Added more to the manual
Dec 29th, 2002
v0.04 -- Fixed a memory leak in mp_to_unsigned_bin
-- optimized invmod code
-- Fixed bug in mp_div
-- use exchange instead of copy for results
-- added a bit more to the manual
Dec 27th, 2002
v0.03 -- Sped up s_mp_mul_high_digs by not computing the carries of the lower digits
-- Fixed a bug where mp_set_int wouldn't zero the value first and set the used member.
-- fixed a bug in s_mp_mul_high_digs where the limit placed on the result digits was not calculated properly
-- fixed bugs in add/sub/mul/sqr_mod functions where if the modulus and dest were the same it wouldn't work
-- fixed a bug in mp_mod and mp_mod_d concerning negative inputs
-- mp_mul_d didn't preserve sign
-- Many many many many fixes
-- Works in LibTomCrypt now :-)
-- Added iterations to the timing demos... more accurate.
-- Tom needs a job.
Dec 26th, 2002
v0.02 -- Fixed a few "slips" in the manual. This is "LibTomMath" afterall :-)
-- Added mp_cmp_mag, mp_neg, mp_abs and mp_radix_size that were missing.
-- Sped up the fast [comba] multipliers more [yahoo!]
Dec 25th,2002
v0.01 -- Initial release. Gimme a break.
-- Todo list,
add details to manual [e.g. algorithms]
more comments in code
example programs
Jan 15th, 2003
v0.11 -- More subtle fixes
-- Moved to gentoo linux [hurrah!] so made *nix specific fixes to the make process
-- Sped up the montgomery reduction code quite a bit
-- fixed up demo so when building timing for the x86 it assumes ELF format now
Jan 9th, 2003
v0.10 -- Pekka Riikonen suggested fixes to the radix conversion code.
-- Added baseline montgomery and comba montgomery reductions, sped up exptmods
[to a point, see bn.h for MONTGOMERY_EXPT_CUTOFF]
Jan 6th, 2003
v0.09 -- Updated the manual to reflect recent changes. :-)
-- Added Jacobi function (mp_jacobi) to supplement the number theory side of the lib
-- Added a Mersenne prime finder demo in ./etc/mersenne.c
Jan 2nd, 2003
v0.08 -- Sped up the multipliers by moving the inner loop variables into a smaller scope
-- Corrected a bunch of small "warnings"
-- Added more comments
-- Made "mtest" be able to use /dev/random, /dev/urandom or stdin for RNG data
-- Corrected some bugs where error messages were potentially ignored
-- add etc/pprime.c program which makes numbers which are provably prime.
Jan 1st, 2003
v0.07 -- Removed alot of heap operations from core functions to speed them up
-- Added a root finding function [and mp_sqrt macro like from MPI]
-- Added more to manual
Dec 31st, 2002
v0.06 -- Sped up the s_mp_add, s_mp_sub which inturn sped up mp_invmod, mp_exptmod, etc...
-- Cleaned up the header a bit more
Dec 30th, 2002
v0.05 -- Builds with MSVC out of the box
-- Fixed a bug in mp_invmod w.r.t. even moduli
-- Made mp_toradix and mp_read_radix use char instead of unsigned char arrays
-- Fixed up exptmod to use fewer multiplications
-- Fixed up mp_init_size to use only one heap operation
-- Note there is a slight "off-by-one" bug in the library somewhere
without the padding (see the source for comment) the library
crashes in libtomcrypt. Anyways a reasonable workaround is to pad the
numbers which will always correct it since as the numbers grow the padding
will still be beyond the end of the number
-- Added more to the manual
Dec 29th, 2002
v0.04 -- Fixed a memory leak in mp_to_unsigned_bin
-- optimized invmod code
-- Fixed bug in mp_div
-- use exchange instead of copy for results
-- added a bit more to the manual
Dec 27th, 2002
v0.03 -- Sped up s_mp_mul_high_digs by not computing the carries of the lower digits
-- Fixed a bug where mp_set_int wouldn't zero the value first and set the used member.
-- fixed a bug in s_mp_mul_high_digs where the limit placed on the result digits was not calculated properly
-- fixed bugs in add/sub/mul/sqr_mod functions where if the modulus and dest were the same it wouldn't work
-- fixed a bug in mp_mod and mp_mod_d concerning negative inputs
-- mp_mul_d didn't preserve sign
-- Many many many many fixes
-- Works in LibTomCrypt now :-)
-- Added iterations to the timing demos... more accurate.
-- Tom needs a job.
Dec 26th, 2002
v0.02 -- Fixed a few "slips" in the manual. This is "LibTomMath" afterall :-)
-- Added mp_cmp_mag, mp_neg, mp_abs and mp_radix_size that were missing.
-- Sped up the fast [comba] multipliers more [yahoo!]
Dec 25th,2002
v0.01 -- Initial release. Gimme a break.
-- Todo list,
add details to manual [e.g. algorithms]
more comments in code
example programs

72
demo.c
View File

@ -19,8 +19,10 @@
#ifdef TIMER_X86
#define TIMER
extern ulong64 rdtsc(void);
extern void reset(void);
extern ulong64 _rdtsc(void);
extern void _reset(void);
ulong64 rdtsc(void) { return _rdtsc(); }
void reset(void) { _reset(); }
#endif
#ifdef TIMER
@ -85,7 +87,6 @@ int main(void)
mp_int a, b, c, d, e, f;
unsigned long expt_n, add_n, sub_n, mul_n, div_n, sqr_n, mul2d_n, div2d_n, gcd_n, lcm_n, inv_n;
int rr;
mp_digit tom;
#ifdef TIMER
int n;
@ -99,42 +100,33 @@ int main(void)
mp_init(&e);
mp_init(&f);
mp_read_radix(&a, "59994534535345535344389423", 10);
mp_read_radix(&b, "49993453555234234565675534", 10);
mp_read_radix(&c, "62398923474472948723847281", 10);
mp_mulmod(&a, &b, &c, &f);
/* setup mont */
mp_montgomery_setup(&c, &tom);
mp_mul(&a, &b, &a);
mp_montgomery_reduce(&a, &c, tom);
mp_montgomery_reduce(&a, &c, tom);
mp_lshd(&a, c.used*2);
mp_mod(&a, &c, &a);
mp_toradix(&a, cmd, 10);
printf("%s\n\n", cmd);
mp_toradix(&f, cmd, 10);
printf("%s\n", cmd);
/* return 0; */
mp_read_radix(&a, "V//////////////////////////////////////////////////////////////////////////////////////", 64);
mp_reduce_setup(&b, &a);
printf("\n\n----\n\n");
mp_toradix(&b, buf, 10);
printf("b == %s\n\n\n", buf);
mp_read_radix(&b, "4982748972349724892742", 10);
mp_sub_d(&a, 1, &c);
mp_exptmod(&b, &c, &a, &d);
mp_toradix(&d, buf, 10);
printf("b^p-1 == %s\n", buf);
#ifdef DEBUG
mp_read_radix(&a, "347743159439876626079252796797422223177535447388206607607181663903045907591201940478223621722118173270898487582987137708656414344685816179420855160986340457973820182883508387588163122354089264395604796675278966117567294812714812796820596564876450716066283126720010859041484786529056457896367683122960411136319", 10);
mp_read_radix(&b, "347743159439876626079252796797422223177535447388206607607181663903045907591201940478223621722118173270898487582987137708656414344685816179420855160986340457973820182883508387588163122354089264395604796675278966117567294812714812796820596564876450716066283126720010859041484786529056457896367683122960411136318", 10);
mp_set(&c, 1);
reset_timings();
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
mp_exptmod(&c, &b, &a, &d);
dump_timings();
return 0;
#endif
#ifdef TIMER
goto expt;
mp_read_radix(&a, "340282366920938463463374607431768211455", 10);
mp_read_radix(&b, "340282366920938463463574607431768211455", 10);
while (a.used * DIGIT_BIT < 8192) {
@ -182,7 +174,7 @@ int main(void)
printf("Multiplying %d-bit took %llu cycles\n", mp_count_bits(&a), tt / ((ulong64)100000));
mp_copy(&b, &a);
}
expt:
{
char *primes[] = {
"17933601194860113372237070562165128350027320072176844226673287945873370751245439587792371960615073855669274087805055507977323024886880985062002853331424203",
@ -206,7 +198,7 @@ int main(void)
mp_mod(&b, &c, &b);
mp_set(&c, 3);
reset();
for (rr = 0; rr < 35; rr++) {
for (rr = 0; rr < 100; rr++) {
mp_exptmod(&c, &b, &a, &d);
}
tt = rdtsc();
@ -219,7 +211,7 @@ int main(void)
draw(&d);
exit(0);
}
printf("Exponentiating %d-bit took %llu cycles\n", mp_count_bits(&a), tt / ((ulong64)35));
printf("Exponentiating %d-bit took %llu cycles\n", mp_count_bits(&a), tt / ((ulong64)100));
}
}

View File

@ -1,13 +1,13 @@
CC = gcc
CFLAGS += -Wall -W -Wshadow -ansi -O3 -fomit-frame-pointer -funroll-loops
CFLAGS += -Wall -W -Wshadow -ansi -O3 -fomit-frame-pointer -funroll-loops
VERSION=0.10
VERSION=0.11
default: test
test: bn.o demo.o
$(CC) bn.o demo.o -o demo
cd mtest ; gcc $(CFLAGS) mtest.c -o mtest.exe -s
cd mtest ; gcc $(CFLAGS) mtest.c -o mtest -s
# builds the x86 demo
test86:
@ -22,9 +22,9 @@ docs: docdvi
rm -f bn.log bn.aux bn.dvi
clean:
rm -f *.pdf *.o *.exe mtest/*.exe etc/*.exe bn.log bn.aux bn.dvi *.s
rm -f *.pdf *.o *.exe demo mtest/mtest mtest/*.exe etc/*.exe bn.log bn.aux bn.dvi *.log *.s etc/pprime etc/mersenne
zipup: clean docs
chdir .. ; rm -rf ltm* libtommath-$(VERSION) ; mkdir libtommath-$(VERSION) ; \
cd .. ; rm -rf ltm* libtommath-$(VERSION) ; mkdir libtommath-$(VERSION) ; \
cp -R ./libtommath/* ./libtommath-$(VERSION)/ ; tar -c libtommath-$(VERSION)/* > ltm-$(VERSION).tar ; \
bzip2 -9vv ltm-$(VERSION).tar ; zip -9 -r ltm-$(VERSION).zip libtommath-$(VERSION)/*
bzip2 -9vv ltm-$(VERSION).tar ; zip -9 -r ltm-$(VERSION).zip libtommath-$(VERSION)/*

View File

@ -1,34 +1,34 @@
; Simple RDTSC reader for NASM
;
; build with "nasm -f ___ timer.asm" where ___ is coff or elf [or whatever]
;
; Most *nix installs use elf so it would be "nasm -f elf timer.asm"
;
; Tom St Denis
[bits 32]
[section .data]
timer dd 0, 0
[section .text]
[global _gettsc]
_gettsc:
rdtsc
ret
[global _rdtsc]
_rdtsc:
rdtsc
sub eax,[timer]
sbb edx,[timer+4]
ret
[global _reset]
_reset:
push eax
push edx
rdtsc
mov [timer],eax
mov [timer+4],edx
pop edx
pop eax
ret
; Simple RDTSC reader for NASM
;
; build with "nasm -f ___ timer.asm" where ___ is coff or elf [or whatever]
;
; Most *nix installs use elf so it would be "nasm -f elf timer.asm"
;
; Tom St Denis
[bits 32]
[section .data]
timer dd 0, 0
[section .text]
[global _gettsc]
_gettsc:
rdtsc
ret
[global _rdtsc]
_rdtsc:
rdtsc
sub eax,[timer]
sbb edx,[timer+4]
ret
[global _reset]
_reset:
push eax
push edx
rdtsc
mov [timer],eax
mov [timer+4],edx
pop edx
pop eax
ret