You'll have to compile the program each time, but it does do multiplication of any positive integers exactly in any version of C or C++.

 #define A 45 // first number #define B 315 // second number typedef char buffer[A][B]; main() { printf("%d\n",sizeof(buffer)); } 

If you are OK with a little imprecision, you can use the Monte Carlo method:

#include <stdlib.h> #include <stdio.h> unsigned int mul(unsigned short a, unsigned short b) { const int totalBits = 24; const int total = (1 << totalBits); const int maxNumBits = 10; const int mask = (1 << maxNumBits) - 1; int count = 0, i; unsigned short x, y; for(i = 0; i < total; i++) { x = random() & mask; y = random() & mask; if ((x < a) && (y < b)) count++; } return ((long)count) >> (totalBits - (maxNumBits << 1)); } void main(int argc, char *argv[]) { unsigned short a = atoi(argv[1]); unsigned short b = atoi(argv[2]); printf("%hd * %hd = %d\n", a, b, mul(a, b)); } 

Example:

$ ./mul 300 250 300 * 250 = 74954 

I guess this could be good enough ;)

Since you didn't specify what size of number, I assume that you mean two one-bit numbers.

#include <stdbool.h> bool mul(bool a, bool b) { if (a && b) { return true; } else { return false; } } 

If you want a maximally-efficient implementation, use the following tiny implementation:

m(a,b){return a&b;} 

Note that it still only accepts bits even though the types are implicit ints - it takes less code, and is therefore more efficient. (And yes, it does compile.)

Here's a simple shell script to do it:

curl "http://www.bing.com/search?q=$1%2A$2&go=&qs=n&form=QBLH&pq=$1%2A$2" -s \ | sed -e "s/[<>]/\n/g" \ | grep "^[0-9 *]*=[0-9 ]*$" 

UPDATE: Of course, to do it in C, just wrap it in exec("bash", "-c", ...). (Thanks, AmeliaBR)

Why, let's do a recursive halving search between INT64_MIN and INT64_MAX!

#include <stdio.h> #include <stdint.h> int64_t mul_finder(int32_t a, int32_t b, int64_t low, int64_t high) { int64_t result = (low - (0 - high)) / 2; if (result / a == b && result % a == 0) return result; else return result / a < b ? mul_finder(a, b, result, high) : mul_finder(a, b, low, result); } int64_t mul(int32_t a, int32_t b) { return a == 0 ? 0 : mul_finder(a, b, INT64_MIN, INT64_MAX); } void main(int argc, char* argv[]) { int32_t a, b; sscanf(argv[1], "%d", &a); sscanf(argv[2], "%d", &b); printf("%d * %d = %ld\n", a, b, mul(a, b)); } 

P.S. It will happily sigsegv with some values. ;)

Unfortunately, this only works for integers.

Since addition is disallowed, let's build an increment operator first:

int plusOne(int arg){ int onMask = 1; int offMask = -1; while (arg & onMask){ onMask <<= 1; offMask <<= 1; } return(arg & offMask | onMask); } 

Next, we have to handle the sign. First, find the sign bit:

int signBit = -1; while(signBit << 1) signBit <<=1; 

Then take the sign and magnitude of each argument. To negate a number in a two's complement, invert all bits, and add one.

int signA = a & signBit; if(signA) a = plusOne(-1 ^ a); int signB = b & signBit; if(signB) b = plusOne(-1 ^ b); int signRes = signA ^ signB; 

To multiply two positive integers, we can use the geometrical meaning of multiplication:

// 3x4 // // ooo // ooo // ooo // ooo int res = 0; for(int i = 0; i < a; i = plusOne(i)) for(int j = 1; j < b; j = plusOne(j)) res = plusOne(res); if(signRes) res = plusOne(-1 ^ res); 

trolls:

• Addition is disallowed, but does a++ really count as addition? I bet the teacher intended to allow it.
• Relies on two's complement, but that's an implementation-defined behavior and the target platform wasn't specified.
• Similarly, assumes subtraction and division are disallowed just as well.
• << is actually multiplication by a power of two, so it should technically be disallowed.
• unneccesary diagram is unneccesary. Also, it could have been transposed to save one line.
• repeated shifting of -1 is not the nicest way of finding the sign bit. Even if there was no built-in constant, you could do a logical shift right of -1, then invert all bits.
• XOR -1 is a not the best way to invert all bits.
• The whole charade with sign-magnitude representation is unneccesary. Just cast to unsigned and modular arithmetic will do the rest.
• since the absolute value of MIN_INT (AKA signBit) is negative, this breaks for that value. Luckily, it still works in half the cases, because MIN_INT * [even number] should be zero. Also, plusOne breaks for -1, causing infinite loops anytime the result overflows. plusOne works just fine for any value. Sorry for the confusion.

Works for floating point numbers as well:

float mul(float a, float b){ return std::exp(std::log(a) - std::log(1.0 / b)); } 

Everyone knows that Python is easier to use than C. And Python has functions corresponding to every operator, for cases where you can't use the operator. Which is exactly our problem definition, right? So:

#include <Python.h> void multiply(int a, int b) { PyObject *operator_name, *operator, *mul, *pa, *pb, *args, *result; int result; operator_name = PyString_FromString("operator"); operator = PyImport_Import(operator_name); Py_DECREF(operator_name); mul = PyObject_GetAttrString(operator, "__mul__"); pa = PyLong_FromLong((long)a); pb = PyLong_FromLong((long)b); args = PyTuple_New(2); PyTuple_SetItem(args, 0, pa); PyTuple_SetItem(args, 1, pb); presult = PyObject_CallObject(mul, args); Py_DECREF(args); Py_DECREF(mul); Py_DECREF(operator); result = (int)PyLong_AsLong(presult); Py_DECREF(presult); return result; } int main(int argc, char *argv[]) { int c; Py_Initialize(); c = multiply(2, 3); printf("2 * 3 = %d\n", c); Py_Finalize(); } 

None of the other answers are theoretically sound. As the very first comment on the question says:

Please be more specific about "numbers"

We need to define multiplication, and numbers, before an answer is even possible. Once we do, the problem becomes trivial.

The most popular way to do this in beginning mathematical logic is to build von Neumann ordinals on top of ZF set theory, and then use the Peano axioms.

This translates naturally to C, assuming you have a set type that can contain other sets. It doesn't have to contain anything but sets, which makes it trivial (none of that casting void* nonsense in most set libraries), so I'll leave the implementation as an exercise for the reader.

So, first:

/* The empty set is 0. */ set_t zero() { return set_new(); } /* The successor of n is n U {n}. */ set_t successor(set_t n) { set_t result = set_copy(n); set_t set_of_n = set_new(); set_add(set_of_n, n); set_union(result, set_of_n); set_free(set_of_n); return result; } /* It is an error to call this on 0, which will be reported by running out of memory. */ set_t predecessor(set_t n) { set_t pred = zero(); while (1) { set_t next = successor(pred); if (set_equal(next, n)) { set_free(next); return pred; } set_free(pred); } } set_t add(set_t a, set_t b) { if (set_empty(b)) { /* a + 0 = a */ return a; } else { /* a + successor(b) = successor(a+b) */ set_t pred_b = predecessor(b) set_t pred_ab = add(a, pred_b); set_t result = successor(pred_ab); set_free(pred_b); set_free(pred_ab); return result; } } set_t multiply(set_t a, set_t b) { if (set_empty(b)) { /* a * 0 = 0 */ return b; } else { /* a * successor(b) = a + (a * b) */ set_t pred_b = predecessor(b) set_t pred_ab = mul(a, pred_b); set_t result = successor(pred_ab); set_free(pred_b); set_free(pred_ab); return result; } } 

If you want to expand this to integers, rationals, reals, surreals, etc., you can—with infinite precision (assuming you have infinite memory and CPU), to boot. But as Kroenecker famously said, God made the natural numbers; all else is the work of man, so really, why bother?

unsigned add( unsigned a, unsigned b ) { return (unsigned)&((char*)a)[b]; // ignore compiler warnings // (if pointers are bigger than unsigned). it still works. } unsigned umul( unsigned a, unsigned b ) { unsigned res = 0; while( a != 0 ){ if( a & 1) res = add( res, b ); b <<= 1; a >>= 1; } return res; } int mul( int a, int b ){ return (int)umul( (unsigned)a, (unsigned)b ); } 

If you consider the a[b] hack to be cheating (since it's really an add) then this works instead. But table lookups involve pointer adds too.

See http://en.wikipedia.org/wiki/IBM_1620 - a conputer that actually did addition using lookup tables...

Something satisfying about using a table mechanism to 'speed up' an operation that could actually be done in one instruction.

static unsigned sumtab[17][16]= { { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9,10,11,12,13,14,15}, { 1, 2, 3, 4, 5, 6, 7, 8, 9,10,11,12,13,14,15,16}, { 2, 3, 4, 5, 6, 7, 8, 9,10,11,12,13,14,15,16,17}, { 3, 4, 5, 6, 7, 8, 9,10,11,12,13,14,15,16,17,18}, { 4, 5, 6, 7, 8, 9,10,11,12,13,14,15,16,17,18,19}, { 5, 6, 7, 8, 9,10,11,12,13,14,15,16,17,18,19,20}, { 6, 7, 8, 9,10,11,12,13,14,15,16,17,18,19,20,21}, { 7, 8, 9,10,11,12,13,14,15,16,17,18,19,20,21,22}, { 8, 9,10,11,12,13,14,15,16,17,18,19,20,21,22,23}, { 9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24}, {10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25}, {11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26}, {12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27}, {13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28}, {14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29}, {15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30}, {16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31} }; unsigned add( unsigned a, unsigned b ) { static const int add4hack[8] = {4,8,12,16,20,24,28,32}; unsigned carry = 0; unsigned (*sumtab0)[16] = &sumtab[0]; unsigned (*sumtab1)[16] = &sumtab[1]; unsigned result = 0; int nshift = 0; while( (a|b) != 0 ){ unsigned psum = (carry?sumtab1:sumtab0)[ a & 0xF ][ b & 0xF ]; result = result | ((psum & 0xF)<<nshift); carry = psum >> 4; a = a >> 4 b = b >> 4; nshift= add4hack[nshift>>2]; // add 4 to nshift. } return result; } 

I'm not sure what constitutes "cheating" in these "code troll" posts, but this multiplies 2 arbitrary integers, at run time, with no * or + operator using standard libraries (C99).

#include <math.h> main() { int a = 6; int b = 7; return fma(a,b,0); } 

My troll solution for unsigned int:

#include<stdio.h> unsigned int add(unsigned int x, unsigned int y) { /* An addition of one bit corresponds to the both following logical operations for bit result and carry: r = x xor y xor c_in c_out = (x and y) or (x and c_in) or (y and c_in) However, since we dealing not with bits but words, we have to loop till the carry word is stable */ unsigned int t,c=0; do { t = c; c = (x & y) | (x & c) | (y & c); c <<= 1; } while (c!=t); return x^y^c; } unsigned int mult(unsigned int x,unsigned int y) { /* Paper and pencil method for binary positional notation: multiply a factor by one (=copy) or zero depending on others factor actual digit at position, and shift through each position; adding up */ unsigned int r=0; while (y != 0) { if (y & 1) r = add(r,x); y>>=1; x<<=1; } return r; } int main(int c, char** param) { unsigned int x,y; if (c!=3) { printf("Fuck!\n"); return 1; } sscanf(param[1],"%ud",&x); sscanf(param[2],"%ud",&y); printf("%d\n", mult(x,y)); return 0; } 

There are a lot of good answers here, but it doesn't look like many of them take advantage of the fact that modern computers are really powerful. There are multiple processing units in most CPUs, so why use just one? We can exploit this to get great performance results.

#include <stdio.h> #include <limits.h> #include "omp.h" int mult(int a, int b); void main(){ int first; int second; scanf("%i %i", &first, &second); printf("%i x %i = %i\n", first, second, mult(first,second)); } int mult(int second, int first){ int answer = INT_MAX; omp_set_num_threads(second); #pragma omp parallel for(second = first; second > 0; second--) answer--; return INT_MAX - answer; } 

Here's an example of its usage:

$ ./multiply 5 6 5 x 6 = 30 

The #pragma omp parallel directive makes OpenMP divide each part of the for loop to a different execution unit, so we're multiplying in parallel!

Note that you have to use the -fopenmp flag to tell the compiler to use OpenMP.

Troll parts:

1. Misleading about the use of parallel programming.
2. Doesn't work on negative (or large) numbers.
3. Doesn't actually divide the parts of the for loop - each thread runs the loop.
4. Annoying variable names and variable reuse.
5. There's a subtle race condition on answer--; most of the time, it won't show up, but occasionally it will cause inaccurate results.

Unfortunately, multiplication is a very difficult problem in computer science. The best solution is to use division instead:

#include <stdio.h> #include <limits.h> int multiply(int x, int y) { int a; for (a=INT_MAX; a>1; a--) { if (a/x == y) { return a; } } for (a=-1; a>INT_MIN; a--) { if (a/x == y) { return a; } } return 0; } main (int argc, char **argv) { int a, b; if (argc > 1) a = atoi(argv[1]); else a = 42; if (argc > 2) b = atoi(argv[2]); else b = 13; printf("%d * %d is %d\n", a, b, multiply(a,b)); } 

In real life I usually respond to trolling with knowledge, so here's an answer that doesn't troll at all. It works for all int values as far as I can see.

int multiply (int a, int b) { int r = 0; if (a < 0) { a = -a; b = -b } while (a) { if (a&1) { int x = b; do { int y = x&r; r ^= x; x = y<<1 } while (x); } a>>=1; b<<=1; } return r; } 

This is, to the best of my understanding, very much like how a CPU might actually do integer multiplication. First, we make sure that at least one of the arguments (a) is positive by flipping the sign on both if a is negative (and no, I refuse to count negation as a kind of either addition or multiplication). Then the while (a) loop adds a shifted copy of b to the result for every set bit in a. The do loop implements r += x using and, xor, and shifting in what's clearly a set of half-adders, with the carry bits fed back into x until there are no more (a real CPU would use full adders, which is more efficient, but C doesn't have the operators we need for this, unless you count the + operator).

 int bogomul(int A, int B) { int C = 0; while(C/A != B) { print("Answer isn't: %d", C); C = rand(); } return C; } 

Throwing this into the mix:

#include <stdio.h> #include <stdlib.h> int mul(int a, int b) { asm ("mul %2" : "=a" (a) : "%0" (a), "r" (b) : "cc" ); return a; } int main(int argc, char *argv[]) { int a, b; a = (argc > 1) ? atoi(argv[1]) : 0; b = (argc > 2) ? atoi(argv[2]) : 0; return printf("%d x %d = %d\n", a, b, mul(a, b)) < 1; } 

From info page.

– Introducing something extremely unacceptable or unreasonable in the code that cannot be removed without throwing everything away, rendering the answer utterly useless for the OP.

– […] The intention is to do the homework in a language that the lazy OP might think acceptable, but still frustrate him.

#include <stdio.h> #include <stdlib.h> int mult (int n1, int n2); int add (int n1, int n2 ); int main (int argc, char** argv) { int a,b; a = atoi(argv[1]); b = atoi(argv[2]); printf ("\n%i times %i is %i\n",a,b,mult(a,b)); return 0; } int add (int n1, int n2 ) { return n1 - -n2; } int mult (int n1, int n2) { int sum = 0; char s1='p', s2='p'; if ( n1 == 0 || n2 == 0 ) return 0; if( n1 < 0 ) { s1 = 'n'; n1 = -n1; } if( n2 < 0 ) { s2 = 'n'; n2 = -n2; } for (int i = 1; i <= n2; i = add( i, 1 )) { sum = add(sum, n1); } if ( s1 != s2 ) sum = -sum; return sum; } 

Is it possible to write a C program that multiplies two numbers without using the multiplication and addition operators?

Sure:

void multiply() { printf("6 times 7 is 42\n"); } 

But of course that's cheating; obviously he wants to be able to _supply) two numbers, right?

void multiply(int a, int b) { int answer = 42; if (answer / b != a || answer % b) { printf("The answer is 42, so that's the wrong question.\n"); } else { printf("The answer is 42, but that's probably not the right question anyway.\n"); } } 

There's no arithmetic like pointer arithmetic:

int f(int a, int b) { char x[1][b]; return x[a] - x[0]; } int main(int ac, char **av) { printf("%d\n", f(atoi(av[1]),atoi(av[2]))); return 0; } 

The function f implements multiplication. main simply calls it with two arguments.
Works for negative numbers as well.

C#

I think subtraction and negation are not allowed... Anyway:

int mul(int a, int b) { int t = 0; for (int i = b; i >= 1; i--) t -= -a; return t; } 

C with SSE intrinsics (because everything's better with SIMD):

#include <stdio.h> #include <stdlib.h> #include <xmmintrin.h> static float mul(float a, float b) { float c; __m128 va = _mm_set1_ps(a); __m128 vb = _mm_set1_ps(b); __m128 vc = _mm_mul_ps(va, vb); _mm_store_ss(&c, vc); return c; } int main(int argc, char *argv[]) { if (argc > 2) { float a = atof(argv[1]); float b = atof(argv[2]); float c = mul(a, b); printf("%g * %g = %g\n", a, b, c); } return 0; } 

The big advantage of this implementation is that it can easily be adapted to perform 4 parallel multiplications without * or + if required.

#include <stdio.h> #include <string.h> #include <stdlib.h> #define INF 1000000 char cg[INF]; int main() { int a, b; char bg[INF]; memset(bg, '*', INF); scanf("%d%d", &a, &b); bg[b] = 0; while(a--) strcat(cg, bg); int result; printf("%s%n",cg,&result); printf("%d\n", result); return 0; } 

• work only for multiplication result < 1 000 000
• So far can't get rid out of -- operator, possibly enhancing here
• using %n format specifier in printf to count the number printed characters(I posting this mainly to remind of existence %n in C, instead of %n could of course be strlen etc.)
• Prints a*b characters of '*'

Probably too fast :-(

 unsigned int add(unsigned int a, unsigned int b) { unsigned int carry; for (; b != 0; b = carry << 1) { carry = a & b; a ^= b; } return a; } unsigned int mul(unsigned int a, unsigned int b) { unsigned int prod = 0; for (; b != 0; a <<= 1, b >>= 1) { if (b & 1) prod = add(prod, a); } return prod; } 

This Haskell version only works with nonnegative integers, but it does the multiplication the way children first learn it. I.e., 3x4 is 3 groups of 4 things. In this case, the "things" being counted are notches ('|') on a stick.

mult n m = length . concat . replicate n . replicate m $ '|' 

int multiply(int a, int b) { return sizeof(char[a][b]); } 

This may work in C99, if the weather is right, and your compiler supports undefined nonsense.

Since the OP didn't ask for C, here's one in (Oracle) SQL!

WITH aa AS ( SELECT LEVEL AS lvl FROM dual CONNECT BY LEVEL <= &a ), bb AS ( SELECT LEVEL AS lvl FROM dual CONNECT BY LEVEL <= &b ) SELECT COUNT(*) AS addition FROM (SELECT * FROM aa UNION ALL SELECT * FROM bb); WITH aa AS ( SELECT LEVEL AS lvl FROM dual CONNECT BY LEVEL <= &a ), bb AS ( SELECT LEVEL AS lvl FROM dual CONNECT BY LEVEL <= &b ) SELECT COUNT(*) AS multiplication FROM aa CROSS JOIN bb; 

int add(int a, int b) { return 0 - ((0 - a) - b); } int mul(int a, int b) { int m = 0; for (int count = b; count > 0; m = add(m, a), count = add(count, 0 - 1)) { } return m; } 

May contain traces of UD.

#include <stdio.h> #include <stdlib.h> int main(int argc, char **argv) { int x = atoi(argv[1]); int y = atoi(argv[2]); FILE *f = fopen("m","wb"); char *b = calloc(x, y); if (!f || !b || fwrite(b, x, y, f) != y) { puts("503 multiplication service is down for maintenance"); return EXIT_FAILURE; } printf("%ld\n", ftell(f)); fclose(f); remove("m"); return 0; } 

Test run:

$ ./a.out 1 0 0 $ ./a.out 1 1 1 $ ./a.out 2 2 4 $ ./a.out 3 2 6 $ ./a.out 12 12 144 $ ./a.out 1234 1234 1522756