Added mulmid and div/divrem/divapprox profiling code.

profile/p-nn_div_divconquer.c

@@ -0,0 +1,114 @@
+/* 
+  Copyright (C) 2013 William Hart
+
+  All rights reserved.
+
+  Redistribution and use in source and binary forms, with or without
+  modification, are permitted provided that the following conditions are met:
+
+  1. Redistributions of source code must retain the above copyright notice, 
+     this list of conditions and the following disclaimer.
+
+  2. Redistributions in binary form must reproduce the above copyright
+     notice, this list of conditions and the following disclaimer in the
+	 documentation and/or other materials provided with the distribution.
+
+  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ``AS IS''
+  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+  DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE
+  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+  DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+  SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+  CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+  OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+  OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <time.h>
+#include <math.h>
+#include "nn.h"
+#include "test.h"
+
+#undef ITER
+#define ITER 10000
+
+rand_t state;
+
+void time_div_divconquer(void)
+{
+   nn_t a, a2, b, q1, q2;
+   len_t size;
+   word_t inv;
+   long count, i;
+   clock_t t;
+
+   TMP_INIT;
+
+   for (size = 370; size < 1000; size = (long) ceil(size*1.1))
+   {
+      TMP_START;
+
+      a = TMP_ALLOC(2*size - 1);
+      a2 = TMP_ALLOC(2*size - 1);
+      b = TMP_ALLOC(size);
+      q1 = TMP_ALLOC(size);
+      q2 = TMP_ALLOC(size);
+
+      printf("size = %ld: ", size);
+
+      t = clock();
+      for (i = 0; i < 10; i++)
+      {
+         randoms_of_len(size, ANY, state, &b, NULL);
+         b[size - 1] |= (1UL<<(WORD_BITS - 1));
+         randoms_of_len(2*size - 1, ANY, state, &a, NULL);
+         inv = precompute_inverse2(b[size - 1], b[size - 2]);
+
+         for (count = 0; count < ITER/10; count++)
+         {
+            nn_copy(a2, a, 2*size - 1);
+            nn_divrem_classical_preinv_c(q1, a2, 2*size - 1, b, size, inv, 0);
+         }
+      }
+      t = clock() - t;
+
+      printf("classical = %gs, ", ((double) t)/CLOCKS_PER_SEC/ITER);
+
+      t = clock();
+      for (i = 0; i < 10; i++)
+      {
+         randoms_of_len(size, ANY, state, &b, NULL);
+         b[size - 1] |= (1UL<<(WORD_BITS - 1));
+         randoms_of_len(2*size - 1, ANY, state, &a, NULL);
+         inv = precompute_inverse2(b[size - 1], b[size - 2]);
+
+         for (count = 0; count < ITER/10; count++)
+         {
+            nn_copy(a2, a, 2*size - 1);
+            nn_div_divconquer_preinv_c(q2, a2, 2*size - 1, b, size, inv, 0);
+         }
+      }
+      t = clock() - t;
+
+      printf("divconquer = %gs\n", ((double) t)/CLOCKS_PER_SEC/ITER);
+
+      TMP_END;
+   }
+}
+
+int main(void)
+{
+   printf("\nTiming nn_div_divconquer vs nn_divrem_classical:\n");
+
+   randinit(&state);
+
+   time_div_divconquer();
+
+   randclear(state);
+
+   return 0;
+}
+

profile/p-nn_divapprox_divconquer.c

@@ -62,9 +62,9 @@ void time_divapprox_divconquer(void)
       t = clock();
       for (i = 0; i < 10; i++)
       {
-         mpn_random(b, size);
-         b[size - 1] |= GMP_LIMB_HIGHBIT;
-         mpn_random(a, 2*size - 1);
+         randoms_of_len(size, ANY, state, &b, NULL);
+         b[size - 1] |= (1UL<<(WORD_BITS - 1));
+         randoms_of_len(2*size - 1, ANY, state, &a, NULL);
          inv = precompute_inverse2(b[size - 1], b[size - 2]);
 
          for (count = 0; count < ITER/10; count++)
@@ -80,9 +80,9 @@ void time_divapprox_divconquer(void)
       t = clock();
       for (i = 0; i < 10; i++)
       {
-         mpn_random(b, size);
-         b[size - 1] |= GMP_LIMB_HIGHBIT;
-         mpn_random(a, 2*size - 1);
+         randoms_of_len(size, ANY, state, &b, NULL);
+         b[size - 1] |= (1UL<<(WORD_BITS - 1));
+         randoms_of_len(2*size - 1, ANY, state, &a, NULL);
          inv = precompute_inverse2(b[size - 1], b[size - 2]);
 
          for (count = 0; count < ITER/10; count++)

profile/p-nn_divrem_divconquer.c

@@ -62,9 +62,9 @@ void time_divrem_divconquer(void)
       t = clock();
       for (i = 0; i < 10; i++)
       {
-         mpn_random(b, size);
+         randoms_of_len(size, ANY, state, &b, NULL);
          b[size - 1] |= (1UL<<(WORD_BITS - 1));
-         mpn_random(a, 2*size - 1);
+         randoms_of_len(2*size - 1, ANY, state, &a, NULL);
          inv = precompute_inverse2(b[size - 1], b[size - 2]);
 
          for (count = 0; count < ITER/10; count++)
@@ -78,9 +78,9 @@ void time_divrem_divconquer(void)
       t += clock();
       for (i = 0; i < 10; i++)
       {
-         mpn_random(b, size);
+         randoms_of_len(size, ANY, state, &b, NULL);
          b[size - 1] |= (1UL<<(WORD_BITS - 1));
-         mpn_random(a, 2*size - 1);
+         randoms_of_len(2*size - 1, ANY, state, &a, NULL);
          inv = precompute_inverse2(b[size - 1], b[size - 2]);
 
          for (count = 0; count < ITER/10; count++)
@@ -97,9 +97,9 @@ void time_divrem_divconquer(void)
          t = clock();
          for (i = 0; i < 10; i++)
          {
-            mpn_random(b, size);
+            randoms_of_len(size, ANY, state, &b, NULL);
             b[size - 1] |= (1UL<<(WORD_BITS - 1));
-            mpn_random(a, 2*size - 1);
+            randoms_of_len(2*size - 1, ANY, state, &a, NULL);
             inv = precompute_inverse2(b[size - 1], b[size - 2]);
 
             for (count = 0; count < ITER/10; count++)

profile/p-nn_mulmid_kara.c

@@ -62,14 +62,14 @@ void time_mulmid_kara(void)
 
       t = clock();
       for (count = 0; count < ITER; count++)
-         mpn_mul(r1, a, 2*size, b, size);
+         nn_mulmid_classical(r1 + size, r1, a, 2*size, b, size);
       t = clock() - t;
 
       printf("classical = %gs, ", ((double) t)/CLOCKS_PER_SEC/ITER);
 
       t = clock();
       for (count = 0; count < ITER; count++)
-         mpn_mulmid(r2, a, 2*size, b, size);
+         nn_mulmid_kara(r2 + size, r2, a, 2*size, b, size);
       t = clock() - t;
 
       printf("kara = %gs\n", ((double) t)/CLOCKS_PER_SEC/ITER);

profile/p-nn_mulmid_vs_mul.c

@@ -0,0 +1,122 @@
+/* 
+  Copyright (C) 2013 William Hart
+
+  All rights reserved.
+
+  Redistribution and use in source and binary forms, with or without
+  modification, are permitted provided that the following conditions are met:
+
+  1. Redistributions of source code must retain the above copyright notice, 
+     this list of conditions and the following disclaimer.
+
+  2. Redistributions in binary form must reproduce the above copyright
+     notice, this list of conditions and the following disclaimer in the
+	 documentation and/or other materials provided with the distribution.
+
+  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ``AS IS''
+  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+  DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE
+  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+  DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+  SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+  CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+  OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+  OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <time.h>
+#include <math.h>
+#include "nn.h"
+#include "test.h"
+
+#undef ITER
+#define ITER 25000
+
+rand_t state;
+
+void time_mulmid_kara(void)
+{
+   nn_t a, b, r1, r2;
+   len_t size;
+   long count;
+   clock_t t;
+
+   TMP_INIT;
+
+   for (size = 2; size < 4000; size = (long) ceil(size*1.1))
+   {
+      TMP_START;
+
+      a = TMP_ALLOC(2*size);
+      b = TMP_ALLOC(size);
+      r1 = TMP_ALLOC(size + 3);
+      r2 = TMP_ALLOC(3*size);
+
+      randoms_of_len(2*size, ANY, state, &a, NULL);
+      randoms_of_len(size, ANY, state, &b, NULL);
+
+      printf("size = %ld: ", size);
+
+      if (size < 100)
+      {
+         t = clock();
+         for (count = 0; count < ITER; count++)
+            nn_mulmid_classical(r1 + size + 1, r1, a, 2*size, b, size);
+         t = clock() - t;
+
+         printf("mulmid_classical = %gs, ", ((double) t)/CLOCKS_PER_SEC/ITER);
+      }
+
+      t = clock();
+      for (count = 0; count < ITER; count++)
+         nn_mulmid_kara(r1 + size + 1, r2, a, 2*size, b, size);
+      t = clock() - t;
+
+      printf("mulmid_kara = %gs\n", ((double) t)/CLOCKS_PER_SEC/ITER);
+
+      printf("size = %ld: ", size);
+
+      if (size < 100)
+      {
+         t = clock();
+         for (count = 0; count < ITER; count++)
+            nn_mul_classical(r2, a, 2*size, b, size);
+         t = clock() - t;
+
+         printf("mul_classical = %gs, ", ((double) t)/CLOCKS_PER_SEC/ITER);
+      }
+
+      t = clock();
+      for (count = 0; count < ITER; count++)
+         nn_mul(r2, a, 2*size, b, size);
+      t = clock() - t;
+
+      printf("mul = %gs, ", ((double) t)/CLOCKS_PER_SEC/ITER);
+
+      t = clock();
+      for (count = 0; count < ITER; count++)
+         nn_mullow(r2 + 2*size, r2, a, 2*size, b, size);
+      t = clock() - t;
+
+      printf("mullow = %gs\n", ((double) t)/CLOCKS_PER_SEC/ITER);
+
+      TMP_END;
+   }
+}
+
+int main(void)
+{
+   printf("\nTiming nn_mulmid_kara vs nn_mulmid_classical:\n");
+
+   randinit(&state);
+
+   time_mulmid_kara();
+
+   randclear(state);
+
+   return 0;
+}
+