benchtests: Switch string benchmarks to use bench-timing.h.

Switch the string benchmarks to using bench-timing.h instead
of hp-timing.h directly. This allows the string benchmarks to
be run usefully on architectures such as ARM that do not have
support for hp-timing.h.

In order to do this the tests have been changed from timing each
individual call and picking the lowest execution time recorded to
timing a number of calls and taking the mean execution time.

ChangeLog:

2013-09-04   Will Newton  <will.newton@linaro.org>

	* benchtests/bench-timing.h (TIMING_PRINT_MEAN): New macro.
	* benchtests/bench-string.h: Include bench-timing.h instead
	of including hp-timing.h directly. (INNER_LOOP_ITERS): New
	define. (HP_TIMING_BEST): Delete macro. (test_init): Remove
	call to HP_TIMING_DIFF_INIT.
	* benchtests/bench-memccpy.c: Use bench-timing.h macros
	instead of hp-timing.h macros.
	* benchtests/bench-memchr.c: Likewise.
	* benchtests/bench-memcmp.c: Likewise.
	* benchtests/bench-memcpy.c: Likewise.
	* benchtests/bench-memmem.c: Likewise.
	* benchtests/bench-memmove.c: Likewise.
	* benchtests/bench-memset.c: Likewise.
	* benchtests/bench-rawmemchr.c: Likewise.
	* benchtests/bench-strcasecmp.c: Likewise.
	* benchtests/bench-strcasestr.c: Likewise.
	* benchtests/bench-strcat.c: Likewise.
	* benchtests/bench-strchr.c: Likewise.
	* benchtests/bench-strcmp.c: Likewise.
	* benchtests/bench-strcpy.c: Likewise.
	* benchtests/bench-strcpy_chk.c: Likewise.
	* benchtests/bench-strlen.c: Likewise.
	* benchtests/bench-strncasecmp.c: Likewise.
	* benchtests/bench-strncat.c: Likewise.
	* benchtests/bench-strncmp.c: Likewise.
	* benchtests/bench-strncpy.c: Likewise.
	* benchtests/bench-strnlen.c: Likewise.
	* benchtests/bench-strpbrk.c: Likewise.
	* benchtests/bench-strrchr.c: Likewise.
	* benchtests/bench-strspn.c: Likewise.
	* benchtests/bench-strstr.c: Likewise.

27 files changed, 334 insertions, 502 deletions

diff --git a/benchtests/bench-memccpy.c b/benchtests/bench-memccpy.c
index 612513c68d..2c47e79376 100644
--- a/benchtests/bench-memccpy.c
+++ b/benchtests/bench-memccpy.c
@@ -59,6 +59,9 @@ do_one_test (impl_t *impl, void *dst, const void *src, int c, size_t len,
 	     size_t n)
 {
   void *expect = len > n ? NULL : (char *) dst + len;
+  size_t i, iters = INNER_LOOP_ITERS;
+  timing_t start, stop, cur;
+
   if (CALL (impl, dst, src, c, n) != expect)
     {
       error (0, 0, "Wrong result in function %s %p %p", impl->name,
@@ -74,23 +77,16 @@ do_one_test (impl_t *impl, void *dst, const void *src, int c, size_t len,
       return;
     }
 
-  if (HP_TIMING_AVAIL)
+  TIMING_NOW (start);
+  for (i = 0; i < iters; ++i)
     {
-      hp_timing_t start __attribute__ ((unused));
-      hp_timing_t stop __attribute__ ((unused));
-      hp_timing_t best_time = ~ (hp_timing_t) 0;
-      size_t i;
-
-      for (i = 0; i < 32; ++i)
-	{
-	  HP_TIMING_NOW (start);
-	  CALL (impl, dst, src, c, n);
-	  HP_TIMING_NOW (stop);
-	  HP_TIMING_BEST (best_time, start, stop);
-	}
-
-      printf ("\t%zd", (size_t) best_time);
+      CALL (impl, dst, src, c, n);
     }
+  TIMING_NOW (stop);
+
+  TIMING_DIFF (cur, start, stop);
+
+  TIMING_PRINT_MEAN ((double) cur, (double) iters);
 }
 
 static void
@@ -121,14 +117,12 @@ do_test (size_t align1, size_t align2, int c, size_t len, size_t n,
   for (i = len; i + align1 < page_size && i < len + 64; ++i)
     s1[i] = 32 + 32 * i % (max_char - 32);
 
-  if (HP_TIMING_AVAIL)
-    printf ("Length %4zd, n %4zd, char %d, alignment %2zd/%2zd:", len, n, c, align1, align2);
+  printf ("Length %4zd, n %4zd, char %d, alignment %2zd/%2zd:", len, n, c, align1, align2);
 
   FOR_EACH_IMPL (impl, 0)
     do_one_test (impl, s2, s1, c, len, n);
 
-  if (HP_TIMING_AVAIL)
-    putchar ('\n');
+  putchar ('\n');
 }
 
 int
diff --git a/benchtests/bench-memchr.c b/benchtests/bench-memchr.c
index 5470ce6c5b..db099ad543 100644
--- a/benchtests/bench-memchr.c
+++ b/benchtests/bench-memchr.c
@@ -39,6 +39,9 @@ static void
 do_one_test (impl_t *impl, const char *s, int c, size_t n, char *exp_res)
 {
   char *res = CALL (impl, s, c, n);
+  size_t i, iters = INNER_LOOP_ITERS;
+  timing_t start, stop, cur;
+
   if (res != exp_res)
     {
       error (0, 0, "Wrong result in function %s %p %p", impl->name,
@@ -47,23 +50,16 @@ do_one_test (impl_t *impl, const char *s, int c, size_t n, char *exp_res)
       return;
     }
 
-  if (HP_TIMING_AVAIL)
+  TIMING_NOW (start);
+  for (i = 0; i < iters; ++i)
     {
-      hp_timing_t start __attribute ((unused));
-      hp_timing_t stop __attribute ((unused));
-      hp_timing_t best_time = ~ (hp_timing_t) 0;
-      size_t i;
-
-      for (i = 0; i < 32; ++i)
-	{
-	  HP_TIMING_NOW (start);
-	  CALL (impl, s, c, n);
-	  HP_TIMING_NOW (stop);
-	  HP_TIMING_BEST (best_time, start, stop);
-	}
-
-      printf ("\t%zd", (size_t) best_time);
+      CALL (impl, s, c, n);
     }
+  TIMING_NOW (stop);
+
+  TIMING_DIFF (cur, start, stop);
+
+  TIMING_PRINT_MEAN ((double) cur, (double) iters);
 }
 
 static void
@@ -96,14 +92,12 @@ do_test (size_t align, size_t pos, size_t len, int seek_char)
       buf1[align + len] = seek_char;
     }
 
-  if (HP_TIMING_AVAIL)
-    printf ("Length %4zd, alignment %2zd:", pos, align);
+  printf ("Length %4zd, alignment %2zd:", pos, align);
 
   FOR_EACH_IMPL (impl, 0)
     do_one_test (impl, (char *) (buf1 + align), seek_char, len, result);
 
-  if (HP_TIMING_AVAIL)
-    putchar ('\n');
+  putchar ('\n');
 }
 
 int
diff --git a/benchtests/bench-memcmp.c b/benchtests/bench-memcmp.c
index 053bb50567..544130b01d 100644
--- a/benchtests/bench-memcmp.c
+++ b/benchtests/bench-memcmp.c
@@ -78,23 +78,19 @@ static void
 do_one_test (impl_t *impl, const CHAR *s1, const CHAR *s2, size_t len,
 	     int exp_result)
 {
-  if (HP_TIMING_AVAIL)
+  size_t i, iters = INNER_LOOP_ITERS;
+  timing_t start, stop, cur;
+
+  TIMING_NOW (start);
+  for (i = 0; i < iters; ++i)
     {
-      hp_timing_t start __attribute ((unused));
-      hp_timing_t stop __attribute ((unused));
-      hp_timing_t best_time = ~ (hp_timing_t) 0;
-      size_t i;
-
-      for (i = 0; i < 32; ++i)
-	{
-	  HP_TIMING_NOW (start);
-	  CALL (impl, s1, s2, len);
-	  HP_TIMING_NOW (stop);
-	  HP_TIMING_BEST (best_time, start, stop);
-	}
-
-      printf ("\t%zd", (size_t) best_time);
+      CALL (impl, s1, s2, len);
     }
+  TIMING_NOW (stop);
+
+  TIMING_DIFF (cur, start, stop);
+
+  TIMING_PRINT_MEAN ((double) cur, (double) iters);
 }
 
 static void
@@ -124,14 +120,12 @@ do_test (size_t align1, size_t align2, size_t len, int exp_result)
   s2[len] = align2;
   s2[len - 1] -= exp_result;
 
-  if (HP_TIMING_AVAIL)
-    printf ("Length %4zd, alignment %2zd/%2zd:", len, align1, align2);
+  printf ("Length %4zd, alignment %2zd/%2zd:", len, align1, align2);
 
   FOR_EACH_IMPL (impl, 0)
     do_one_test (impl, s1, s2, len, exp_result);
 
-  if (HP_TIMING_AVAIL)
-    putchar ('\n');
+  putchar ('\n');
 }
 
 int
diff --git a/benchtests/bench-memcpy.c b/benchtests/bench-memcpy.c
index 1b126711f8..8cd9c23b44 100644
--- a/benchtests/bench-memcpy.c
+++ b/benchtests/bench-memcpy.c
@@ -52,6 +52,9 @@ static void
 do_one_test (impl_t *impl, char *dst, const char *src,
 	     size_t len)
 {
+  size_t i, iters = INNER_LOOP_ITERS;
+  timing_t start, stop, cur;
+
   if (CALL (impl, dst, src, len) != MEMCPY_RESULT (dst, len))
     {
       error (0, 0, "Wrong result in function %s %p %p", impl->name,
@@ -68,23 +71,16 @@ do_one_test (impl_t *impl, char *dst, const char *src,
       return;
     }
 
-  if (HP_TIMING_AVAIL)
+  TIMING_NOW (start);
+  for (i = 0; i < iters; ++i)
     {
-      hp_timing_t start __attribute ((unused));
-      hp_timing_t stop __attribute ((unused));
-      hp_timing_t best_time = ~ (hp_timing_t) 0;
-      size_t i;
-
-      for (i = 0; i < 32; ++i)
-	{
-	  HP_TIMING_NOW (start);
-	  CALL (impl, dst, src, len);
-	  HP_TIMING_NOW (stop);
-	  HP_TIMING_BEST (best_time, start, stop);
-	}
-
-      printf ("\t%zd", (size_t) best_time);
+      CALL (impl, dst, src, len);
     }
+  TIMING_NOW (stop);
+
+  TIMING_DIFF (cur, start, stop);
+
+  TIMING_PRINT_MEAN ((double) cur, (double) iters);
 }
 
 static void
@@ -107,14 +103,12 @@ do_test (size_t align1, size_t align2, size_t len)
   for (i = 0, j = 1; i < len; i++, j += 23)
     s1[i] = j;
 
-  if (HP_TIMING_AVAIL)
-    printf ("Length %4zd, alignment %2zd/%2zd:", len, align1, align2);
+  printf ("Length %4zd, alignment %2zd/%2zd:", len, align1, align2);
 
   FOR_EACH_IMPL (impl, 0)
     do_one_test (impl, s2, s1, len);
 
-  if (HP_TIMING_AVAIL)
-    putchar ('\n');
+  putchar ('\n');
 }
 
 int
diff --git a/benchtests/bench-memmem.c b/benchtests/bench-memmem.c
index ca758a84f2..b8f8a8b2a3 100644
--- a/benchtests/bench-memmem.c
+++ b/benchtests/bench-memmem.c
@@ -60,23 +60,19 @@ static void
 do_one_test (impl_t *impl, const void *haystack, size_t haystack_len,
 	     const void *needle, size_t needle_len, const void *expected)
 {
-  if (HP_TIMING_AVAIL)
+  size_t i, iters = INNER_LOOP_ITERS;
+  timing_t start, stop, cur;
+
+  TIMING_NOW (start);
+  for (i = 0; i < iters; ++i)
     {
-      hp_timing_t start __attribute ((unused));
-      hp_timing_t stop __attribute ((unused));
-      hp_timing_t best_time = ~ (hp_timing_t) 0;
-      size_t i;
+      CALL (impl, haystack, haystack_len, needle, needle_len);
+    }
+  TIMING_NOW (stop);
 
-      for (i = 0; i < 32; ++i)
-	{
-	  HP_TIMING_NOW (start);
-	  CALL (impl, haystack, haystack_len, needle, needle_len);
-	  HP_TIMING_NOW (stop);
-	  HP_TIMING_BEST (best_time, start, stop);
-	}
+  TIMING_DIFF (cur, start, stop);
 
-      printf ("\t%zd", (size_t) best_time);
-    }
+  TIMING_PRINT_MEAN ((double) cur, (double) iters);
 }
 
 static void
@@ -87,16 +83,14 @@ do_test (const char *str, size_t len, size_t idx)
   memcpy (tmpbuf, buf1 + idx, len);
   memcpy (buf1 + idx, str, len);
 
-  if (HP_TIMING_AVAIL)
-    printf ("String %s, offset %zd:", str, idx);
+  printf ("String %s, offset %zd:", str, idx);
 
   FOR_EACH_IMPL (impl, 0)
     do_one_test (impl, buf1, BUF1PAGES * page_size, str, len, buf1 + idx);
 
   memcpy (buf1 + idx, tmpbuf, len);
 
-  if (HP_TIMING_AVAIL)
-    putchar ('\n');
+  putchar ('\n');
 }
 
 static void
@@ -120,15 +114,13 @@ do_random_tests (void)
 	  buf1[idx + off] = ch;
 	}
 
-      if (HP_TIMING_AVAIL)
-	printf ("String %.*s, offset %zd:", (int) len, buf1 + idx, idx);
+      printf ("String %.*s, offset %zd:", (int) len, buf1 + idx, idx);
 
       FOR_EACH_IMPL (impl, 0)
 	do_one_test (impl, buf1, BUF1PAGES * page_size, buf1 + idx, len,
 		     buf1 + idx);
 
-      if (HP_TIMING_AVAIL)
-	putchar ('\n');
+      putchar ('\n');
 
       memcpy (buf1 + idx, tmpbuf, len);
     }
diff --git a/benchtests/bench-memmove.c b/benchtests/bench-memmove.c
index 8925606af8..332d6af263 100644
--- a/benchtests/bench-memmove.c
+++ b/benchtests/bench-memmove.c
@@ -67,6 +67,9 @@ static void
 do_one_test (impl_t *impl, char *dst, char *src, const char *orig_src,
 	     size_t len)
 {
+  size_t i, iters = INNER_LOOP_ITERS;
+  timing_t start, stop, cur;
+
   memcpy (src, orig_src, len);
 #ifdef TEST_BCOPY
   CALL (impl, src, dst, len);
@@ -91,27 +94,20 @@ do_one_test (impl_t *impl, char *dst, char *src, const char *orig_src,
       return;
     }
 
-  if (HP_TIMING_AVAIL)
+  TIMING_NOW (start);
+  for (i = 0; i < iters; ++i)
     {
-      hp_timing_t start __attribute ((unused));
-      hp_timing_t stop __attribute ((unused));
-      hp_timing_t best_time = ~ (hp_timing_t) 0;
-      size_t i;
-
-      for (i = 0; i < 32; ++i)
-	{
-	  HP_TIMING_NOW (start);
 #ifdef TEST_BCOPY
-	  CALL (impl, src, dst, len);
+      CALL (impl, src, dst, len);
 #else
-	  CALL (impl, dst, src, len);
+      CALL (impl, dst, src, len);
 #endif
-	  HP_TIMING_NOW (stop);
-	  HP_TIMING_BEST (best_time, start, stop);
-	}
-
-      printf ("\t%zd", (size_t) best_time);
     }
+  TIMING_NOW (stop);
+
+  TIMING_DIFF (cur, start, stop);
+
+  TIMING_PRINT_MEAN ((double) cur, (double) iters);
 }
 
 static void
@@ -134,14 +130,12 @@ do_test (size_t align1, size_t align2, size_t len)
   for (i = 0, j = 1; i < len; i++, j += 23)
     s1[i] = j;
 
-  if (HP_TIMING_AVAIL)
-    printf ("Length %4zd, alignment %2zd/%2zd:", len, align1, align2);
+  printf ("Length %4zd, alignment %2zd/%2zd:", len, align1, align2);
 
   FOR_EACH_IMPL (impl, 0)
     do_one_test (impl, s2, (char *) (buf2 + align1), s1, len);
 
-  if (HP_TIMING_AVAIL)
-    putchar ('\n');
+  putchar ('\n');
 }
 
 int
diff --git a/benchtests/bench-memset.c b/benchtests/bench-memset.c
index ea29cf37fc..e45807c0ee 100644
--- a/benchtests/bench-memset.c
+++ b/benchtests/bench-memset.c
@@ -75,6 +75,8 @@ simple_memset (char *s, int c, size_t n)
 static void
 do_one_test (impl_t *impl, char *s, int c __attribute ((unused)), size_t n)
 {
+  size_t i, iters = INNER_LOOP_ITERS;
+  timing_t start, stop, cur;
   char tstbuf[n];
 #ifdef TEST_BZERO
   simple_bzero (tstbuf, n);
@@ -92,28 +94,20 @@ do_one_test (impl_t *impl, char *s, int c __attribute ((unused)), size_t n)
       return;
     }
 
-  if (HP_TIMING_AVAIL)
+  TIMING_NOW (start);
+  for (i = 0; i < iters; ++i)
     {
-      hp_timing_t start __attribute ((unused));
-      hp_timing_t stop __attribute ((unused));
-      hp_timing_t best_time = ~ (hp_timing_t) 0;
-      size_t i;
-
-      for (i = 0; i < 32; ++i)
-	{
-	  HP_TIMING_NOW (start);
 #ifdef TEST_BZERO
-	  CALL (impl, s, n);
+      CALL (impl, s, n);
 #else
-	  CALL (impl, s, c, n);
+      CALL (impl, s, c, n);
 #endif
+    }
+  TIMING_NOW (stop);
 
-	  HP_TIMING_NOW (stop);
-	  HP_TIMING_BEST (best_time, start, stop);
-	}
+  TIMING_DIFF (cur, start, stop);
 
-      printf ("\t%zd", (size_t) best_time);
-    }
+  TIMING_PRINT_MEAN ((double) cur, (double) iters);
 }
 
 static void
@@ -123,14 +117,12 @@ do_test (size_t align, int c, size_t len)
   if (align + len > page_size)
     return;
 
-  if (HP_TIMING_AVAIL)
-    printf ("Length %4zd, alignment %2zd, c %2d:", len, align, c);
+  printf ("Length %4zd, alignment %2zd, c %2d:", len, align, c);
 
   FOR_EACH_IMPL (impl, 0)
     do_one_test (impl, (char *) buf1 + align, c, len);
 
-  if (HP_TIMING_AVAIL)
-    putchar ('\n');
+  putchar ('\n');
 }
 
 int
diff --git a/benchtests/bench-rawmemchr.c b/benchtests/bench-rawmemchr.c
index a6b29d7ae3..df6a310e2b 100644
--- a/benchtests/bench-rawmemchr.c
+++ b/benchtests/bench-rawmemchr.c
@@ -40,6 +40,8 @@ simple_rawmemchr (const char *s, int c)
 static void
 do_one_test (impl_t *impl, const char *s, int c, char *exp_res)
 {
+  size_t i, iters = INNER_LOOP_ITERS;
+  timing_t start, stop, cur;
   char *res = CALL (impl, s, c);
   if (res != exp_res)
     {
@@ -49,23 +51,16 @@ do_one_test (impl_t *impl, const char *s, int c, char *exp_res)
       return;
     }
 
-  if (HP_TIMING_AVAIL)
+  TIMING_NOW (start);
+  for (i = 0; i < iters; ++i)
     {
-      hp_timing_t start __attribute ((unused));
-      hp_timing_t stop __attribute ((unused));
-      hp_timing_t best_time = ~ (hp_timing_t) 0;
-      size_t i;
-
-      for (i = 0; i < 32; ++i)
-	{
-	  HP_TIMING_NOW (start);
-	  CALL (impl, s, c);
-	  HP_TIMING_NOW (stop);
-	  HP_TIMING_BEST (best_time, start, stop);
-	}
-
-      printf ("\t%zd", (size_t) best_time);
+      CALL (impl, s, c);
     }
+  TIMING_NOW (stop);
+
+  TIMING_DIFF (cur, start, stop);
+
+  TIMING_PRINT_MEAN ((double) cur, (double) iters);
 }
 
 static void
@@ -92,14 +87,12 @@ do_test (size_t align, size_t pos, size_t len, int seek_char)
   buf1[align + len] = -seek_char;
   result = (char *) (buf1 + align + pos);
 
-  if (HP_TIMING_AVAIL)
-    printf ("Length %4zd, alignment %2zd:", pos, align);
+  printf ("Length %4zd, alignment %2zd:", pos, align);
 
   FOR_EACH_IMPL (impl, 0)
     do_one_test (impl, (char *) (buf1 + align), seek_char, result);
 
-  if (HP_TIMING_AVAIL)
-    putchar ('\n');