Re: [PATCH] Analyze niter for until-wrap condition [PR101145]

[guojiufu via Gcc-patches]

On 2021-07-01 20:35, Richard Biener wrote:
On Thu, 1 Jul 2021, Jiufu Guo wrote:
For code like:
unsigned foo(unsigned val, unsigned start)
{
  unsigned cnt = 0;
  for (unsigned i = start; i > val; ++i)
    cnt++;
  return cnt;
}

The number of iterations should be about UINT_MAX - start.

For

unsigned foo(unsigned val, unsigned start)
{
  unsigned cnt = 0;
  for (unsigned i = start; i >= val; ++i)
    cnt++;
  return cnt;
}

and val == 0 the loop never terminates.  I don't see anywhere
in the patch that you disregard GE_EXPR and I remember
the code handles GE as well as GT?  From a quick look this is
also not covered by a testcase you add - not exactly sure
how it would materialize in a miscompilation.

In number_of_iterations_cond, there is code:
   if (code == GE_EXPR || code == GT_EXPR
        || (code == NE_EXPR && integer_zerop (iv0->step)))
      {
        std::swap (iv0, iv1);
        code = swap_tree_comparison (code);
      }
It converts "GT/GE" (i >= val) to "LT/LE" (val <= i),
and LE (val <= i) is converted to LT (val - 1 < i).
So, the code is added to number_of_iterations_lt.

But, this patch leads mis-compilation for unsigned "i >= val" as
above transforms: converting LE (val <= i) to LT (val - 1 < i)
seems not appropriate (e.g where val=0).
Thanks for pointing out this!!!

I would investigate a way to handle this correctly.
A possible way maybe just to return false for this kind of LE.

Any suggestions?


There is function adjust_cond_for_loop_until_wrap which
handles similar work for const bases.
Like adjust_cond_for_loop_until_wrap, this patch enhance
function number_of_iterations_cond/number_of_iterations_lt
to analyze number of iterations for this kind of loop.

Bootstrap and regtest pass on powerpc64le, is this ok for trunk?

gcc/ChangeLog:

        PR tree-optimization/101145
        * tree-ssa-loop-niter.c
        (number_of_iterations_until_wrap): New function.
        (number_of_iterations_lt): Invoke above function.
        (adjust_cond_for_loop_until_wrap):
        Merge to number_of_iterations_until_wrap.
        (number_of_iterations_cond): Update invokes for
        adjust_cond_for_loop_until_wrap and number_of_iterations_lt.

gcc/testsuite/ChangeLog:

        PR tree-optimization/101145
        * gcc.dg/vect/pr101145.c: New test.
        * gcc.dg/vect/pr101145.inc: New test.
        * gcc.dg/vect/pr101145_1.c: New test.
        * gcc.dg/vect/pr101145_2.c: New test.
        * gcc.dg/vect/pr101145_3.c: New test.
---
 gcc/testsuite/gcc.dg/vect/pr101145.c   | 187 
+++++++++++++++++++++++++
 gcc/testsuite/gcc.dg/vect/pr101145.inc |  63 +++++++++
 gcc/testsuite/gcc.dg/vect/pr101145_1.c |  15 ++
 gcc/testsuite/gcc.dg/vect/pr101145_2.c |  15 ++
 gcc/testsuite/gcc.dg/vect/pr101145_3.c |  15 ++
 gcc/tree-ssa-loop-niter.c              | 150 +++++++++++---------
 6 files changed, 380 insertions(+), 65 deletions(-)
 create mode 100644 gcc/testsuite/gcc.dg/vect/pr101145.c
 create mode 100644 gcc/testsuite/gcc.dg/vect/pr101145.inc
 create mode 100644 gcc/testsuite/gcc.dg/vect/pr101145_1.c
 create mode 100644 gcc/testsuite/gcc.dg/vect/pr101145_2.c
 create mode 100644 gcc/testsuite/gcc.dg/vect/pr101145_3.c

diff --git a/gcc/testsuite/gcc.dg/vect/pr101145.c 
b/gcc/testsuite/gcc.dg/vect/pr101145.c
new file mode 100644
index 00000000000..74031b031cf
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/vect/pr101145.c
@@ -0,0 +1,187 @@
+/* { dg-require-effective-target vect_int } */
+/* { dg-options "-O3 -fdump-tree-vect-details" } */
+#include <limits.h>
+
+unsigned __attribute__ ((noinline))
+foo (int *__restrict__ a, int *__restrict__ b, unsigned l, unsigned 
n)
+{
+  while (n < ++l)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+unsigned __attribute__ ((noinline))
+foo_1 (int *__restrict__ a, int *__restrict__ b, unsigned l, 
unsigned)
+{
+  while (UINT_MAX - 64 < ++l)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+unsigned __attribute__ ((noinline))
+foo_2 (int *__restrict__ a, int *__restrict__ b, unsigned l, unsigned 
n)
+{
+  l = UINT_MAX - 32;
+  while (n < ++l)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+unsigned __attribute__ ((noinline))
+foo_3 (int *__restrict__ a, int *__restrict__ b, unsigned l, unsigned 
n)
+{
+  while (n <= ++l)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+unsigned __attribute__ ((noinline))
+foo_4 (int *__restrict__ a, int *__restrict__ b, unsigned l, unsigned 
n)
+{  // infininate
+  while (0 <= ++l)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+unsigned __attribute__ ((noinline))
+foo_5 (int *__restrict__ a, int *__restrict__ b, unsigned l, unsigned 
n)
+{
+  //no loop
+  l = UINT_MAX;
+  while (n < ++l)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+unsigned __attribute__ ((noinline))
+bar (int *__restrict__ a, int *__restrict__ b, unsigned l, unsigned 
n)
+{
+  while (--l < n)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+unsigned __attribute__ ((noinline))
+bar_1 (int *__restrict__ a, int *__restrict__ b, unsigned l, 
unsigned)
+{
+  while (--l < 64)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+unsigned __attribute__ ((noinline))
+bar_2 (int *__restrict__ a, int *__restrict__ b, unsigned l, unsigned 
n)
+{
+  l = 32;
+  while (--l < n)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+
+int a[3200], b[3200];
+int fail;
+
+int
+main ()
+{
+  unsigned l, n;
+  unsigned res;
+  /* l > n*/
+  n = UINT_MAX - 64;
+  l = n + 32;
+  res = foo (a, b, l, n);
+  if (res != 0)
+    fail++;
+
+  l = n;
+  res = foo (a, b, l, n);
+  if (res != 0)
+    fail++;
+
+  l = n - 1;
+  res = foo (a, b, l, n);
+  if (res != l + 1)
+    fail++;
+
+  l = n - 32;
+  res = foo (a, b, l, n);
+  if (res != l + 1)
+    fail++;
+
+  l = UINT_MAX;
+  res = foo (a, b, l, n);
+  if (res != 0)
+    fail++;
+
+  l = n + 32;
+  res = foo_1 (a, b, l, n);
+  if (res != 0)
+    fail++;
+
+  l = n + 32;
+  res = foo_2 (a, b, l, n);
+  if (res != 0)
+    fail++;
+
+  l = n;
+  res = foo_3 (a, b, l, n);
+  if (res != 0)
+    fail++;
+
+  l = n - 1;
+  res = foo_3 (a, b, l, n);
+  if (res != 0)
+    fail++;
+
+  l = n - 2;
+  res = foo_3 (a, b, l, n);
+  if (res != l + 1)
+    fail++;
+
+  res = foo_5 (a, b, l, n);
+  if (res != 0)
+    fail++;
+
+  n = 64;
+  l = n - 32;
+  res = bar (a, b, l, n);
+  res++;
+  if (res != 0)
+    fail++;
+
+  l = n;
+  res = bar (a, b, l, n);
+  res++;
+  if (res != 0)
+    fail++;
+
+  l = n + 1;
+  res = bar (a, b, l, n);
+  res++;
+  if (res != l)
+    fail++;
+
+  l = 0;
+  res = bar (a, b, l, n);
+  res++;
+  if (res != l)
+    fail++;
+
+  l = 32;
+  res = bar_1 (a, b, l, n);
+  res++;
+  if (res != 0)
+    fail++;
+
+  res = bar_1 (a, b, l, n);
+  res++;
+  if (res != 0)
+    fail++;
+
+  if (fail)
+    __builtin_abort ();
+  return 0;
+}
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 7 "vect" } 
} */
diff --git a/gcc/testsuite/gcc.dg/vect/pr101145.inc 
b/gcc/testsuite/gcc.dg/vect/pr101145.inc
new file mode 100644
index 00000000000..6eed3fa8aca
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/vect/pr101145.inc
@@ -0,0 +1,63 @@
+TYPE __attribute__ ((noinline))
+foo_sign (int *__restrict__ a, int *__restrict__ b, TYPE l, TYPE n)
+{
+  for (l = L_BASE; n < l; l += C)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+TYPE __attribute__ ((noinline))
+bar_sign (int *__restrict__ a, int *__restrict__ b, TYPE l, TYPE n)
+{
+  for (l = L_BASE_DOWN; l < n; l -= C)
+    *a++ = *b++ + 1;
+  return l;
+}
+
+int __attribute__ ((noinline)) neq (int a, int b) { return a != b; }
+
+int a[1000], b[1000];
+int fail;
+
+int
+main ()
+{
+  TYPE res;
+  TYPE l;
+  TYPE n;
+  n = N_BASE;
+  l = n - C;
+  res = foo_sign (a, b, l, n);
+  if (res != l)
+    fail++;
+
+  l = n;
+  res = foo_sign (a, b, l, n);
+  if (res != l)
+    fail++;
+
+  l = n + C;
+  res = foo_sign (a, b, l, n);
+  if (neq ((res - MIN) / C, 0))
+    fail++;
+
+  n = N_BASE_DOWN;
+  l = n - C;
+  res = bar_sign (a, b, l, n);
+  if (neq ((MAX - res) / C, 0))
+    fail++;
+
+  l = n;
+  res = bar_sign (a, b, l, n);
+  if (res != l)
+    fail++;
+
+  l = n + C;
+  res = bar_sign (a, b, l, n);
+  if (res != l)
+    fail++;
+
+  if (fail)
+    __builtin_abort ();
+  return 0;
+}
diff --git a/gcc/testsuite/gcc.dg/vect/pr101145_1.c 
b/gcc/testsuite/gcc.dg/vect/pr101145_1.c
new file mode 100644
index 00000000000..94f6b99b893
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/vect/pr101145_1.c
@@ -0,0 +1,15 @@
+/* { dg-require-effective-target vect_int } */
+/* { dg-options "-O3 -fdump-tree-vect-details" } */
+#define TYPE signed char
+#define MIN -128
+#define MAX 127
+#define N_BASE (MAX - 32)
+#define N_BASE_DOWN (MIN + 32)
+
+#define C 3
+#define L_BASE l
+#define L_BASE_DOWN l
+
+#include "pr101145.inc"
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 2 "vect" } 
} */
diff --git a/gcc/testsuite/gcc.dg/vect/pr101145_2.c 
b/gcc/testsuite/gcc.dg/vect/pr101145_2.c
new file mode 100644
index 00000000000..d3cfc9e01e1
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/vect/pr101145_2.c
@@ -0,0 +1,15 @@
+/* { dg-require-effective-target vect_int } */
+/* { dg-options "-O3 -fdump-tree-vect-details" } */
+#define TYPE unsigned char
+#define MIN 0
+#define MAX 255
+#define N_BASE (MAX - 32 + 1)
+#define N_BASE_DOWN (MIN + 32)
+
+#define C 2
+#define L_BASE l
+#define L_BASE_DOWN l
+
+#include "pr101145.inc"
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 2 "vect" } 
} */
diff --git a/gcc/testsuite/gcc.dg/vect/pr101145_3.c 
b/gcc/testsuite/gcc.dg/vect/pr101145_3.c
new file mode 100644
index 00000000000..e2cf9b1f7e6
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/vect/pr101145_3.c
@@ -0,0 +1,15 @@
+/* { dg-require-effective-target vect_int } */
+/* { dg-options "-O3 -fdump-tree-vect-details" } */
+#define TYPE int *
+#define MIN ((TYPE)0)
+#define MAX ((TYPE)((long long)-1))
+#define N_BASE ((TYPE) (-32))
+#define N_BASE_DOWN (MIN + 32)
+
+#define C 1
+#define L_BASE l
+#define L_BASE_DOWN l
+
+#include "pr101145.inc"
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 2 "vect" } 
} */
diff --git a/gcc/tree-ssa-loop-niter.c b/gcc/tree-ssa-loop-niter.c
index b5add827018..06db6a36ef8 100644
--- a/gcc/tree-ssa-loop-niter.c
+++ b/gcc/tree-ssa-loop-niter.c
@@ -1473,6 +1473,86 @@ assert_loop_rolls_lt (tree type, affine_iv 
*iv0, affine_iv *iv1,
     }
 }

+/* Determines number of iterations of loop whose ending condition
+   is IV0 < IV1 which likes:  {base, -C} < n,  or n < {base, C}.
+   The number of iterations is stored to NITER.  */
+
+static bool
+number_of_iterations_until_wrap (class loop *, tree type, affine_iv 
*iv0,
+                                affine_iv *iv1, class tree_niter_desc *niter)
+{
+  tree niter_type = unsigned_type_for (type);
+  tree max, min;
+
+  if (POINTER_TYPE_P (type))
+    {
+      max = fold_convert (type, TYPE_MAX_VALUE (niter_type));
+      min = fold_convert (type, TYPE_MIN_VALUE (niter_type));
+    }
+  else
+    {
+      max = TYPE_MAX_VALUE (type);
+      min = TYPE_MIN_VALUE (type);
+    }

It would be cheaper to use wide_int here - wi::min/max_value
(TYPE_PRECISION (...), TYPE_SIGN (...)) and in the simplification
code below.

Thanks!  I think you mean to use wide_int for "tree high = max, low = 
min,".
I would check and update the patch to use wide_int.

The MAX/MIN would be in type 'tree' when binding to MINUS_EXPR with 
iv0/iv1->base,
for this, we may still need to use "TYPE_MAX/MIN_VALUE (type)", right?


+  tree high = max, low = min, one = build_int_cst (niter_type, 1);
+  tree step;
+
+  /* n < {base, C}. */
+  if (integer_zerop (iv0->step) && TREE_CODE (iv1->step) == 
INTEGER_CST
+      && !tree_int_cst_sign_bit (iv1->step))
+    {
+      step = iv1->step;
+      niter->niter = fold_build2 (MINUS_EXPR, niter_type, max, 
iv1->base);
+      if (TREE_CODE (iv1->base) == INTEGER_CST)
+       low = fold_build2 (MINUS_EXPR, type, iv1->base, one);
+      else if (TREE_CODE (iv0->base) == INTEGER_CST)
+       low = iv0->base;
+    }
+  /* {base, -C} < n. */
+  else if (TREE_CODE (iv0->step) == INTEGER_CST
+	   && tree_int_cst_sign_bit (iv0->step) && integer_zerop 
(iv1->step))
+    {
+      step = fold_build1 (NEGATE_EXPR, TREE_TYPE (iv0->step), 
iv0->step);
+      niter->niter = fold_build2 (MINUS_EXPR, niter_type, iv0->base, 
min);
+      if (TREE_CODE (iv0->base) == INTEGER_CST)
+       high = fold_build2 (PLUS_EXPR, type, iv0->base, one);
+      else if (TREE_CODE (iv1->base) == INTEGER_CST)
+       high = iv1->base;
+    }
+  else
+    return false;
+
+  /* (delta + step - 1) / step */
+  step = fold_convert (niter_type, step);
+  niter->niter = fold_convert (niter_type, niter->niter);
+  niter->niter = fold_build2 (PLUS_EXPR, niter_type, niter->niter, 
step);
+  niter->niter = fold_build2 (FLOOR_DIV_EXPR, niter_type, 
niter->niter, step);
+
+  tree m = fold_build2 (MINUS_EXPR, niter_type, high, low);
+  m = fold_convert (niter_type, m);
+  mpz_t mstep, tmp, mmax;
+  mpz_init (mstep);
+  mpz_init (tmp);
+  mpz_init (mmax);
+  wi::to_mpz (wi::to_wide (step), mstep, UNSIGNED);
+  wi::to_mpz (wi::to_wide (m), mmax, UNSIGNED);
+  mpz_add (tmp, mmax, mstep);
+  mpz_sub_ui (tmp, tmp, 1);
+  mpz_fdiv_q (tmp, tmp, mstep);
+  niter->max = widest_int::from (wi::from_mpz (niter_type, tmp, 
false),
+                                TYPE_SIGN (niter_type));
+  mpz_clear (mstep);
+  mpz_clear (tmp);

You forget to clear mmax?  I wonder why you need to involve GMP here
Oh, thanks for pointing out this :) need to clear mmax.

at all since this is just add/sub/divide - you could use
widest_int directly I think.  At least the back-and-forth between
trees, wide-int and mpz looks odd.
I once use 'trees' to do the calculation, while I encounter the issue
that the may not hold the precision for some cases:
"mmax" is MAX(1<<64-1); step is 4, then mmax + 4 becomes 3.

I would check to use widest_int (e.g. long[4]) which seems enough to 
hold all types.


BR,
Jiufu Guo.


Richard.

+  niter->may_be_zero
+    = fold_build2 (LE_EXPR, boolean_type_node, iv1->base, iv0->base);
+
+  niter->control.no_overflow = false;
+
+  return true;
+}
+
 /* Determines number of iterations of loop whose ending condition
    is IV0 < IV1.  TYPE is the type of the iv.  The number of
    iterations is stored to NITER.  BNDS bounds the difference
@@ -1501,6 +1581,11 @@ number_of_iterations_lt (class loop *loop, tree 
type, affine_iv *iv0,
       niter->bound = iv0->base;
     }

+  /* {base, -C} < n,  or n < {base, C} */
+  if (tree_int_cst_sign_bit (iv0->step)
+      || (!integer_zerop (iv1->step) && !tree_int_cst_sign_bit 
(iv1->step)))
+    return number_of_iterations_until_wrap (loop, type, iv0, iv1, 
niter);
+
   delta = fold_build2 (MINUS_EXPR, niter_type,
                       fold_convert (niter_type, iv1->base),
                       fold_convert (niter_type, iv0->base));
@@ -1665,62 +1750,6 @@ dump_affine_iv (FILE *file, affine_iv *iv)
     }
 }

-/* Given exit condition IV0 CODE IV1 in TYPE, this function adjusts
-   the condition for loop-until-wrap cases.  For example:
-     (unsigned){8, -1}_loop < 10        => {0, 1} != 9
-     10 < (unsigned){0, max - 7}_loop   => {0, 1} != 8
-   Return true if condition is successfully adjusted.  */
-
-static bool
-adjust_cond_for_loop_until_wrap (tree type, affine_iv *iv0, tree_code 
*code,
-                                affine_iv *iv1)
-{
-  /* Only support simple cases for the moment.  */
-  if (TREE_CODE (iv0->base) != INTEGER_CST
-      || TREE_CODE (iv1->base) != INTEGER_CST)
-    return false;
-
-  tree niter_type = unsigned_type_for (type), high, low;
-  /* Case: i-- < 10.  */
-  if (integer_zerop (iv1->step))
-    {
-      /* TODO: Should handle case in which abs(step) != 1.  */
-      if (!integer_minus_onep (iv0->step))
-       return false;
-      /* Give up on infinite loop.  */
-      if (*code == LE_EXPR
-         && tree_int_cst_equal (iv1->base, TYPE_MAX_VALUE (type)))
-       return false;
-      high = fold_build2 (PLUS_EXPR, niter_type,
-                         fold_convert (niter_type, iv0->base),
-                         build_int_cst (niter_type, 1));
-      low = fold_convert (niter_type, TYPE_MIN_VALUE (type));
-    }
-  else if (integer_zerop (iv0->step))
-    {
-      /* TODO: Should handle case in which abs(step) != 1.  */
-      if (!integer_onep (iv1->step))
-       return false;
-      /* Give up on infinite loop.  */
-      if (*code == LE_EXPR
-         && tree_int_cst_equal (iv0->base, TYPE_MIN_VALUE (type)))
-       return false;
-      high = fold_convert (niter_type, TYPE_MAX_VALUE (type));
-      low = fold_build2 (MINUS_EXPR, niter_type,
-                        fold_convert (niter_type, iv1->base),
-                        build_int_cst (niter_type, 1));
-    }
-  else
-    gcc_unreachable ();
-
-  iv0->base = low;
-  iv0->step = fold_convert (niter_type, integer_one_node);
-  iv1->base = high;
-  iv1->step = build_int_cst (niter_type, 0);
-  *code = NE_EXPR;
-  return true;
-}
-
 /* Determine the number of iterations according to condition (for 
staying
    inside loop) which compares two induction variables using 
comparison
    operator CODE.  The induction variable on left side of the 
comparison
@@ -1855,15 +1884,6 @@ number_of_iterations_cond (class loop *loop,
       return true;
     }

-  /* Handle special case loops: while (i-- < 10) and while (10 < i++) 
by
-     adjusting iv0, iv1 and code.  */
-  if (code != NE_EXPR
-      && (tree_int_cst_sign_bit (iv0->step)
-         || (!integer_zerop (iv1->step)
-             && !tree_int_cst_sign_bit (iv1->step)))
-      && !adjust_cond_for_loop_until_wrap (type, iv0, &code, iv1))
-    return false;
-
   /* OK, now we know we have a senseful loop.  Handle several cases, 
depending
      on what comparison operator is used.  */
   bound_difference (loop, iv1->base, iv0->base, &bnds);