re PR libfortran/80602 (Reduce stack usage for blocked matmul)

2017-05-08  Thomas Koenig  <tkoenig@gcc.gnu.org>

	PR fortran/80602
	* m4/matmul_internal.m4:  'matmul_name`:  Change
	t1 to a VLA of the required size.
        * generated/matmul_c10.c: Regenerated.
        * generated/matmul_c16.c: Regenerated.
        * generated/matmul_c4.c: Regenerated.
        * generated/matmul_c8.c: Regenerated.
        * generated/matmul_i1.c: Regenerated.
        * generated/matmul_i16.c: Regenerated.
        * generated/matmul_i2.c: Regenerated.
        * generated/matmul_i4.c: Regenerated.
        * generated/matmul_i8.c: Regenerated.
        * generated/matmul_r10.c: Regenerated.
        * generated/matmul_r16.c: Regenerated.
        * generated/matmul_r4.c: Regenerated.
        * generated/matmul_r8.c: Regenerated.

2017-05-08  Thomas Koenig  <tkoenig@gcc.gnu.org>

	PR fortran/80602
	* gfortran.dg/matmul_15.f90:  New test case.

From-SVN: r247753

1 files changed, 60 insertions, 10 deletions

diff --git a/libgfortran/generated/matmul_i4.c b/libgfortran/generated/matmul_i4.c
index 1e724d54ffa..eab6c773d99 100644
--- a/libgfortran/generated/matmul_i4.c
+++ b/libgfortran/generated/matmul_i4.c
@@ -286,8 +286,7 @@ matmul_i4_avx (gfc_array_i4 * const restrict retarray,
 		 i1, i2, i3, i4, i5, i6;
 
       /* Local variables */
-      GFC_INTEGER_4 t1[65536], /* was [256][256] */
-		 f11, f12, f21, f22, f31, f32, f41, f42,
+      GFC_INTEGER_4 f11, f12, f21, f22, f31, f32, f41, f42,
 		 f13, f14, f23, f24, f33, f34, f43, f44;
       index_type i, j, l, ii, jj, ll;
       index_type isec, jsec, lsec, uisec, ujsec, ulsec;
@@ -311,6 +310,17 @@ matmul_i4_avx (gfc_array_i4 * const restrict retarray,
       if (m == 0 || n == 0 || k == 0)
 	return;
 
+      /* Adjust size of t1 to what is needed.  */
+      index_type t1_dim;
+      t1_dim = (a_dim1-1) * 256 + b_dim1;
+      if (t1_dim > 65536)
+	t1_dim = 65536;
+
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wvla"
+      GFC_INTEGER_4 t1[t1_dim]; /* was [256][256] */
+#pragma GCC diagnostic pop
+
       /* Empty c first.  */
       for (j=1; j<=n; j++)
 	for (i=1; i<=m; i++)
@@ -829,8 +839,7 @@ matmul_i4_avx2 (gfc_array_i4 * const restrict retarray,
 		 i1, i2, i3, i4, i5, i6;
 
       /* Local variables */
-      GFC_INTEGER_4 t1[65536], /* was [256][256] */
-		 f11, f12, f21, f22, f31, f32, f41, f42,
+      GFC_INTEGER_4 f11, f12, f21, f22, f31, f32, f41, f42,
 		 f13, f14, f23, f24, f33, f34, f43, f44;
       index_type i, j, l, ii, jj, ll;
       index_type isec, jsec, lsec, uisec, ujsec, ulsec;
@@ -854,6 +863,17 @@ matmul_i4_avx2 (gfc_array_i4 * const restrict retarray,
       if (m == 0 || n == 0 || k == 0)
 	return;
 
+      /* Adjust size of t1 to what is needed.  */
+      index_type t1_dim;
+      t1_dim = (a_dim1-1) * 256 + b_dim1;
+      if (t1_dim > 65536)
+	t1_dim = 65536;
+
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wvla"
+      GFC_INTEGER_4 t1[t1_dim]; /* was [256][256] */
+#pragma GCC diagnostic pop
+
       /* Empty c first.  */
       for (j=1; j<=n; j++)
 	for (i=1; i<=m; i++)
@@ -1372,8 +1392,7 @@ matmul_i4_avx512f (gfc_array_i4 * const restrict retarray,
 		 i1, i2, i3, i4, i5, i6;
 
       /* Local variables */
-      GFC_INTEGER_4 t1[65536], /* was [256][256] */
-		 f11, f12, f21, f22, f31, f32, f41, f42,
+      GFC_INTEGER_4 f11, f12, f21, f22, f31, f32, f41, f42,
 		 f13, f14, f23, f24, f33, f34, f43, f44;
       index_type i, j, l, ii, jj, ll;
       index_type isec, jsec, lsec, uisec, ujsec, ulsec;
@@ -1397,6 +1416,17 @@ matmul_i4_avx512f (gfc_array_i4 * const restrict retarray,
       if (m == 0 || n == 0 || k == 0)
 	return;
 
+      /* Adjust size of t1 to what is needed.  */
+      index_type t1_dim;
+      t1_dim = (a_dim1-1) * 256 + b_dim1;
+      if (t1_dim > 65536)
+	t1_dim = 65536;
+
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wvla"
+      GFC_INTEGER_4 t1[t1_dim]; /* was [256][256] */
+#pragma GCC diagnostic pop
+
       /* Empty c first.  */
       for (j=1; j<=n; j++)
 	for (i=1; i<=m; i++)
@@ -1911,8 +1941,7 @@ matmul_i4_vanilla (gfc_array_i4 * const restrict retarray,
 		 i1, i2, i3, i4, i5, i6;
 
       /* Local variables */
-      GFC_INTEGER_4 t1[65536], /* was [256][256] */
-		 f11, f12, f21, f22, f31, f32, f41, f42,
+      GFC_INTEGER_4 f11, f12, f21, f22, f31, f32, f41, f42,
 		 f13, f14, f23, f24, f33, f34, f43, f44;
       index_type i, j, l, ii, jj, ll;
       index_type isec, jsec, lsec, uisec, ujsec, ulsec;
@@ -1936,6 +1965,17 @@ matmul_i4_vanilla (gfc_array_i4 * const restrict retarray,
       if (m == 0 || n == 0 || k == 0)
 	return;
 
+      /* Adjust size of t1 to what is needed.  */
+      index_type t1_dim;
+      t1_dim = (a_dim1-1) * 256 + b_dim1;
+      if (t1_dim > 65536)
+	t1_dim = 65536;
+
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wvla"
+      GFC_INTEGER_4 t1[t1_dim]; /* was [256][256] */
+#pragma GCC diagnostic pop
+
       /* Empty c first.  */
       for (j=1; j<=n; j++)
 	for (i=1; i<=m; i++)
@@ -2508,8 +2548,7 @@ matmul_i4 (gfc_array_i4 * const restrict retarray,
 		 i1, i2, i3, i4, i5, i6;
 
       /* Local variables */
-      GFC_INTEGER_4 t1[65536], /* was [256][256] */
-		 f11, f12, f21, f22, f31, f32, f41, f42,
+      GFC_INTEGER_4 f11, f12, f21, f22, f31, f32, f41, f42,
 		 f13, f14, f23, f24, f33, f34, f43, f44;
       index_type i, j, l, ii, jj, ll;
       index_type isec, jsec, lsec, uisec, ujsec, ulsec;
@@ -2533,6 +2572,17 @@ matmul_i4 (gfc_array_i4 * const restrict retarray,
       if (m == 0 || n == 0 || k == 0)
 	return;
 
+      /* Adjust size of t1 to what is needed.  */
+      index_type t1_dim;
+      t1_dim = (a_dim1-1) * 256 + b_dim1;
+      if (t1_dim > 65536)
+	t1_dim = 65536;
+
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wvla"
+      GFC_INTEGER_4 t1[t1_dim]; /* was [256][256] */
+#pragma GCC diagnostic pop
+
       /* Empty c first.  */
       for (j=1; j<=n; j++)
 	for (i=1; i<=m; i++)