From: Kirill Yukhin <kirill.yukhin@gmail.com>
To: Jakub Jelinek <jakub@redhat.com>
Cc: gcc-patches List <gcc-patches@gcc.gnu.org>,
Uros Bizjak <ubizjak@gmail.com>, "H.J. Lu" <hjl.tools@gmail.com>
Subject: Re: [PATCH, i386 tests] New tests to check vectorization for AVX2 insns.
Date: Fri, 14 Oct 2011 14:05:00 -0000 [thread overview]
Message-ID: <CAGs3RfuYWxe6m6wy=CNJeU0VmuE=T_nTyVrssaJRwmmpsUPJdQ@mail.gmail.com> (raw)
In-Reply-To: <20111014115346.GW2210@tyan-ft48-01.lab.bos.redhat.com>
[-- Attachment #1: Type: text/plain, Size: 651 bytes --]
Thanks, done.
Anything else?
K
On Fri, Oct 14, 2011 at 3:53 PM, Jakub Jelinek <jakub@redhat.com> wrote:
> On Fri, Oct 14, 2011 at 03:13:45PM +0400, Kirill Yukhin wrote:
>
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/avx2-vpaddb-3.c
> @@ -0,0 +1,49 @@
> +/* { dg-do run } */
> +/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
> +/* { dg-require-effective-target avx2 } */
> +
> ...
> +
> +/* { dg-final { scan-assembler-times "vpaddb\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
>
> You need
> /* { dg-final { cleanup-saved-temps } } */
> in all the testcases compiled with -save-temps.
>
> Jakub
>
[-- Attachment #2: avx2.vect-2.tests.gcc.patch --]
[-- Type: application/octet-stream, Size: 17295 bytes --]
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpaddb-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpaddb-3.c
new file mode 100644
index 0000000..c99c967
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpaddb-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+char a[SIZE];
+char b[SIZE];
+char c[SIZE];
+volatile char c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_paddb ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] + b[i];
+}
+
+void
+check_paddb ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] + b[i];
+}
+
+void static
+avx2_test (void)
+{
+ int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_paddb ();
+ check_paddb ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (char) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpaddb\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpaddd-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpaddd-3.c
new file mode 100644
index 0000000..ab6b658
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpaddd-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+int a[SIZE];
+int b[SIZE];
+int c[SIZE];
+volatile int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_paddd ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] + b[i];
+}
+
+void
+check_paddd ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] + b[i];
+}
+
+void static
+avx2_test (void)
+{
+ int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_paddd ();
+ check_paddd ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpaddd\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpaddq-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpaddq-3.c
new file mode 100644
index 0000000..066c6fa
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpaddq-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+long long int a[SIZE];
+long long int b[SIZE];
+long long int c[SIZE];
+volatile long long int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_paddq ()
+{
+ long long int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] + b[i];
+}
+
+void
+check_paddq ()
+{
+ long long int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] + b[i];
+}
+
+void static
+avx2_test (void)
+{
+ long long int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_paddq ();
+ check_paddq ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (long long int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpaddq\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpaddw-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpaddw-3.c
new file mode 100644
index 0000000..34ba45e
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpaddw-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+short int a[SIZE];
+short int b[SIZE];
+short int c[SIZE];
+volatile short int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_paddw ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] + b[i];
+}
+
+void
+check_paddw ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] + b[i];
+}
+
+void static
+avx2_test (void)
+{
+ int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_paddw ();
+ check_paddw ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof(short) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpaddw\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpand-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpand-3.c
new file mode 100644
index 0000000..8f1b6c0
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpand-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+unsigned int a[SIZE];
+unsigned int b[SIZE];
+unsigned int c[SIZE];
+volatile unsigned int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_pand ()
+{
+ unsigned int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] & b[i];
+}
+
+void
+check_pand ()
+{
+ unsigned int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] & b[i];
+}
+
+void static
+avx2_test (void)
+{
+ unsigned int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_pand ();
+ check_pand ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (unsigned int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpand\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpmulld-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpmulld-3.c
new file mode 100644
index 0000000..5f656c3
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpmulld-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+int a[SIZE];
+int b[SIZE];
+int c[SIZE];
+volatile int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_pmulld ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] * b[i];
+}
+
+void
+check_pmulld ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] * b[i];
+}
+
+void static
+avx2_test (void)
+{
+ int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_pmulld ();
+ check_pmulld ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpmulld\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpmullw-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpmullw-3.c
new file mode 100644
index 0000000..48ea215
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpmullw-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+short int a[SIZE];
+short int b[SIZE];
+short int c[SIZE];
+volatile short int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_pmulld ()
+{
+ short int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] * b[i];
+}
+
+void
+check_pmulld ()
+{
+ short int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] * b[i];
+}
+
+void static
+avx2_test (void)
+{
+ short int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_pmulld ();
+ check_pmulld ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (short int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpmullw\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpsrad-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpsrad-3.c
new file mode 100644
index 0000000..008c86a
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpsrad-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+int a[SIZE];
+int b[SIZE];
+int c[SIZE];
+volatile int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_psrad ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] >> 5;
+}
+
+void
+check_psrad ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] >> 5;
+}
+
+void static
+avx2_test (void)
+{
+ int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_psrad ();
+ check_psrad ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpsrad\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpsraw-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpsraw-3.c
new file mode 100644
index 0000000..c805c00
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpsraw-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+short int a[SIZE];
+short int b[SIZE];
+short int c[SIZE];
+volatile short int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_psraw ()
+{
+ short int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] >> 5;
+}
+
+void
+check_psraw ()
+{
+ short int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] >> 5;
+}
+
+void static
+avx2_test (void)
+{
+ short int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_psraw ();
+ check_psraw ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (short int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpsraw\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpsrld-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpsrld-3.c
new file mode 100644
index 0000000..09427b2
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpsrld-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+unsigned int a[SIZE];
+unsigned int b[SIZE];
+unsigned int c[SIZE];
+volatile unsigned int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_psrld ()
+{
+ unsigned int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] >> 5;
+}
+
+void
+check_psrld ()
+{
+ unsigned int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] >> 5;
+}
+
+void static
+avx2_test (void)
+{
+ unsigned int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_psrld ();
+ check_psrld ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (unsigned int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpsrld\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpsrlw-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpsrlw-3.c
new file mode 100644
index 0000000..439a784
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpsrlw-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+unsigned short a[SIZE];
+unsigned short b[SIZE];
+unsigned short c[SIZE];
+volatile unsigned short c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_psrld ()
+{
+ unsigned short i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] >> 5;
+}
+
+void
+check_psrld ()
+{
+ unsigned short i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] >> 5;
+}
+
+void static
+avx2_test (void)
+{
+ unsigned short i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_psrld ();
+ check_psrld ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (unsigned short) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpsrlw\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpsubb-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpsubb-3.c
new file mode 100644
index 0000000..6e48971
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpsubb-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+char a[SIZE];
+char b[SIZE];
+char c[SIZE];
+volatile char c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_psubb ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] - b[i];
+}
+
+void
+check_psubb ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] - b[i];
+}
+
+void static
+avx2_test (void)
+{
+ int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_psubb ();
+ check_psubb ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (char) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpsubb\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpsubd-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpsubd-3.c
new file mode 100644
index 0000000..d037d7e
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpsubd-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+int a[SIZE];
+int b[SIZE];
+int c[SIZE];
+volatile int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_psubd ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] - b[i];
+}
+
+void
+check_psubd ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] - b[i];
+}
+
+void static
+avx2_test (void)
+{
+ int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_psubd ();
+ check_psubd ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpsubd\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpsubq-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpsubq-3.c
new file mode 100644
index 0000000..92e1e6f
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpsubq-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+long long int a[SIZE];
+long long int b[SIZE];
+long long int c[SIZE];
+volatile long long int c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_psubq ()
+{
+ long long int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] - b[i];
+}
+
+void
+check_psubq ()
+{
+ long long int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] - b[i];
+}
+
+void static
+avx2_test (void)
+{
+ long long int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_psubq ();
+ check_psubq ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (long long int) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpsubq\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx2-vpsubw-3.c b/gcc/testsuite/gcc.target/i386/avx2-vpsubw-3.c
new file mode 100644
index 0000000..52fc88b
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx2-vpsubw-3.c
@@ -0,0 +1,50 @@
+/* { dg-do run } */
+/* { dg-options "-mavx2 -O2 -ftree-vectorize -save-temps" } */
+/* { dg-require-effective-target avx2 } */
+
+#include "avx2-check.h"
+
+#define SIZE 256
+
+short a[SIZE];
+short b[SIZE];
+short c[SIZE];
+volatile short c_ref[SIZE];
+
+__attribute__ ((__noinline__))
+void
+gen_psubw ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c[i] = a[i] - b[i];
+}
+
+void
+check_psubw ()
+{
+ int i;
+ for (i = 0; i < SIZE; ++i)
+ c_ref[i] = a[i] - b[i];
+}
+
+void static
+avx2_test (void)
+{
+ int i, j;
+ for (i = 0; i < 4; ++i ) {
+ for ( j = 0; j < SIZE; ++j ) {
+ a[i] = i*i+i;
+ b[i] = i*i*i;
+ }
+
+ gen_psubw ();
+ check_psubw ();
+
+ if ( memcmp(c, c_ref, SIZE * sizeof (short) ) )
+ abort();
+ }
+}
+
+/* { dg-final { scan-assembler-times "vpsubw\[ \\t\]+\[^\n\]*%ymm\[0-9\]" 1 } } */
+/* { dg-final { cleanup-saved-temps } } */
next prev parent reply other threads:[~2011-10-14 13:53 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-10-14 11:54 Kirill Yukhin
2011-10-14 12:10 ` Jakub Jelinek
2011-10-14 14:05 ` Kirill Yukhin [this message]
2011-10-14 14:44 ` Jakub Jelinek
2011-10-17 14:39 ` Kirill Yukhin
2011-10-17 14:49 ` Jakub Jelinek
2011-10-17 16:03 ` Kirill Yukhin
2011-10-18 16:49 ` H.J. Lu
2011-10-18 16:54 ` Kirill Yukhin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAGs3RfuYWxe6m6wy=CNJeU0VmuE=T_nTyVrssaJRwmmpsUPJdQ@mail.gmail.com' \
--to=kirill.yukhin@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=hjl.tools@gmail.com \
--cc=jakub@redhat.com \
--cc=ubizjak@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).