> -----Original Message-----
> From: Zhenqiang Chen [mailto:zhenqiang.c...@linaro.org]
> Sent: 11 April 2013 02:33
> To: Marcus Shawcroft
> Cc: James Greenhalgh; gcc-patches@gcc.gnu.org; Marcus Shawcroft;
> pins...@gmail.com; Richard Earnshaw
> Subject: Re: [PATCH, AARCH64] Fix unrecognizable insn issue
>
> On 10 April 2013 18:48, Marcus Shawcroft <marcus.shawcr...@gmail.com>
> wrote:
> > Zhenqiang, Does Jame's patch fix your test case?
>
> Thank you all. The patch fixes my test case.
I'm pleased to hear it.
Marcus/Richard, I've committed the attached as r197741. Could you
OK a backport to the 4.8 release branch? (The backport is clean,
and regresses without issue on aarch64-none-elf). I'll also put it
on arm/aarch64-4.7-branch.
Thanks,
James Greenhalgh
---
gcc/
2013-04-11 James Greenhalgh <james.greenha...@arm.com>
* config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
floating-point vector comparisons against 0.
gcc/testsuite/
2013-04-11 James Greenhalgh <james.greenha...@arm.com>
* gcc.target/aarch64/vect-fcm.x: Add check for zero forms of
inverse operands.
* gcc.target/aarch64/vect-fcm-eq-d.c: Check that new zero form
loop is vectorized.
* gcc.target/aarch64/vect-fcm-eq-f.c: Likewise.
* gcc.target/aarch64/vect-fcm-ge-d.c: Check that new zero form
loop is vectorized and that the correct instruction is generated.
* gcc.target/aarch64/vect-fcm-ge-f.c: Likewise.
* gcc.target/aarch64/vect-fcm-gt-d.c: Likewise.
* gcc.target/aarch64/vect-fcm-gt-f.c: Likewise.
diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
index 0f00095..79c3093 100644
--- a/gcc/config/aarch64/aarch64-simd.md
+++ b/gcc/config/aarch64/aarch64-simd.md
@@ -1586,6 +1586,7 @@
"TARGET_SIMD"
{
int inverse = 0;
+ int use_zero_form = 0;
int swap_bsl_operands = 0;
rtx mask = gen_reg_rtx (<V_cmp_result>mode);
rtx tmp = gen_reg_rtx (<V_cmp_result>mode);
@@ -1596,12 +1597,16 @@
switch (GET_CODE (operands[3]))
{
case GE:
+ case GT:
case LE:
+ case LT:
case EQ:
- if (!REG_P (operands[5])
- && (operands[5] != CONST0_RTX (<MODE>mode)))
- operands[5] = force_reg (<MODE>mode, operands[5]);
- break;
+ if (operands[5] == CONST0_RTX (<MODE>mode))
+ {
+ use_zero_form = 1;
+ break;
+ }
+ /* Fall through. */
default:
if (!REG_P (operands[5]))
operands[5] = force_reg (<MODE>mode, operands[5]);
@@ -1652,7 +1657,26 @@
a GT b -> a GT b
a LE b -> b GE a
a LT b -> b GT a
- a EQ b -> a EQ b */
+ a EQ b -> a EQ b
+ Note that there also exist direct comparison against 0 forms,
+ so catch those as a special case. */
+ if (use_zero_form)
+ {
+ inverse = 0;
+ switch (GET_CODE (operands[3]))
+ {
+ case LT:
+ base_comparison = gen_aarch64_cmlt<mode>;
+ break;
+ case LE:
+ base_comparison = gen_aarch64_cmle<mode>;
+ break;
+ default:
+ /* Do nothing, other zero form cases already have the correct
+ base_comparison. */
+ break;
+ }
+ }
if (!inverse)
emit_insn (base_comparison (mask, operands[4], operands[5]));
diff --git a/gcc/testsuite/gcc.target/aarch64/vect-fcm-eq-d.c b/gcc/testsuite/gcc.target/aarch64/vect-fcm-eq-d.c
index b6fb5ae..19ecd63 100644
--- a/gcc/testsuite/gcc.target/aarch64/vect-fcm-eq-d.c
+++ b/gcc/testsuite/gcc.target/aarch64/vect-fcm-eq-d.c
@@ -7,7 +7,7 @@
#include "vect-fcm.x"
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 3 "vect" } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 4 "vect" } } */
/* { dg-final { scan-assembler "fcmeq\\tv\[0-9\]+\.2d, v\[0-9\]+\.2d, v\[0-9\]+\.2d" } } */
/* { dg-final { scan-assembler "fcmeq\\tv\[0-9\]+\.2d, v\[0-9\]+\.2d, 0" } } */
/* { dg-final { cleanup-tree-dump "vect" } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/vect-fcm-eq-f.c b/gcc/testsuite/gcc.target/aarch64/vect-fcm-eq-f.c
index 283d34f..30be5ad 100644
--- a/gcc/testsuite/gcc.target/aarch64/vect-fcm-eq-f.c
+++ b/gcc/testsuite/gcc.target/aarch64/vect-fcm-eq-f.c
@@ -7,7 +7,7 @@
#include "vect-fcm.x"
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 3 "vect" } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 4 "vect" } } */
/* { dg-final { scan-assembler "fcmeq\\tv\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s" } } */
/* { dg-final { scan-assembler "fcmeq\\tv\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s, 0" } } */
/* { dg-final { cleanup-tree-dump "vect" } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/vect-fcm-ge-d.c b/gcc/testsuite/gcc.target/aarch64/vect-fcm-ge-d.c
index 868e1f8..b922833 100644
--- a/gcc/testsuite/gcc.target/aarch64/vect-fcm-ge-d.c
+++ b/gcc/testsuite/gcc.target/aarch64/vect-fcm-ge-d.c
@@ -7,8 +7,9 @@
#include "vect-fcm.x"
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 3 "vect" } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 4 "vect" } } */
/* { dg-final { scan-assembler "fcmge\\tv\[0-9\]+\.2d, v\[0-9\]+\.2d, v\[0-9\]+\.2d" } } */
/* { dg-final { scan-assembler "fcmge\\tv\[0-9\]+\.2d, v\[0-9\]+\.2d, 0" } } */
+/* { dg-final { scan-assembler "fcmlt\\tv\[0-9\]+\.2d, v\[0-9\]+\.2d, 0" } } */
/* { dg-final { cleanup-tree-dump "vect" } } */
/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/vect-fcm-ge-f.c b/gcc/testsuite/gcc.target/aarch64/vect-fcm-ge-f.c
index e3258f3..04d3533 100644
--- a/gcc/testsuite/gcc.target/aarch64/vect-fcm-ge-f.c
+++ b/gcc/testsuite/gcc.target/aarch64/vect-fcm-ge-f.c
@@ -7,8 +7,9 @@
#include "vect-fcm.x"
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 3 "vect" } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 4 "vect" } } */
/* { dg-final { scan-assembler "fcmge\\tv\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s" } } */
/* { dg-final { scan-assembler "fcmge\\tv\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s, 0" } } */
+/* { dg-final { scan-assembler "fcmlt\\tv\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s, 0" } } */
/* { dg-final { cleanup-tree-dump "vect" } } */
/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/vect-fcm-gt-d.c b/gcc/testsuite/gcc.target/aarch64/vect-fcm-gt-d.c
index ed8b452..421a04a 100644
--- a/gcc/testsuite/gcc.target/aarch64/vect-fcm-gt-d.c
+++ b/gcc/testsuite/gcc.target/aarch64/vect-fcm-gt-d.c
@@ -7,8 +7,9 @@
#include "vect-fcm.x"
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 3 "vect" } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 4 "vect" } } */
/* { dg-final { scan-assembler "fcmgt\\tv\[0-9\]+\.2d, v\[0-9\]+\.2d, v\[0-9\]+\.2d" } } */
/* { dg-final { scan-assembler "fcmgt\\tv\[0-9\]+\.2d, v\[0-9\]+\.2d, 0" } } */
+/* { dg-final { scan-assembler "fcmle\\tv\[0-9\]+\.2d, v\[0-9\]+\.2d, 0" } } */
/* { dg-final { cleanup-tree-dump "vect" } } */
/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/vect-fcm-gt-f.c b/gcc/testsuite/gcc.target/aarch64/vect-fcm-gt-f.c
index e90a875..cdeab14 100644
--- a/gcc/testsuite/gcc.target/aarch64/vect-fcm-gt-f.c
+++ b/gcc/testsuite/gcc.target/aarch64/vect-fcm-gt-f.c
@@ -7,8 +7,9 @@
#include "vect-fcm.x"
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 3 "vect" } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 4 "vect" } } */
/* { dg-final { scan-assembler "fcmgt\\tv\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s" } } */
/* { dg-final { scan-assembler "fcmgt\\tv\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s, 0" } } */
+/* { dg-final { scan-assembler "fcmle\\tv\[0-9\]+\.\[24\]s, v\[0-9\]+\.\[24\]s, 0" } } */
/* { dg-final { cleanup-tree-dump "vect" } } */
/* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/vect-fcm.x b/gcc/testsuite/gcc.target/aarch64/vect-fcm.x
index 7e51bef..803861b 100644
--- a/gcc/testsuite/gcc.target/aarch64/vect-fcm.x
+++ b/gcc/testsuite/gcc.target/aarch64/vect-fcm.x
@@ -40,6 +40,15 @@ foobar (FTYPE *in1, FTYPE *in2, FTYPE *output)
output[i] = (in1[i] OP 0.0) ? 4.0 : 2.0;
}
+void
+foobarbar (FTYPE *in1, FTYPE *in2, FTYPE *output)
+{
+ int i = 0;
+ /* Vectorizable. */
+ for (i = 0; i < N; i++)
+ output[i] = (in1[i] INV_OP 0.0) ? 4.0 : 2.0;
+}
+
int
main (int argc, char **argv)
{
@@ -51,6 +60,11 @@ main (int argc, char **argv)
for (i = 0; i < N; i++)
if (out1[i] != out2[i])
abort ();
+ foobar (input1, input2, out1);
+ foobarbar (input1, input2, out2);
+ for (i = 0; i < N; i++)
+ if (out1[i] == out2[i])
+ abort ();
return 0;
}
