* [PATCH v2 2/2] aarch64: Lower vcombine to GIMPLE
@ 2022-07-12 14:43 Andrew Carlotti
2022-07-13 8:30 ` Richard Sandiford
0 siblings, 1 reply; 2+ messages in thread
From: Andrew Carlotti @ 2022-07-12 14:43 UTC (permalink / raw)
To: gcc-patches; +Cc: Richard Biener, Richard Sandiford
This lowers vcombine intrinsics to a GIMPLE vector constructor, which enables
better optimisation during GIMPLE passes.
gcc/
* config/aarch64/aarch64-builtins.c
(aarch64_general_gimple_fold_builtin): Add combine.
gcc/testsuite/
* gcc.target/aarch64/advsimd-intrinsics/combine.c:
New test.
diff --git a/gcc/config/aarch64/aarch64-builtins.cc b/gcc/config/aarch64/aarch64-builtins.cc
index 5753988a9964967c27a03aca5fddb9025fd8ed6e..a25756cfed5fab3a98ebf3e2ee29a5e117cbd2aa 100644
--- a/gcc/config/aarch64/aarch64-builtins.cc
+++ b/gcc/config/aarch64/aarch64-builtins.cc
@@ -2857,6 +2857,28 @@ aarch64_general_gimple_fold_builtin (unsigned int fcode, gcall *stmt,
gimple_call_set_lhs (new_stmt, gimple_call_lhs (stmt));
break;
+ BUILTIN_VDC (BINOP, combine, 0, AUTO_FP)
+ BUILTIN_VD_I (BINOPU, combine, 0, NONE)
+ BUILTIN_VDC_P (BINOPP, combine, 0, NONE)
+ {
+ tree first_part, second_part;
+ if (BYTES_BIG_ENDIAN)
+ {
+ second_part = args[0];
+ first_part = args[1];
+ }
+ else
+ {
+ first_part = args[0];
+ second_part = args[1];
+ }
+ tree ret_type = TREE_TYPE (gimple_call_lhs (stmt));
+ tree ctor = build_constructor_va (ret_type, 2, NULL_TREE, first_part,
+ NULL_TREE, second_part);
+ new_stmt = gimple_build_assign (gimple_call_lhs (stmt), ctor);
+ }
+ break;
+
/*lower store and load neon builtins to gimple. */
BUILTIN_VALL_F16 (LOAD1, ld1, 0, LOAD)
BUILTIN_VDQ_I (LOAD1_U, ld1, 0, LOAD)
diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/combine.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/combine.c
new file mode 100644
index 0000000000000000000000000000000000000000..d08faf7a4a160a1e83428ed9b270731bbf7b8c8a
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/combine.c
@@ -0,0 +1,18 @@
+/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-final { check-function-bodies "**" "" {-O[^0]} } } */
+/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
+
+#include <arm_neon.h>
+
+/*
+** foo:
+** umov w0, v1\.s\[1\]
+** ret
+*/
+
+int32_t foo (int32x2_t a, int32x2_t b)
+{
+ int32x4_t c = vcombine_s32(a, b);
+ return vgetq_lane_s32(c, 3);
+}
+
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH v2 2/2] aarch64: Lower vcombine to GIMPLE
2022-07-12 14:43 [PATCH v2 2/2] aarch64: Lower vcombine to GIMPLE Andrew Carlotti
@ 2022-07-13 8:30 ` Richard Sandiford
0 siblings, 0 replies; 2+ messages in thread
From: Richard Sandiford @ 2022-07-13 8:30 UTC (permalink / raw)
To: Andrew Carlotti; +Cc: gcc-patches, Richard Biener
Andrew Carlotti <andrew.carlotti@arm.com> writes:
> This lowers vcombine intrinsics to a GIMPLE vector constructor, which enables
> better optimisation during GIMPLE passes.
>
> gcc/
>
> * config/aarch64/aarch64-builtins.c
> (aarch64_general_gimple_fold_builtin): Add combine.
>
> gcc/testsuite/
>
> * gcc.target/aarch64/advsimd-intrinsics/combine.c:
> New test.
>
> diff --git a/gcc/config/aarch64/aarch64-builtins.cc b/gcc/config/aarch64/aarch64-builtins.cc
> index 5753988a9964967c27a03aca5fddb9025fd8ed6e..a25756cfed5fab3a98ebf3e2ee29a5e117cbd2aa 100644
> --- a/gcc/config/aarch64/aarch64-builtins.cc
> +++ b/gcc/config/aarch64/aarch64-builtins.cc
> @@ -2857,6 +2857,28 @@ aarch64_general_gimple_fold_builtin (unsigned int fcode, gcall *stmt,
> gimple_call_set_lhs (new_stmt, gimple_call_lhs (stmt));
> break;
>
> + BUILTIN_VDC (BINOP, combine, 0, AUTO_FP)
> + BUILTIN_VD_I (BINOPU, combine, 0, NONE)
> + BUILTIN_VDC_P (BINOPP, combine, 0, NONE)
> + {
> + tree first_part, second_part;
> + if (BYTES_BIG_ENDIAN)
> + {
> + second_part = args[0];
> + first_part = args[1];
> + }
> + else
> + {
> + first_part = args[0];
> + second_part = args[1];
> + }
> + tree ret_type = TREE_TYPE (gimple_call_lhs (stmt));
Just repeating what we discussed off-list for the record: this needs
to be gimple_call_return_type. LGTM with that change.
Thanks,
Richard
> + tree ctor = build_constructor_va (ret_type, 2, NULL_TREE, first_part,
> + NULL_TREE, second_part);
> + new_stmt = gimple_build_assign (gimple_call_lhs (stmt), ctor);
> + }
> + break;
> +
> /*lower store and load neon builtins to gimple. */
> BUILTIN_VALL_F16 (LOAD1, ld1, 0, LOAD)
> BUILTIN_VDQ_I (LOAD1_U, ld1, 0, LOAD)
> diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/combine.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/combine.c
> new file mode 100644
> index 0000000000000000000000000000000000000000..d08faf7a4a160a1e83428ed9b270731bbf7b8c8a
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/combine.c
> @@ -0,0 +1,18 @@
> +/* { dg-do compile { target { aarch64*-*-* } } } */
> +/* { dg-final { check-function-bodies "**" "" {-O[^0]} } } */
> +/* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
> +
> +#include <arm_neon.h>
> +
> +/*
> +** foo:
> +** umov w0, v1\.s\[1\]
> +** ret
> +*/
> +
> +int32_t foo (int32x2_t a, int32x2_t b)
> +{
> + int32x4_t c = vcombine_s32(a, b);
> + return vgetq_lane_s32(c, 3);
> +}
> +
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2022-07-13 8:30 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-07-12 14:43 [PATCH v2 2/2] aarch64: Lower vcombine to GIMPLE Andrew Carlotti
2022-07-13 8:30 ` Richard Sandiford
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).