* [AArch64_be] Fix vldX/vstX AdvSIMD intrinsics
@ 2015-09-02 13:18 Christophe Lyon
2015-09-02 13:53 ` James Greenhalgh
0 siblings, 1 reply; 2+ messages in thread
From: Christophe Lyon @ 2015-09-02 13:18 UTC (permalink / raw)
To: gcc-patches
[-- Attachment #1: Type: text/plain, Size: 677 bytes --]
Hi,
The aarch64_vldX/aarch64_vstX expanders used for the vldX/vstX AdvSIMD
intrisics in Q mode called vec_load_lanes, witch shuffles the vectors
to match the layout expected by the vectorizer.
We do not want this to happen when the intrinsics are called directly
by the end-user code.
This patch fixes this, by calling gen_aarch64_simd_ldX/gen_aarch64_simd_stX.
With this patch, the following tests now pass in advsimd-intrinsics
(target aarch64_be):
vldX_lane.c, vtrn, vuzp, vzip
as well as aarch64/vldN_1.c and aarch64/vstN_1.c
It fixes PR 59810, 63652, 63653.
No regression, and tested on aarch64 and aarch64_be using the Foundation Model.
OK for trunk?
Christophe.
[-- Attachment #2: vldX-vstX.patch --]
[-- Type: text/x-patch, Size: 1091 bytes --]
2015-09-02 Christophe Lyon <christophe.lyon@linaro.org>
PR target/59810
PR target/63652
PR target/63653
* config/aarch64/aarch64-simd.md
(aarch64_ld<VSTRUCT:nregs><VQ:mode>): Call
gen_aarch64_simd_ld<VSTRUCT:nregs><VQ:mode>.
(aarch64_st<VSTRUCT:nregs><VQ:mode>): Call
gen_aarch64_simd_st<VSTRUCT:nregs><VQ:mode>.
diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
index 9777418..75fa0ab 100644
--- a/gcc/config/aarch64/aarch64-simd.md
+++ b/gcc/config/aarch64/aarch64-simd.md
@@ -4566,7 +4566,7 @@
machine_mode mode = <VSTRUCT:MODE>mode;
rtx mem = gen_rtx_MEM (mode, operands[1]);
- emit_insn (gen_vec_load_lanes<VSTRUCT:mode><VQ:mode> (operands[0], mem));
+ emit_insn (gen_aarch64_simd_ld<VSTRUCT:nregs><VQ:mode> (operands[0], mem));
DONE;
})
@@ -4849,7 +4849,7 @@
machine_mode mode = <VSTRUCT:MODE>mode;
rtx mem = gen_rtx_MEM (mode, operands[0]);
- emit_insn (gen_vec_store_lanes<VSTRUCT:mode><VQ:mode> (mem, operands[1]));
+ emit_insn (gen_aarch64_simd_st<VSTRUCT:nregs><VQ:mode> (mem, operands[1]));
DONE;
})
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [AArch64_be] Fix vldX/vstX AdvSIMD intrinsics
2015-09-02 13:18 [AArch64_be] Fix vldX/vstX AdvSIMD intrinsics Christophe Lyon
@ 2015-09-02 13:53 ` James Greenhalgh
0 siblings, 0 replies; 2+ messages in thread
From: James Greenhalgh @ 2015-09-02 13:53 UTC (permalink / raw)
To: Christophe Lyon; +Cc: gcc-patches
On Wed, Sep 02, 2015 at 02:18:03PM +0100, Christophe Lyon wrote:
> Hi,
>
> The aarch64_vldX/aarch64_vstX expanders used for the vldX/vstX AdvSIMD
> intrisics in Q mode called vec_load_lanes, witch shuffles the vectors
> to match the layout expected by the vectorizer.
>
> We do not want this to happen when the intrinsics are called directly
> by the end-user code.
>
> This patch fixes this, by calling gen_aarch64_simd_ldX/gen_aarch64_simd_stX.
>
> With this patch, the following tests now pass in advsimd-intrinsics
> (target aarch64_be):
> vldX_lane.c, vtrn, vuzp, vzip
> as well as aarch64/vldN_1.c and aarch64/vstN_1.c
>
> It fixes PR 59810, 63652, 63653.
Great!
>
> No regression, and tested on aarch64 and aarch64_be using the Foundation Model.
>
> OK for trunk?
OK.
Thanks,
James
> 2015-09-02 Christophe Lyon <christophe.lyon@linaro.org>
>
> PR target/59810
> PR target/63652
> PR target/63653
> * config/aarch64/aarch64-simd.md
> (aarch64_ld<VSTRUCT:nregs><VQ:mode>): Call
> gen_aarch64_simd_ld<VSTRUCT:nregs><VQ:mode>.
> (aarch64_st<VSTRUCT:nregs><VQ:mode>): Call
> gen_aarch64_simd_st<VSTRUCT:nregs><VQ:mode>.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2015-09-02 13:53 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-09-02 13:18 [AArch64_be] Fix vldX/vstX AdvSIMD intrinsics Christophe Lyon
2015-09-02 13:53 ` James Greenhalgh
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).