From: Christophe Lyon <christophe.lyon@arm.com>
To: gcc-patches@gcc.gnu.org,
Stam Markianos-Wright <stam.markianos-wright@arm.com>
Cc: kyrylo.tkachov@arm.com, Richard.Earnshaw@arm.com,
Andrea Corallo <andrea.corallo@arm.com>
Subject: Re: [PATCH 16/35] arm: Add integer vector overloading of vsubq_x instrinsic
Date: Tue, 22 Nov 2022 11:00:25 +0100 [thread overview]
Message-ID: <37d38564-70d2-a64f-f5d7-f9800c7419bc@arm.com> (raw)
In-Reply-To: <20221117163809.1009526-17-andrea.corallo@arm.com>
On 11/17/22 17:37, Andrea Corallo via Gcc-patches wrote:
> From: Stam Markianos-Wright <stam.markianos-wright@arm.com>
>
> In the past we had only defined the vsubq_x generic overload of the
> vsubq_x_* intrinsics for float vector types. This would cause them
> to fall back to the `__ARM_undef` failure state if they was called
> through the generic version.
> This patch simply adds these overloads.
>
> gcc/ChangeLog:
>
> * config/arm/arm_mve.h (__arm_vsubq_x FP): New overloads.
> (__arm_vsubq_x Integer): New.
Hi Stam,
To hopefully help Kyrill in the review, I think this fix is tested by
patch #19, where we now have
+/* { dg-final { scan-assembler-not "__ARM_undef" } } */
(this line explains why this bug was not noticed so far)
Thanks,
Christophe
> ---
> gcc/config/arm/arm_mve.h | 28 ++++++++++++++++++++++++++++
> 1 file changed, 28 insertions(+)
>
> diff --git a/gcc/config/arm/arm_mve.h b/gcc/config/arm/arm_mve.h
> index f6b42dc3fab..09167ec118e 100644
> --- a/gcc/config/arm/arm_mve.h
> +++ b/gcc/config/arm/arm_mve.h
> @@ -38259,6 +38259,18 @@ extern void *__ARM_undef;
> #define __arm_vsubq_x(p1,p2,p3) ({ __typeof(p1) __p1 = (p1); \
> __typeof(p2) __p2 = (p2); \
> _Generic( (int (*)[__ARM_mve_typeid(__p1)][__ARM_mve_typeid(__p2)])0, \
> + int (*)[__ARM_mve_type_int8x16_t][__ARM_mve_type_int8x16_t]: __arm_vsubq_x_s8 (__ARM_mve_coerce(__p1, int8x16_t), __ARM_mve_coerce(__p2, int8x16_t), p3), \
> + int (*)[__ARM_mve_type_int16x8_t][__ARM_mve_type_int16x8_t]: __arm_vsubq_x_s16 (__ARM_mve_coerce(__p1, int16x8_t), __ARM_mve_coerce(__p2, int16x8_t), p3), \
> + int (*)[__ARM_mve_type_int32x4_t][__ARM_mve_type_int32x4_t]: __arm_vsubq_x_s32 (__ARM_mve_coerce(__p1, int32x4_t), __ARM_mve_coerce(__p2, int32x4_t), p3), \
> + int (*)[__ARM_mve_type_int8x16_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_s8 (__ARM_mve_coerce(__p1, int8x16_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_int16x8_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_s16 (__ARM_mve_coerce(__p1, int16x8_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_int32x4_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_s32 (__ARM_mve_coerce(__p1, int32x4_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_uint8x16_t]: __arm_vsubq_x_u8 (__ARM_mve_coerce(__p1, uint8x16_t), __ARM_mve_coerce(__p2, uint8x16_t), p3), \
> + int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_uint16x8_t]: __arm_vsubq_x_u16 (__ARM_mve_coerce(__p1, uint16x8_t), __ARM_mve_coerce(__p2, uint16x8_t), p3), \
> + int (*)[__ARM_mve_type_uint32x4_t][__ARM_mve_type_uint32x4_t]: __arm_vsubq_x_u32 (__ARM_mve_coerce(__p1, uint32x4_t), __ARM_mve_coerce(__p2, uint32x4_t), p3), \
> + int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_u8 (__ARM_mve_coerce(__p1, uint8x16_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_u16 (__ARM_mve_coerce(__p1, uint16x8_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_uint32x4_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_u32 (__ARM_mve_coerce(__p1, uint32x4_t), __ARM_mve_coerce3(p2, int), p3), \
> int (*)[__ARM_mve_type_float16x8_t][__ARM_mve_type_float16x8_t]: __arm_vsubq_x_f16 (__ARM_mve_coerce(__p1, float16x8_t), __ARM_mve_coerce(__p2, float16x8_t), p3), \
> int (*)[__ARM_mve_type_float32x4_t][__ARM_mve_type_float32x4_t]: __arm_vsubq_x_f32 (__ARM_mve_coerce(__p1, float32x4_t), __ARM_mve_coerce(__p2, float32x4_t), p3), \
> int (*)[__ARM_mve_type_float16x8_t][__ARM_mve_type_fp_n]: __arm_vsubq_x_n_f16 (__ARM_mve_coerce(__p1, float16x8_t), __ARM_mve_coerce2(p2, double), p3), \
> @@ -40223,6 +40235,22 @@ extern void *__ARM_undef;
> int (*)[__ARM_mve_type_uint16_t_ptr]: __arm_vld4q_u16 (__ARM_mve_coerce1(p0, uint16_t *)), \
> int (*)[__ARM_mve_type_uint32_t_ptr]: __arm_vld4q_u32 (__ARM_mve_coerce1(p0, uint32_t *))))
>
> +#define __arm_vsubq_x(p1,p2,p3) ({ __typeof(p1) __p1 = (p1); \
> + __typeof(p2) __p2 = (p2); \
> + _Generic( (int (*)[__ARM_mve_typeid(__p1)][__ARM_mve_typeid(__p2)])0, \
> + int (*)[__ARM_mve_type_int8x16_t][__ARM_mve_type_int8x16_t]: __arm_vsubq_x_s8 (__ARM_mve_coerce(__p1, int8x16_t), __ARM_mve_coerce(__p2, int8x16_t), p3), \
> + int (*)[__ARM_mve_type_int16x8_t][__ARM_mve_type_int16x8_t]: __arm_vsubq_x_s16 (__ARM_mve_coerce(__p1, int16x8_t), __ARM_mve_coerce(__p2, int16x8_t), p3), \
> + int (*)[__ARM_mve_type_int32x4_t][__ARM_mve_type_int32x4_t]: __arm_vsubq_x_s32 (__ARM_mve_coerce(__p1, int32x4_t), __ARM_mve_coerce(__p2, int32x4_t), p3), \
> + int (*)[__ARM_mve_type_int8x16_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_s8 (__ARM_mve_coerce(__p1, int8x16_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_int16x8_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_s16 (__ARM_mve_coerce(__p1, int16x8_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_int32x4_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_s32 (__ARM_mve_coerce(__p1, int32x4_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_uint8x16_t]: __arm_vsubq_x_u8 (__ARM_mve_coerce(__p1, uint8x16_t), __ARM_mve_coerce(__p2, uint8x16_t), p3), \
> + int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_uint16x8_t]: __arm_vsubq_x_u16 (__ARM_mve_coerce(__p1, uint16x8_t), __ARM_mve_coerce(__p2, uint16x8_t), p3), \
> + int (*)[__ARM_mve_type_uint32x4_t][__ARM_mve_type_uint32x4_t]: __arm_vsubq_x_u32 (__ARM_mve_coerce(__p1, uint32x4_t), __ARM_mve_coerce(__p2, uint32x4_t), p3), \
> + int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_u8 (__ARM_mve_coerce(__p1, uint8x16_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_u16 (__ARM_mve_coerce(__p1, uint16x8_t), __ARM_mve_coerce3(p2, int), p3), \
> + int (*)[__ARM_mve_type_uint32x4_t][__ARM_mve_type_int_n]: __arm_vsubq_x_n_u32 (__ARM_mve_coerce(__p1, uint32x4_t), __ARM_mve_coerce3(p2, int), p3));})
> +
> #define __arm_vgetq_lane(p0,p1) ({ __typeof(p0) __p0 = (p0); \
> _Generic( (int (*)[__ARM_mve_typeid(__p0)])0, \
> int (*)[__ARM_mve_type_int8x16_t]: __arm_vgetq_lane_s8 (__ARM_mve_coerce(__p0, int8x16_t), p1), \
next prev parent reply other threads:[~2022-11-22 10:00 UTC|newest]
Thread overview: 82+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-17 16:37 [PATCH 00/35] arm: rework MVE testsuite and rework backend where necessary (1st chunk) Andrea Corallo
2022-11-17 16:37 ` [PATCH 01/35] arm: improve vcreateq* tests Andrea Corallo
2022-11-18 9:47 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 02/35] arm: fix 'vmsr' spacing and register capitalization Andrea Corallo
2022-11-18 16:33 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 03/35] arm: improve tests and fix vddupq* Andrea Corallo
2022-11-18 16:34 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 04/35] arm: improve tests and fix vdwdupq* Andrea Corallo
2022-11-18 16:35 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 05/35] arm: improve vidupq* tests Andrea Corallo
2022-11-18 16:36 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 06/35] arm: improve tests and fix vdupq* Andrea Corallo
2022-11-18 16:37 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 07/35] arm: improve tests and fix vcmp* Andrea Corallo
2022-11-18 16:40 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 08/35] arm: improve tests for vmin* Andrea Corallo
2022-11-18 16:41 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 09/35] arm: improve tests for vmax* Andrea Corallo
2022-11-18 16:42 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 10/35] arm: improve tests for vabavq* Andrea Corallo
2022-11-18 16:43 ` Kyrylo Tkachov
2022-11-21 14:49 ` Andrea Corallo
2022-11-17 16:37 ` [PATCH 11/35] arm: improve tests for vabdq* Andrea Corallo
2022-11-18 16:44 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 12/35] arm: improve tests and fix vabsq* Andrea Corallo
2022-11-18 16:45 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 13/35] arm: further fix overloading of MVE vaddq[_m]_n intrinsic Andrea Corallo
2022-11-18 16:49 ` Kyrylo Tkachov
2022-11-21 10:45 ` Stam Markianos-Wright
2022-11-17 16:37 ` [PATCH 14/35] arm: propagate fixed overloading of MVE intrinsic scalar parameters Andrea Corallo
2022-11-18 16:51 ` Kyrylo Tkachov
2022-11-21 10:46 ` Stam Markianos-Wright
2022-11-17 16:37 ` [PATCH 15/35] arm: Explicitly specify other float types for _Generic overloading [PR107515] Andrea Corallo
2022-11-18 16:58 ` Kyrylo Tkachov
2022-11-20 22:49 ` Ramana Radhakrishnan
2022-11-21 14:11 ` Stam Markianos-Wright
2022-11-21 10:45 ` Stam Markianos-Wright
2022-11-17 16:37 ` [PATCH 16/35] arm: Add integer vector overloading of vsubq_x instrinsic Andrea Corallo
2022-11-22 10:00 ` Christophe Lyon [this message]
2022-11-22 10:54 ` Andrea Corallo
2022-11-22 16:48 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 17/35] arm: improve tests and fix vadd* Andrea Corallo
2022-11-22 16:49 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 18/35] arm: improve tests for vmulq* Andrea Corallo
2022-11-22 16:51 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 19/35] arm: improve tests and fix vsubq* Andrea Corallo
2022-11-22 16:51 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 20/35] arm: improve tests for vfmasq_m* Andrea Corallo
2022-11-22 16:52 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 21/35] arm: improve tests for vhaddq_m* Andrea Corallo
2022-11-22 16:53 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 22/35] arm: improve tests for vhsubq_m* Andrea Corallo
2022-11-22 16:53 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 23/35] arm: improve tests for viwdupq* Andrea Corallo
2022-11-22 16:54 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 24/35] arm: improve tests for vmladavaq* Andrea Corallo
2022-11-22 16:54 ` Kyrylo Tkachov
2022-11-17 16:37 ` [PATCH 25/35] arm: improve tests and fix vmlaldavaxq* Andrea Corallo
2022-11-22 16:56 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 26/35] arm: improve tests for vmlasq* Andrea Corallo
2022-11-22 16:56 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 27/35] arm: improve tests for vqaddq_m* Andrea Corallo
2022-11-22 16:57 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 28/35] arm: improve tests for vqdmlahq_m* Andrea Corallo
2022-11-22 16:57 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 29/35] arm: improve tests for vqdmul* Andrea Corallo
2022-11-22 16:58 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 30/35] arm: improve tests for vqrdmlahq* Andrea Corallo
2022-11-22 17:01 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 31/35] arm: improve tests for vqrdmlashq_m* Andrea Corallo
2022-11-22 17:02 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 32/35] arm: improve tests for vqsubq* Andrea Corallo
2022-11-22 17:03 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 33/35] arm: improve tests and fix vrmlaldavhaq* Andrea Corallo
2022-11-22 17:03 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 34/35] arm: improve tests for vrshlq* Andrea Corallo
2022-11-22 17:04 ` Kyrylo Tkachov
2022-11-17 16:38 ` [PATCH 35/35] arm: improve tests for vsetq_lane* Andrea Corallo
2022-11-22 17:06 ` Kyrylo Tkachov
2022-11-24 14:43 ` [PATCH 35/35 V2] " Andrea Corallo
2022-11-24 15:28 ` Kyrylo Tkachov
2022-11-28 9:20 ` [PATCH 00/35] arm: rework MVE testsuite and rework backend where necessary (1st chunk) Andrea Corallo
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=37d38564-70d2-a64f-f5d7-f9800c7419bc@arm.com \
--to=christophe.lyon@arm.com \
--cc=Richard.Earnshaw@arm.com \
--cc=andrea.corallo@arm.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=kyrylo.tkachov@arm.com \
--cc=stam.markianos-wright@arm.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).