From: Richard Sandiford <richard.sandiford@arm.com>
To: Christophe LYON <christophe.lyon@foss.st.com>
Cc: Christophe Lyon via Gcc-patches <gcc-patches@gcc.gnu.org>
Subject: Re: [PATCH v2 12/14] arm: Convert more load/store MVE builtins to predicate qualifiers
Date: Fri, 15 Oct 2021 16:56:51 +0100 [thread overview]
Message-ID: <mptlf2u9t8c.fsf@arm.com> (raw)
In-Reply-To: <d7a72be0-dd39-ebd5-2b0b-6f9ae9e25716@foss.st.com> (Christophe LYON's message of "Fri, 15 Oct 2021 17:33:35 +0200")
Christophe LYON <christophe.lyon@foss.st.com> writes:
> On 15/10/2021 17:08, Richard Sandiford wrote:
>> Christophe Lyon via Gcc-patches <gcc-patches@gcc.gnu.org> writes:
>>> This patch covers a few builtins where we do not use the <mode>
>>> iterator and thus we cannot use <MVE_vpred>.
>>>
>>> For v2di instructions, we use the V8BI mode for predicates.
>> Why V8BI though, when VPRED uses HI?
>
>
> Hmm.. I used your suggestion:
>
> https://gcc.gnu.org/pipermail/gcc-patches/2021-October/581362.html
>
> Maybe I misinterpreted it?
Gah! Sorry, that was a typo for V2BI :-(
>> Would it make sense to define a V2BI? Or doesn't that work?
>
> I didn't try, I'll have a look. wrt the architecture, I'm not sure what
> this would mean?
Hmm, yeah, I guess V2BI is wrong because every 4th bit matters.
It would need to be V4BI instead.
I guess if the number of predicate elements doesn't match the number
of data elements, there's probably not much point using a dedicated
predicate mode.
So yeah, maybe keeping HImode is better after all. Sorry for the
misdirection.
Richard
next prev parent reply other threads:[~2021-10-15 15:56 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-13 10:15 [PATCH v2 00/14] ARM/MVE use vectors of boolean for predicates Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 01/14] arm: Add new tests for comparison vectorization with Neon and MVE Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 02/14] arm: Add tests for PR target/100757 Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 03/14] arm: Add tests for PR target/101325 Christophe Lyon
2021-10-15 14:49 ` Richard Sandiford
2021-10-13 10:15 ` [PATCH v2 04/14] arm: Add GENERAL_AND_VPR_REGS regclass Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 05/14] arm: Add support for VPR_REG in arm_class_likely_spilled_p Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 06/14] arm: Fix mve_vmvnq_n_<supf><mode> argument mode Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 07/14] arm: Implement MVE predicates as vectors of booleans Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 08/14] arm: Implement auto-vectorized MVE comparisons with vectors of boolean predicates Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 09/14] arm: Fix vcond_mask expander for MVE (PR target/100757) Christophe Lyon
2021-10-15 14:55 ` Richard Sandiford
2021-10-13 10:15 ` [PATCH v2 10/14] arm: Convert remaining MVE vcmp builtins to predicate qualifiers Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 11/14] arm: Convert more MVE " Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 12/14] arm: Convert more load/store " Christophe Lyon
2021-10-15 15:08 ` Richard Sandiford
2021-10-15 15:33 ` Christophe LYON
2021-10-15 15:56 ` Richard Sandiford [this message]
2021-10-13 10:15 ` [PATCH v2 13/14] arm: Convert more MVE/CDE " Christophe Lyon
2021-10-13 10:15 ` [PATCH v2 14/14] arm: Add VPR_REG to ALL_REGS Christophe Lyon
2021-10-15 15:10 ` [PATCH v2 00/14] ARM/MVE use vectors of boolean for predicates Richard Sandiford
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=mptlf2u9t8c.fsf@arm.com \
--to=richard.sandiford@arm.com \
--cc=christophe.lyon@foss.st.com \
--cc=gcc-patches@gcc.gnu.org \
/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).