From: Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP>
Subject: Re: Forcing GNAT to use 32-bit load/store instructions on ARM?
Date: Fri, 11 Jul 2014 20:22:11 +0000 (UTC)
Date: 2014-07-11T20:22:11+00:00 [thread overview]
Message-ID: <lpph1i$7q2$1@dont-email.me> (raw)
In-Reply-To: lybnsv3h5n.fsf@pushface.org
On 2014-07-11, Simon Wright <simon@pushface.org> wrote:
> Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> writes:
>
>> On 2014-07-10, Simon Wright <simon@pushface.org> wrote:
>>> Simon Wright <simon@pushface.org> writes:
>>>
>>>> The x86_64 code accesses all 32 bits of V. If it is more efficient on
>>>> arm to access the relevant bytes for the component assignments, I
>>>> expect that's what would happen (as reported).
>>>
>>> Just tried this on Raspberry Pi (GCC 4.6.3): yes, it only accesses the
>>> relevant bytes.
>>
>> Just to make sure: you mean it uses ldrb/strb instead of ldr/str
>> in the generated code ?
>
> Yes. I don't understand all the asm, but that part is clear!
>
> _ada_atom:
> @ args = 0, pretend = 0, frame = 0
> @ frame_needed = 0, uses_anonymous_args = 0
> @ link register save eliminated.
> ldr r2, .L2
> ldr r3, .L2+4
> ldmia r2, {r1, r2}
> str r1, [r3, #0]
> ldrb r1, [r3, #0] @ zero_extendqisi2
> bic r1, r1, #1
> strb r1, [r3, #0]
> ldrb r1, [r3, #3] @ zero_extendqisi2
^^^
Eeeek! :-)
That's an "interesting" variant I have not seen before.
It looks like gcc is generating code to do a byte offset load directly
from the top byte of the 32-bit word. In my bitfield struct experiments
in C I only ever saw gcc use ldrb when the bitfield was in the low
8 bits of the 32-bit word. I never managed to get gcc to do the above.
Congratulations on finding yet another variant to this. :-)
> bic r1, r1, #128
> strb r1, [r3, #3]
> str r2, [r3, #0]
> bx lr
Simon.
--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world
next prev parent reply other threads:[~2014-07-11 20:22 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-06-30 22:11 Forcing GNAT to use 32-bit load/store instructions on ARM? daniel.dmk
2014-06-30 23:41 ` Jeffrey Carter
2014-07-01 12:06 ` Simon Clubley
2014-07-01 15:44 ` Niklas Holsti
2014-07-01 17:26 ` Simon Clubley
2014-07-01 17:18 ` Simon Wright
2014-07-01 19:43 ` Simon Wright
2014-07-01 17:28 ` Jeffrey Carter
2014-07-01 0:55 ` anon
2014-07-01 4:30 ` Niklas Holsti
2014-07-01 8:11 ` Dmitry A. Kazakov
2014-07-01 12:09 ` Simon Clubley
2014-07-01 12:20 ` Dmitry A. Kazakov
2014-07-01 17:00 ` Simon Clubley
2014-07-01 19:36 ` Dmitry A. Kazakov
2014-07-01 20:08 ` Simon Clubley
2014-07-02 22:24 ` Randy Brukardt
2014-07-06 20:40 ` MatthiasR
2014-07-07 0:25 ` Simon Clubley
2014-07-07 22:38 ` Randy Brukardt
2014-07-08 6:51 ` Simon Wright
2014-07-10 11:47 ` Simon Wright
2014-07-10 13:06 ` Simon Clubley
2014-07-11 18:05 ` Simon Wright
2014-07-11 20:22 ` Simon Clubley [this message]
2014-07-08 8:50 ` Brian Drummond
2014-07-08 12:12 ` Simon Clubley
2014-07-08 13:26 ` G.B.
2014-07-08 17:13 ` Simon Clubley
2014-07-08 15:36 ` Adam Beneschan
2014-07-08 15:40 ` Adam Beneschan
2014-07-08 20:34 ` Randy Brukardt
2014-07-09 7:31 ` Dmitry A. Kazakov
2014-07-10 0:11 ` Simon Clubley
2014-07-20 11:35 ` MatthiasR
2014-07-20 15:49 ` Simon Clubley
2014-07-26 11:05 ` MatthiasR
2014-08-10 11:20 ` MatthiasR
2014-07-01 12:03 ` Simon Clubley
2014-07-01 19:52 ` daniel.dmk
2014-07-01 20:40 ` Simon Clubley
2014-07-01 20:55 ` Simon Clubley
2014-07-01 21:01 ` Niklas Holsti
2014-07-01 21:20 ` Simon Clubley
2014-07-01 22:38 ` Niklas Holsti
2014-07-02 16:49 ` Simon Clubley
2014-07-01 21:55 ` daniel.dmk
2014-07-02 7:30 ` Simon Wright
2014-07-02 18:52 ` daniel.dmk
2014-07-04 23:51 ` Niklas Holsti
2014-07-05 0:18 ` Niklas Holsti
replies disabled
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox