From: Richard Guenther <richard.guenther@gmail.com>
To: Richard Guenther <richard.guenther@gmail.com>,
gcc-patches@gcc.gnu.org, patches@linaro.org,
richard.sandiford@linaro.org
Subject: Re: [5/9] Main target-independent support for direct interleaving
Date: Mon, 18 Apr 2011 13:22:00 -0000 [thread overview]
Message-ID: <BANLkTincW9V92sbkoty3NnPsgKj95XdyVQ@mail.gmail.com> (raw)
In-Reply-To: <g47harah0u.fsf@linaro.org>
On Mon, Apr 18, 2011 at 2:19 PM, Richard Sandiford
<richard.sandiford@linaro.org> wrote:
> Richard Guenther <richard.guenther@gmail.com> writes:
>> On Mon, Apr 18, 2011 at 1:24 PM, Richard Sandiford
>> <richard.sandiford@linaro.org> wrote:
>>> Richard Guenther <richard.guenther@gmail.com> writes:
>>>> On Tue, Apr 12, 2011 at 3:59 PM, Richard Sandiford
>>>> <richard.sandiford@linaro.org> wrote:
>>>>> Index: gcc/doc/md.texi
>>>>> ===================================================================
>>>>> --- gcc/doc/md.texi 2011-04-12 12:16:46.000000000 +0100
>>>>> +++ gcc/doc/md.texi 2011-04-12 14:48:28.000000000 +0100
>>>>> @@ -3846,6 +3846,48 @@ into consecutive memory locations. Oper
>>>>> consecutive memory locations, operand 1 is the first register, and
>>>>> operand 2 is a constant: the number of consecutive registers.
>>>>>
>>>>> +@cindex @code{vec_load_lanes@var{m}@var{n}} instruction pattern
>>>>> +@item @samp{vec_load_lanes@var{m}@var{n}}
>>>>> +Perform an interleaved load of several vectors from memory operand 1
>>>>> +into register operand 0. Both operands have mode @var{m}. The register
>>>>> +operand is viewed as holding consecutive vectors of mode @var{n},
>>>>> +while the memory operand is a flat array that contains the same number
>>>>> +of elements. The operation is equivalent to:
>>>>> +
>>>>> +@smallexample
>>>>> +int c = GET_MODE_SIZE (@var{m}) / GET_MODE_SIZE (@var{n});
>>>>> +for (j = 0; j < GET_MODE_NUNITS (@var{n}); j++)
>>>>> + for (i = 0; i < c; i++)
>>>>> + operand0[i][j] = operand1[j * c + i];
>>>>> +@end smallexample
>>>>> +
>>>>> +For example, @samp{vec_load_lanestiv4hi} loads 8 16-bit values
>>>>> +from memory into a register of mode @samp{TI}@. The register
>>>>> +contains two consecutive vectors of mode @samp{V4HI}@.
>>>>
>>>> So vec_load_lanestiv2qi would load ... ? c == 8 here. Intuitively
>>>> such operation would have adjacent blocks of siv2qi memory. But
>>>> maybe you want to constrain the mode size to GET_MODE_SIZE (@var{n})
>>>> * GET_MODE_NUNITS (@var{n})? In which case the mode m is
>>>> redundant? You could specify that we load NUNITS adjacent vectors into
>>>> an integer mode of appropriate size.
>>>
>>> Like you say, vec_load_lanestiv2qi would load 16 QImode elements into
>>> 8 consecutive V2QI registers. The first element from register vector I
>>> would come from operand1[I] and the second element would come from
>>> operand1[I + 8]. That's meant to be a valid combination.
>>
>> Ok, but the C loop from the example doesn't seem to match. Or I couldn't
>> wrap my head around it despite looking for 5 minutes and already having
>> coffee ;) I would have expected the vectors being in memory as
>>
>> v0[0], v1[0], v0[1], v1[1], v2[0], v3[1]. v2[1], v3[1], ...
>>
>> not
>>
>> v0[0], v1[0], v2[0], ...
>>
>> as I would have thought the former is more useful (simple unrolling for
>> stride 2).
>
> The second one's right. All lane 0 elements, followed by all lane 1
> elements, etc. I think that's what the C loop says.
>
>> We'd need a separate set of optabs for such an interleaving
>> scheme? In which case we might want to come up with a more
>> specific name than load_lane?
>
> Yeah, if someone has a single instruction that does your first example,
> then it would need a new optab. The individual vector pairs could be
> represented using the current optab though, if each pair needs a
> separate instruction. E.g. with your v2qi example, vec_load_lanessiv2qi
> would load:
>
> v0[0], v1[0], v0[1], v1[1]
>
> and you could repeat for the others. So load_lanes (as defined here)
> could be treated as a primitive, and your first example could be something
> like "repeat_load_lanes".
>
> If you don't like the name "load_lanes" though, I'm happy to use
> something else.
Ah, no - repeat_load_lanes sounds a good name for the new optab if
we need it at any point.
Richard.
> Richard
>
next prev parent reply other threads:[~2011-04-18 12:59 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-04-12 13:21 [0/9] Direct support for loads and stores of interleaved vectors Richard Sandiford
2011-04-12 13:25 ` [1/9] Generalise vect_create_data_ref_ptr Richard Sandiford
2011-04-12 13:30 ` Richard Guenther
2011-04-12 13:28 ` [2/9] Reindent parts of vectorizable_load and vectorizable_store Richard Sandiford
2011-04-12 13:33 ` Richard Guenther
2011-04-12 14:39 ` Richard Sandiford
2011-04-12 13:40 ` [3/9] STMT_VINFO_RELATED_STMT handling in vectorizable_store Richard Sandiford
2011-04-17 10:25 ` Ira Rosen
2011-04-12 13:44 ` [4/9] Move power-of-two checks for interleaving Richard Sandiford
2011-04-12 13:57 ` Richard Guenther
2011-04-12 13:59 ` [5/9] Main target-independent support for direct interleaving Richard Sandiford
2011-04-17 14:26 ` Ira Rosen
2011-04-18 11:54 ` Richard Guenther
2011-04-18 11:57 ` Richard Sandiford
2011-04-18 12:54 ` Richard Guenther
2011-04-18 12:58 ` Richard Sandiford
2011-04-18 13:22 ` Richard Guenther [this message]
2011-04-12 14:01 ` [6/9] NEON vec_load_lanes and vec_store_lanes patterns Richard Sandiford
2011-04-15 13:20 ` Richard Earnshaw
2011-04-12 14:14 ` [7/9] Testsuite: remove vect_{extract_even_odd,strided}_wide Richard Sandiford
2011-04-15 12:43 ` Richard Guenther
2011-04-12 14:19 ` [8/9] Testsuite: split tests for strided accesses Richard Sandiford
2011-04-15 12:44 ` Richard Guenther
2011-04-12 14:29 ` [9/9] Testsuite: Replace vect_strided with vect_stridedN Richard Sandiford
2011-04-15 12:44 ` Richard Guenther
2011-04-12 14:34 ` [10/9] Add tests for stride-3 accesses Richard Sandiford
2011-04-15 12:45 ` Richard Guenther
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=BANLkTincW9V92sbkoty3NnPsgKj95XdyVQ@mail.gmail.com \
--to=richard.guenther@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=patches@linaro.org \
--cc=richard.sandiford@linaro.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).