From: Segher Boessenkool <segher@kernel.crashing.org>
To: Xionghu Luo <luoxhu@linux.ibm.com>
Cc: gcc-patches@gcc.gnu.org, dje.gcc@gmail.com,
wschmidt@linux.ibm.com, guojiufu@linux.ibm.com,
linkw@gcc.gnu.org
Subject: Re: [PATCH 2/4] rs6000: Support variable insert and Expand vec_insert in expander [PR79251]
Date: Tue, 24 Nov 2020 16:37:46 -0600 [thread overview]
Message-ID: <20201124223746.GY2672@gate.crashing.org> (raw)
In-Reply-To: <20201010080825.3599892-3-luoxhu@linux.ibm.com>
Hi!
On Sat, Oct 10, 2020 at 03:08:23AM -0500, Xionghu Luo wrote:
> vec_insert accepts 3 arguments, arg0 is input vector, arg1 is the value
> to be insert, arg2 is the place to insert arg1 to arg0. Current expander
> generates stxv+stwx+lxv if arg2 is variable instead of constant, which
> causes serious store hit load performance issue on Power. This patch tries
> 1) Build VIEW_CONVERT_EXPR for vec_insert (i, v, n) like v[n&3] = i to
> unify the gimple code, then expander could use vec_set_optab to expand.
> 2) Expand the IFN VEC_SET to fast instructions: lvsr+insert+lvsl.
> In this way, "vec_insert (i, v, n)" and "v[n&3] = i" won't be expanded too
> early in gimple stage if arg2 is variable, avoid generating store hit load
> instructions.
>
> For Power9 V4SI:
> addi 9,1,-16
> rldic 6,6,2,60
> stxv 34,-16(1)
> stwx 5,9,6
> lxv 34,-16(1)
> =>
> rlwinm 6,6,2,28,29
> mtvsrwz 0,5
> lvsr 1,0,6
> lvsl 0,0,6
> xxperm 34,34,33
> xxinsertw 34,0,12
> xxperm 34,34,32
It still takes me quite some time to verify this, tricky bit-fiddling!
But the code that generates this is easier to read :-)
> +/* Insert VAL into IDX of TARGET, VAL size is same of the vector element, IDX
> + is variable and also counts by vector element size. */
"Set vector element IDX of TARGET to VAL. IDX is not a constant
integer."?
Okay for trunk (with an improved comment). Thanks!
Segher
next prev parent reply other threads:[~2020-11-24 22:38 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-10 8:08 [PATCH 0/4] rs6000: Enable variable vec_insert with IFN VEC_SET Xionghu Luo
2020-10-10 8:08 ` [PATCH 1/4] rs6000: Change rs6000_expand_vector_set param Xionghu Luo
2020-11-24 19:44 ` Segher Boessenkool
2020-10-10 8:08 ` [PATCH 2/4] rs6000: Support variable insert and Expand vec_insert in expander [PR79251] Xionghu Luo
2020-11-24 22:37 ` Segher Boessenkool [this message]
2020-10-10 8:08 ` [PATCH 3/4] rs6000: Enable vec_insert for P8 with rs6000_expand_vector_set_var_p8 Xionghu Luo
2020-11-27 1:04 ` Xionghu Luo
2020-12-03 14:16 ` Xionghu Luo
2020-12-10 3:32 ` Xionghu Luo
2020-12-23 2:18 ` Ping ^ 3: " Xionghu Luo
2021-01-15 2:48 ` Ping ^ 4: " Xionghu Luo
2021-01-21 23:48 ` Segher Boessenkool
2021-01-22 20:08 ` David Edelsohn
2020-10-10 8:08 ` [PATCH 4/4] rs6000: Update testcases' instruction count Xionghu Luo
2021-01-22 0:17 ` Segher Boessenkool
2021-01-22 20:02 ` David Edelsohn
2021-01-23 1:01 ` Segher Boessenkool
2021-01-23 1:24 ` David Edelsohn
2020-11-05 1:34 ` Ping: [PATCH 0/4] rs6000: Enable variable vec_insert with IFN VEC_SET Xionghu Luo
2020-11-13 2:05 ` Ping^2: " Xionghu Luo
2020-11-24 2:29 ` Ping^3: " Xionghu Luo
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=20201124223746.GY2672@gate.crashing.org \
--to=segher@kernel.crashing.org \
--cc=dje.gcc@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=guojiufu@linux.ibm.com \
--cc=linkw@gcc.gnu.org \
--cc=luoxhu@linux.ibm.com \
--cc=wschmidt@linux.ibm.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).