From: Richard Sandiford <richard.sandiford@linaro.org>
To: Robin Dapp <rdapp@linux.vnet.ibm.com>
Cc: Richard Biener <richard.guenther@gmail.com>,
GCC Patches <gcc-patches@gcc.gnu.org>,
"Bin.Cheng" <amker.cheng@gmail.com>,
Andreas Krebbel <krebbel@linux.vnet.ibm.com>
Subject: Re: [PATCH 2/5 v3] Vect peeling cost model
Date: Wed, 24 May 2017 07:53:00 -0000 [thread overview]
Message-ID: <87inkqzo5y.fsf@linaro.org> (raw)
In-Reply-To: <5fe88af1-b9dc-ad8d-95e3-9843872f2e58@linux.vnet.ibm.com> (Robin Dapp's message of "Wed, 24 May 2017 09:21:47 +0200")
Robin Dapp <rdapp@linux.vnet.ibm.com> writes:
>> Not sure I've understood the series TBH, but is the npeel == vf / 2
>> there specifically for the "unknown number of peels" case? How do
>> we distinguish that from the case in which the number of peels is
>> known to be vf / 2 at compile time? Or have I missed the point
>> completely? (probably yes, sorry!)
>
> Good point, that's not totally waterproof for future uses of
> vect_get_peeling_costs_all_drs (). Currently, however, only when
> peeling for unknown alignment vf != 0 will be passed to it (and vf == 0
> for the known alignment case), so we can distinguish the cases.
Ah, makes sense now, thanks. Would you mind putting something like
that last sentence in a comment?
> In future, the whole vf/2 handling should be improved anyway since e.g.
> it is hardcoded here as well as in tree-vect-loop.c. npeel = 0 also has
> a double meaning, namely not peeling when peeling for known alignment
> and peeling vf/2 iters when peeling for unknown alignment. Room for
> improvement I guess :)
Yeah :-) But thanks for the series, looks like a nice improvement.
Richard
next prev parent reply other threads:[~2017-05-24 7:53 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-04-11 14:38 [RFC] S/390: Alignment peeling prolog generation Robin Dapp
2017-04-11 14:57 ` Bin.Cheng
2017-04-11 15:03 ` Robin Dapp
2017-04-11 15:07 ` Bin.Cheng
2017-04-11 16:25 ` Richard Biener
2017-04-12 7:51 ` Robin Dapp
2017-04-12 7:58 ` Richard Biener
2017-05-04 9:04 ` [PATCH 1/3] Vect peeling cost model Robin Dapp
2017-05-05 10:32 ` Richard Biener
2017-05-04 9:04 ` [RFC] S/390: Alignment peeling prolog generation Robin Dapp
2017-05-05 11:04 ` Richard Biener
2017-05-08 16:12 ` Robin Dapp
2017-05-09 10:38 ` Richard Biener
2017-05-11 11:17 ` Robin Dapp
2017-05-11 12:15 ` Richard Biener
2017-05-11 12:16 ` Richard Biener
2017-05-11 12:48 ` Richard Biener
2017-05-11 11:17 ` [PATCH 1/5] Vect peeling cost model Robin Dapp
2017-05-11 11:18 ` [PATCH 2/5] " Robin Dapp
2017-05-11 11:19 ` [PATCH 3/5] " Robin Dapp
2017-05-11 11:20 ` [PATCH 4/5] " Robin Dapp
2017-05-11 15:30 ` [PATCH 4/5 v2] " Robin Dapp
2017-05-12 9:36 ` Richard Biener
2017-05-23 15:58 ` [PATCH 0/5 v3] " Robin Dapp
2017-05-24 7:51 ` Richard Biener
2017-05-24 11:57 ` Robin Dapp
2017-05-24 13:56 ` Richard Biener
2017-06-03 17:12 ` Andreas Schwab
2017-06-06 7:13 ` Robin Dapp
2017-06-06 17:26 ` Andreas Schwab
2017-06-07 10:50 ` Robin Dapp
2017-06-07 11:43 ` Andreas Schwab
2017-05-23 15:58 ` [PATCH 1/5 " Robin Dapp
2017-05-23 15:58 ` [PATCH 2/5 " Robin Dapp
2017-05-23 19:25 ` Richard Sandiford
2017-05-24 7:37 ` Robin Dapp
2017-05-24 7:53 ` Richard Sandiford [this message]
2017-05-23 15:59 ` [PATCH 5/5 " Robin Dapp
2017-05-23 15:59 ` [PATCH 4/5 " Robin Dapp
2017-05-31 13:56 ` Christophe Lyon
2017-05-31 14:37 ` Robin Dapp
2017-05-31 14:49 ` Christophe Lyon
2017-05-23 16:02 ` [PATCH 3/5 " Robin Dapp
2017-05-11 11:59 ` [PATCH 5/5] " Robin Dapp
2017-05-08 16:13 ` [PATCH 3/4] " Robin Dapp
2017-05-09 10:41 ` Richard Biener
2017-05-08 16:27 ` [PATCH 4/4] " Robin Dapp
2017-05-09 10:55 ` Richard Biener
2017-05-04 9:07 ` [PATCH 2/3] " Robin Dapp
2017-05-05 10:37 ` Richard Biener
2017-05-04 9:14 ` [PATCH 3/3] " Robin Dapp
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=87inkqzo5y.fsf@linaro.org \
--to=richard.sandiford@linaro.org \
--cc=amker.cheng@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=krebbel@linux.vnet.ibm.com \
--cc=rdapp@linux.vnet.ibm.com \
--cc=richard.guenther@gmail.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).