From: Bernd Schmidt <bschmidt@redhat.com>
To: Abe <abe_skolnik@yahoo.com>
Cc: Sebastian Pop <sebpop@gmail.com>,
Kyrill Tkachov <kyrylo.tkachov@arm.com>,
"gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>
Subject: Re: using scratchpads to enhance RTL-level if-conversion: revised patch
Date: Tue, 13 Oct 2015 20:16:00 -0000 [thread overview]
Message-ID: <561D66AB.9090003@redhat.com> (raw)
In-Reply-To: <561D5CC4.8030502@yahoo.com>
> _Potentially_ so, yes. However, GCC is free to put the allocation into
> an otherwise-unused part of the stack frame.
Well, I looked at code generation changes, and it usually seems to come
with an increase in stack frame size - sometimes causing extra
instructions to be emitted.
>> However, why do we need to allocate anything in the first place?
> > If you want to store something that will be thrown away,
> > just pick an address below the stack pointer.
>
> Because allocating a scratchpad should work on all relevant targets. We
> do not have the resources to test on all GCC-supported
> CPU ISAs and on all GCC-supported OSes, and we would like to have an
> optimization that works on as many targets as makes sense
> [those with cmove-like ability and withOUT full-blown conditional
> execution].
Yeah, but if you put in a new facility like this, chances are
maintainers for active targets will pick it up and add the necessary
hooks. That's certainly what happened with shrink-wrapping. So I don't
think this is a concern.
> I agree that your suggestion of having one global default scratchpad
> allocation policy plus per-target
> overrides that are more efficient _is_ a good one, but it will have to
> wait a while for implementation
> if that`s to be done by me. In the meantime, the existing allocation
> policy is compatible with
> multiple targets and costs very little space in the stack frame, if and
> when any at all.
I'm afraid I'll have to reject the patch then, on these grounds:
* it may pessimize code
* it does not even estimate costs to attempt avoiding this
* a much simpler, more efficient implementation is possible.
>>> + MEM_NOTRAP_P (mem) = true;
>> So I'm still not entirely sure which cases you are trying to optimize
>> and which ones not,
>
> The current patch focuses entirely on half-hammock writes with stores to
> addresses
> about which GCC "feels nervous", i.e. "may trap or fault"; for example:
>
> if (condition)
> *pointer = 9;
> // no "else" or "else if"
>
>
>> but couldn't this technique allow a trapping store here?
>
> The purpose of the new if-conversion is to take a may-trap-or-fault
> store and replace it with a store
> that will be OK if the original program was OK with respect to the
> current execution`s inputs,
> environment, PRNG results, etc. For example, the only way the
> if-converted code would dereference a
> null pointer is if/when the original program would have done the same
> thing under the same conditions.
Yeah, but it could still trap if the original program had an error. So I
don't think setting MEM_NOTRAP_P is right.
Bernd
next prev parent reply other threads:[~2015-10-13 20:16 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-07 23:29 Abe
2015-10-08 13:09 ` Sebastian Pop
2015-10-08 13:20 ` Sebastian Pop
2015-10-08 13:26 ` Bernd Schmidt
2015-10-08 13:23 ` Bernd Schmidt
2015-10-13 19:34 ` Abe
2015-10-13 20:16 ` Bernd Schmidt [this message]
2015-10-14 17:43 ` Jeff Law
2015-10-14 19:15 ` Bernd Schmidt
2015-10-15 8:52 ` Richard Biener
2015-10-20 5:52 ` Jeff Law
2015-10-20 9:37 ` Richard Biener
2015-10-14 1:05 ` Richard Henderson
2015-10-14 1:11 ` Richard Henderson
2015-10-14 8:29 ` Eric Botcazou
2015-10-14 17:46 ` Jeff Law
2015-10-13 20:05 Abe
[not found] <024301d11106$2379b5f0$6a6d21d0$@samsung.com>
2015-10-27 23:02 ` Abe
2015-10-30 14:09 ` Bernd Schmidt
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=561D66AB.9090003@redhat.com \
--to=bschmidt@redhat.com \
--cc=abe_skolnik@yahoo.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=kyrylo.tkachov@arm.com \
--cc=sebpop@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).