All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Michel Dänzer" <michel@daenzer.net>
To: "Dieter Nützel" <Dieter@nuetzel-hh.de>
Cc: Dri Devel <dri-devel@lists.freedesktop.org>
Subject: Re: LLVM / Clang git for libclc: optimized build or debug?
Date: Tue, 27 Jan 2015 12:23:24 +0900	[thread overview]
Message-ID: <54C704AC.20804@daenzer.net> (raw)
In-Reply-To: <e0d838ba67975b46c99ead293cfafca6@mail.ud03.udmedia.de>

On 27.01.2015 04:20, Dieter Nützel wrote:
> make[1]: Leaving directory '/opt/llvm/bindings'
> llvm[0]: ***** Completed Debug+Asserts Build
> llvm[0]: ***** Note: Debug build can be 10 times slower than an
> llvm[0]: ***** optimized build. Use make ENABLE_OPTIMIZED=1 to
> llvm[0]: ***** make an optimized build. Alternatively you can
> llvm[0]: ***** configure with --enable-optimized.
> 304.322u 15.792s 1:50.25 290.3% 0+0k 6156432+1518784io 942pf+0w

FWIW, I use these debug vs. optimization related options for building LLVM:

'--enable-optimized' '--with-optimize-option=-O2  -ftree-vectorize
-march=amdfam10 -mtune=native' '--enable-assertions'
'--enable-debug-runtime' '--enable-debug-symbols'
'CPPFLAGS=-fno-omit-frame-pointer' 'CFLAGS=-fno-omit-frame-pointer'
'CXXFLAGS=-fno-omit-frame-pointer'

That gives good performance while preserving assertions and debugging
symbols. -march=amdfam10 (as opposed to -march=native or a newer AMD CPU
family) avoids generating machine code which valgrind can't handle, and
-fno-omit-frame-pointer allows profile call graphs to make sense.


> Second:
> 'make -j 4' do NOT work in the first run. --- Known?

Not to me.


-- 
Earthling Michel Dänzer               |               http://www.amd.com
Libre software enthusiast             |             Mesa and X developer
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

      reply	other threads:[~2015-01-27  3:23 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-01-26 19:20 LLVM / Clang git for libclc: optimized build or debug? Dieter Nützel
2015-01-27  3:23 ` Michel Dänzer [this message]

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=54C704AC.20804@daenzer.net \
    --to=michel@daenzer.net \
    --cc=Dieter@nuetzel-hh.de \
    --cc=dri-devel@lists.freedesktop.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.