Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Glibc overlay with amd64 performance patches! (obsolete)
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 20, 21, 22, 23  Next  
Reply to topic    Gentoo Forums Forum Index Unsupported Software
View previous topic :: View next topic  
Author Message
nesl247
Veteran
Veteran


Joined: 15 Jun 2004
Posts: 1614
Location: Florida

PostPosted: Tue Feb 07, 2006 10:00 pm    Post subject: Reply with quote

Here is mine:

Code:
alex@alex32 ~ $ ./a.out 2200 1000 1048576
Memory to memory copy rate = 3343.110352 MBytes / sec. Block size = 1048576.
Back to top
View user's profile Send private message
teh-steve
n00b
n00b


Joined: 28 Jan 2005
Posts: 27
Location: Worcester

PostPosted: Tue Feb 07, 2006 10:33 pm    Post subject: Reply with quote

Hm, changing my CHOST to i686 did indeed fix the problem. Strange that I didn't notice it before. I suppose that I thought the ebuild was failing to pick up on my actual CHOST and defaulting to i386. Anyway, works great now. Thanks. :-)
Back to top
View user's profile Send private message
TripleM666
n00b
n00b


Joined: 09 Feb 2006
Posts: 11

PostPosted: Thu Feb 09, 2006 4:41 am    Post subject: Reply with quote

i have experienced something really weird with this little benchmark memcpy. i use your glibc-overlay (2.3.6-r2). at the moment my gentoo runs in a vm on xp x64, and with heavy load (recompiling some kde-stuff). the first time, i used memcpy (without -O?), i got some result, as i expected:
Code:
$ ./memcpy 2200 1000 1048576
Memory to memory copy rate = 1269.119141 MBytes / sec. Block size = 1048576.
after that i compiled it with -O1, as someone recommended:
Code:
$ ./memcpy 2200 1000 1048576
Memory to memory copy rate = 4394.828125 MBytes / sec. Block size = 1048576.

what? 8O
-Os shows the similar result as without -O?
-O2/O3 similar to O1

how can it be that fast? in a virtual machine and under load


so long,
Back to top
View user's profile Send private message
nxsty
Veteran
Veteran


Joined: 23 Jun 2004
Posts: 1556
Location: .se

PostPosted: Thu Feb 09, 2006 8:48 am    Post subject: Reply with quote

TripleM666 wrote:
how can it be that fast? in a virtual machine and under load


I've no idea. :) I don't know if that program is very accurate either, I just got it of a suse mailinglist. But it's intel CPU right? Those usually give the dubble result over amd.
Back to top
View user's profile Send private message
Config
Retired Dev
Retired Dev


Joined: 25 May 2003
Posts: 187
Location: Zurich, Switzerland

PostPosted: Thu Feb 09, 2006 4:00 pm    Post subject: Reply with quote

Is there something preventing the inclusion of the madvise-glibc patch?
_________________
Config - caught by a chronic disease called tuxmania....
Back to top
View user's profile Send private message
TripleM666
n00b
n00b


Joined: 09 Feb 2006
Posts: 11

PostPosted: Thu Feb 09, 2006 6:25 pm    Post subject: Reply with quote

nxsty wrote:
TripleM666 wrote:
how can it be that fast? in a virtual machine and under load


I've no idea. :) I don't know if that program is very accurate either, I just got it of a suse mailinglist. But it's intel CPU right? Those usually give the dubble result over amd.
no, its an athlon64 3000+. i think, the virtual machine of vmware does some strange memory magic :lol: , cause i have redone the test under real life conditions:
Code:
$ ./memcpy 2200 1000 1048576
Memory to memory copy rate = 2383.830811 MBytes / sec. Block size = 1048576.
i have choosen 2200 as freq.para to better compare the results.

btw. my emerge --info:
Code:
Gentoo Base System version 1.12.0_pre15
Portage 2.1_pre4-r1 (default-linux/amd64/2005.1, gcc-4.1.0-beta20060127, glibc-2.3.6-r2, 2.6.15 x86_64)
=================================================================
System uname: 2.6.15 x86_64 AMD Athlon(tm) 64 Processor 3000+
dev-lang/python:     2.3.5, 2.4.2-r1
sys-apps/sandbox:    1.2.17
sys-devel/autoconf:  2.13, 2.59-r7
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1
sys-devel/binutils:  2.16.1-r1
sys-devel/libtool:   1.5.22
virtual/os-headers:  2.6.11-r3
ACCEPT_KEYWORDS="amd64 ~amd64"
AUTOCLEAN="yes"
CBUILD="x86_64-pc-linux-gnu"
CFLAGS="-march=athlon64 -O2 -fweb -frename-registers -fomit-frame-pointer -fno-ident -pipe"
CHOST="x86_64-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/kde/3/share/config /usr/lib/X11/xkb /usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-march=athlon64 -O2 -fweb -frename-registers -fomit-frame-pointer -fno-ident -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig distlocks sandbox sfperms strict"
GENTOO_MIRRORS="ftp://ftp.tu-clausthal.de/pub/linux/gentoo/ ftp://sunsite.informatik.rwth-aachen.de/pub/Linux/gentoo ftp://linux.rz.ruhr-uni-bochum.de/gentoo-mirror/ ftp://ftp.uni-erlangen.de/pub/mirrors/gentoo ftp://ftp.wh2.tu-dresden.de/pub/mirrors/gentoo ftp://ftp-stud.fht-esslingen.de/pub/Mirrors/gentoo/"
MAKEOPTS="-j1"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY="/usr/local/portage"
SYNC="rsync://rsync.de.gentoo.org/gentoo-portage"
USE="amd64 X alsa arts avi berkdb bitmap-fonts crypt cups dvd eds emboss encode foomaticdb fortran gif gpm gstreamer imlib ipv6 jpeg kde lzw lzw-tiff mp3 mpeg ncurses nls nptl nptlonly opengl pam pdflib perl png python qt quicktime readline sdl spell ssl tcpd threads tiff truetype-fonts type1-fonts unicode usb userlocales xpm xv zlib elibc_glibc kernel_linux userland_GNU"
Unset:  ASFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, LANG, LC_ALL, LDFLAGS, LINGUAS

so, no probs with gcc4.1 here


so long,
Back to top
View user's profile Send private message
xbb
Tux's lil' helper
Tux's lil' helper


Joined: 26 Nov 2005
Posts: 85

PostPosted: Thu Feb 09, 2006 7:34 pm    Post subject: Reply with quote

I can't start openoffice (compiled with the flags below), it crashes while it's loading, only when LD_BIND_DIRECT is set. Any suggestions?

Code:

CFLAGS="-O2 -mtune=prescott -march=prescott -msse3 -pipe -fno-strict-aliasing"
CXXFLAGS="-O2 -mtune=prescott -march=prescott -msse3 -pipe -fvisibility-inlines-hidden -fno-strict-aliasing"
LDFLAGS="-Wl,-O1 -Wl,--sort-common -Wl,--enable-new-dtags -Wl,-Bdirect -Wl,-hashvals -Wl,-zdynsort"


Last edited by xbb on Fri Feb 10, 2006 3:37 am; edited 1 time in total
Back to top
View user's profile Send private message
n0rbi666
l33t
l33t


Joined: 04 Mar 2005
Posts: 703
Location: Poland -> Krakow

PostPosted: Thu Feb 09, 2006 10:48 pm    Post subject: Reply with quote

Every time, when I emerge
Code:
sys-libs/glibc-2.3.90.20060207  USE="gcc4ssp glibc-compat20 glibc-omitfp linuxthreads-tls nls nptl nptlonly pic userlocales -build -erandom -hardened -nomalloccheck -profile

Everything goes fine, but when I want to compile anything, I've got following error:
Code:
checking whether the C compiler (gcc -march=athlon-xp -O2 -pipe ) works... no
configure: error: installation or configuration problem: C compiler cannot create executables.

!!! ERROR: sys-devel/gcc-4.1.0_beta20060203 failed.
Call stack:
  ebuild.sh, line 1894:   Called dyn_compile
  ebuild.sh, line 941:   Called src_compile
  ebuild.sh, line 1609:   Called toolchain_src_compile
  toolchain.eclass, line 24:   Called gcc_src_compile
  toolchain.eclass, line 1456:   Called gcc_do_configure

Code:

/usr/lib/gcc/i686-pc-linux-gnu/4.1.0-beta20060203/../../../../ i686-pc-linux-gnu/bin/ld:/usr/lib/gcc/i686-pc-linux-gnu/4.1.0-beta20060203/../../../libc.so: file format not recognized; treating as linker script
/usr/lib/gcc/i686-pc-linux-gnu/4.1.0-beta20060203/../../../../ i686-pc-linux-gnu/bin/ld:/usr/lib/gcc/i686-pc-linux-gnu/4.1.0-beta20060203/../../../libc.so:5: syntax error
collect2: ld returned 1 exit status
make: *** [macierz] B³±d 1


Code:
n0rbi ~ # /usr/lib/libc.so
/usr/lib/libc.so: line 1: /bin: is a directory
/usr/lib/libc.so: line 2: Use: command not found
/usr/lib/libc.so: line 3: the: command not found
/usr/lib/libc.so: line 4: syntax error near unexpected token `elf32-i386'
/usr/lib/libc.so: line 4: `OUTPUT_FORMAT(elf32-i386)


My emerge info :
Code:
Portage 2.1_pre4-r1 (default-linux/x86/2005.1, gcc-4.1.0-beta20060203, glibc-2.3.90.20060207-r0, 2.6.15-gentoo-r2 i686)
=================================================================
System uname: 2.6.15-gentoo-r2 i686 AMD Athlon(tm)
Gentoo Base System version 1.12.0_pre15
dev-lang/python:     2.3.5-r2, 2.4.2-r1
sys-apps/sandbox:    1.2.17
sys-devel/autoconf:  2.13, 2.59-r7
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1
sys-devel/binutils:  2.15.92.0.2-r10, 2.16.1-r1
sys-devel/libtool:   1.5.22
virtual/os-headers:  2.6.11-r4
ACCEPT_KEYWORDS="x86 ~x86"
AUTOCLEAN="yes"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-march=athlon-xp -Os -fomit-frame-pointer -ffast-math -ftracer -pipe -s -DNDEBUG -DG_DISABLE_ASSERT -DNO_DEBUG -DG_DISABLE_CHECKS -DG_DISABLE_CAST_CHECKS"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/kde/3/share/config /usr/share/X11/xkb /usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-march=athlon-xp -Os -fomit-frame-pointer -ffast-math -ftracer -pipe -s -DNDEBUG -DG_DISABLE_ASSERT -DNO_DEBUG -DG_DISABLE_CHECKS -DG_DISABLE_CAST_CHECKS -fno-enforce-eh-specs -fvisibility-inlines-hidden"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig distlocks sandbox sfperms strict"
GENTOO_MIRRORS="http://distfiles.gentoo.org http://distro.ibiblio.org/pub/linux/distributions/gentoo"
LANG="pl_PL"
LC_ALL="pl_PL"
LDFLAGS="-Wl,-O1 -Wl,--sort-common -Wl,--as-needed  -Wl,--enable-new-dtags -s"
LINGUAS="pl"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY="/usr/local/portage"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="x86 3dnow 3dnowext S3TC X a52 aac aalib acpi alsa amarok apm artswrappersuid asf audiofile bash-completion berkdb bitmap-fonts bl bmp browserplugin bzip2 c++ cdda cddb cdparanoia clanVoice cpudetection crypt css cups custom-cflags dbus dri dvd dvdr dvdread eds emboss encode expat fame fat firefox flac foomaticdb fortran gdbm gg gif gimpprint glibc-compat20 glibc-omitfp glx gpm gtk gtk2 hpn id3 idn ieee1394 imlib insecure-drivers ithreads jack java javascript jp2 jpeg jpeg2k kde kdeenablefinal kdepim lame libg++ libvisual libwww linuxthreads-tls lirc lm_sensors lzo lzw mad matroska mikmod mjpeg mmx mmxext mng motif mozcalendar mozdevelop mozilla mozsvg mp3 mpeg mpeg2 mpeg4 mplayer multislot musepack ncurses nforce2 nls nptl nptlonly ntfs ogg oggvorbis opengl openssh pam pcre pdflib perl pic pmu png python qt quicktime rar readline real rtc sblive sdl sensord speech spell sse ssl svga tcltk tcpd threads tiff transcode truetype truetype-fonts type1-fonts usb userlocales v4l v4l2 vcd vorbis win32codecs wma123 xinerama xml2 xmms xv zlib elibc_glibc input_devices_keyboard input_devices_mouse input_devices_evdev kernel_linux linguas_pl userland_GNU video_cards_ati"
Unset:  ASFLAGS, CTARGET, EMERGE_DEFAULT_OPTS


I tried with gcc 3.4.5 and latest 4.1beta - no success
Have I missed something ? :roll:

--add--

Ok, i Solved it :)
I was using old binutils :oops: I forgot about it :?

BTW. Great ebuild :)
_________________
Trying to forget (Wasting my time)
We're falling right through
Lying to forget (Telling more lies)
We're raising our truth
Back to top
View user's profile Send private message
xbb
Tux's lil' helper
Tux's lil' helper


Joined: 26 Nov 2005
Posts: 85

PostPosted: Fri Feb 10, 2006 3:01 pm    Post subject: Reply with quote

xbb wrote:
I can't start openoffice (compiled with the flags below), it crashes while it's loading, only when LD_BIND_DIRECT is set. Any suggestions?

Code:

CFLAGS="-O2 -mtune=prescott -march=prescott -msse3 -pipe -fno-strict-aliasing"
CXXFLAGS="-O2 -mtune=prescott -march=prescott -msse3 -pipe -fvisibility-inlines-hidden -fno-strict-aliasing"
LDFLAGS="-Wl,-O1 -Wl,--sort-common -Wl,--enable-new-dtags -Wl,-Bdirect -Wl,-hashvals -Wl,-zdynsort"


I have tried also with openoffice-bin (while LD_BIND_DIRECT=1 is set). It starts ok but when I exit the program it crashes and says in console:
Code:

*** glibc detected *** double free or corruption (out): 0xa34d3850 ***
KCrash: Application 'soffice.bin' crashing...


If I start it with --widgets-set gtk, starts and close ok without errors, while when I compiled it by myself it was crashing at start also with gtk widgets.

While LD_BIND_DIRECT is not set, no crashes, no errors with both kde and gtk widgets.
Back to top
View user's profile Send private message
FireBurn
Apprentice
Apprentice


Joined: 19 Sep 2004
Posts: 165
Location: Edinburgh, UK

PostPosted: Sat Feb 11, 2006 11:16 am    Post subject: Reply with quote

Has anyone noticed problems with compiling using --as-needed with the experimental glibc 2.4 and gcc 4.1?

I've been having issues with anything that tries to link against libcairo.so with the error:

i686-pc-linux-gnu-gcc -pedantic -Wall -Wstrict-prototypes -Wnested-externs -std=c89 -U__STRICT_ANSI__ -Wl,-O1 -Wl,--sort-common -Wl,--enable-new-dtags -Wl,-Bdirect -Wl,-hashvals -Wl,-zdynsort -Wl,--as-needed -L/usr/lib -o xscreensaver-demo-Gtk prefs.o dpms.o remote.o exec.o ../utils/resources.o ../utils/usleep.o ../utils/visual.o demo-Gtk.o demo-Gtk-conf.o -L/usr/lib \
-Wl,--export-dynamic -lglade-2.0 -lgtk-x11-2.0 -lxml2 -lz -lgdk-x11-2.0 -latk-1.0 -lpangocairo-1.0 -lpango-1.0 -lcairo -lgdk_pixbuf_xlib-2.0 -lgdk_pixbuf-2.0 -lm -lgobject-2.0 -lgmodule-2.0 -ldl -lglib-2.0 -lXmu -lXpm -lgdk_pixbuf_xlib-2.0 -lgdk_pixbuf-2.0 -lm -lgobject-2.0 -lgmodule-2.0 -ldl -lglib-2.0 -lSM -lICE \
-lXt -lX11 -lXext
/usr/lib/libcairo.so.2: undefined reference to `__popcountsi2@GCC_3.4'
collect2: ld returned 1 exit status
make[1]: *** [xscreensaver-demo-Gtk] Error 1
make[1]: Leaving directory `/var/tmp/portage/xscreensaver-4.23-r1/work/xscreensaver-4.23/driver'
make: *** [default] Error 5

I used to use this flag with gcc-4.1 before moving to the new glibc and if the flag is removed it works just fine

Any one have any ideas or even better patches?
Back to top
View user's profile Send private message
PrakashP
Veteran
Veteran


Joined: 27 Oct 2003
Posts: 1246
Location: C.C.A.A., Germania

PostPosted: Sun Feb 12, 2006 3:09 am    Post subject: Reply with quote

I noticed a big fat performance issue with the 2.3.90 snapshots. On AMD64 and x86 using a SAT solver performance is nearly halfed. :( Any idea? Is some debugging code built? 2.3.6-r2 from the overlay (I hope) doesn't seem to be affected.
Back to top
View user's profile Send private message
rhill
Developer
Developer


Joined: 22 Oct 2004
Posts: 1629
Location: sk.ca

PostPosted: Sun Feb 12, 2006 3:29 pm    Post subject: Reply with quote

Code:
[06:31] <CIA-2> azarah * gentoo/src/patchsets/glibc/2.3.6/ (16 files):
[06:31] <CIA-2> Add -Bdirect and amd64 string/math patches among other things.
[06:31] <CIA-2> Many thanks to Simon Strandman for his work on maintaining the patchset for 2.3.90
[06:31] <CIA-2> I used to port these patches to 2.3.6.


:o
_________________
by design, by neglect
for a fact or just for effect
Back to top
View user's profile Send private message
nxsty
Veteran
Veteran


Joined: 23 Jun 2004
Posts: 1556
Location: .se

PostPosted: Sun Feb 12, 2006 3:45 pm    Post subject: Reply with quote

dirtyepic wrote:
Code:
[06:31] <CIA-2> azarah * gentoo/src/patchsets/glibc/2.3.6/ (16 files):
[06:31] <CIA-2> Add -Bdirect and amd64 string/math patches among other things.
[06:31] <CIA-2> Many thanks to Simon Strandman for his work on maintaining the patchset for 2.3.90
[06:31] <CIA-2> I used to port these patches to 2.3.6.


:o


Nice! :P
Back to top
View user's profile Send private message
duby2291
Guru
Guru


Joined: 17 Oct 2004
Posts: 426

PostPosted: Mon Feb 13, 2006 5:21 am    Post subject: Reply with quote

No problems here with me. I'm running ~amd64 and everything seems to work great. Although I think the speed ups are very subjective. The systems does indeed "feel" faster, but that may very well be becouse I've been told that it is. Mind over matter and all.....

With that said though I am using bdirect and the libm and string patches without a single hiccup here.... Thanks for all the effort nxsty :D

edit.... Also worth noting that I'm using openoffice-bin..... not compiled with bdirect...... and it opens in literally less than a second...... Used to take like ten seconds or more.....
Back to top
View user's profile Send private message
PrakashP
Veteran
Veteran


Joined: 27 Oct 2003
Posts: 1246
Location: C.C.A.A., Germania

PostPosted: Mon Feb 13, 2006 9:50 am    Post subject: Reply with quote

Well, I'll ask whether I can give out the sources for the solver, than you can try yourself. I'll also provide a static compile with goog glibc, so you can see the difference.
Back to top
View user's profile Send private message
immudium
Guru
Guru


Joined: 12 Oct 2004
Posts: 332
Location: Utah

PostPosted: Mon Feb 13, 2006 6:36 pm    Post subject: Reply with quote

PrakashP wrote:
Well, I'll ask whether I can give out the sources for the solver, than you can try yourself. I'll also provide a static compile with goog glibc, so you can see the difference.


I'd be interested in giving that a try. I have three Pentium 4 machines all very close in hardware specs. Two of them have the patched 2.3.90 glibc and one has an unpatched 2.3.6-r2 glibc. The unpatched Pentium 4 kicks the other two machines upside down and sideways in nbench scores. I am currently working on getting one of the two 2.3.90 machines switched to nxsty's 2.3.6 patched glibc so that I can compare performance between gentoo's 2.3.6 glibc, nxsty's patched 2.3.6 and the patched 2.3.90 glibc. I'm hoping 2.3.90 may also explain why Java has been running so pathetically crappy for me as well.
Back to top
View user's profile Send private message
mbar
Veteran
Veteran


Joined: 19 Jan 2005
Posts: 1952
Location: Poland

PostPosted: Mon Feb 13, 2006 8:32 pm    Post subject: Reply with quote

What's this thing with JB_BP and mozilla-firefox 1.5.0.1-r1 in portage? I can't compile it. :?
Back to top
View user's profile Send private message
nesl247
Veteran
Veteran


Joined: 15 Jun 2004
Posts: 1614
Location: Florida

PostPosted: Mon Feb 13, 2006 8:38 pm    Post subject: Reply with quote

Look in the Conrad Install Guide: Dev Edition thread. I think at page 3 is the info on what to do. Then you just need to get the patch from one of the older overlays in the thread.. I would give it to you myself but I just booted to windows for some game time.
Back to top
View user's profile Send private message
nesl247
Veteran
Veteran


Joined: 15 Jun 2004
Posts: 1614
Location: Florida

PostPosted: Mon Feb 13, 2006 8:44 pm    Post subject: Reply with quote

Look in the Conrad Install Guide: Dev Edition thread. I think at page 3 is the info on what to do. Then you just need to get the patch from one of the older overlays in the thread.. I would give it to you myself but I just booted to windows for some game time.
Back to top
View user's profile Send private message
Cinder6
l33t
l33t


Joined: 05 Aug 2004
Posts: 767
Location: California

PostPosted: Tue Feb 14, 2006 1:38 am    Post subject: Reply with quote

What does -hashvals do? I assume it is also an LDFLAG (-Wl,-hashvals?), but I have only been able to read a little about it. Can it be used along with -Bdirect? If not, which is faster?
_________________
Knowledge is power.
Power corrupts.
Study hard.
Be evil.

Ugly Overload
Back to top
View user's profile Send private message
n0rbi666
l33t
l33t


Joined: 04 Mar 2005
Posts: 703
Location: Poland -> Krakow

PostPosted: Tue Feb 14, 2006 4:56 pm    Post subject: Reply with quote

Hi !

I've got another problem - can't compile libggi

Code:
if i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../include   -I/usr/include  -march=athlon-xp -Os -fomit-frame-pointer -ffast-math -ftracer -pipe -s -DNDEBUG -DG_DISABLE_ASSERT -DNO_DEBUG -DG_DISABLE_CHECKS -DG_DISABLE_CAST_CHECKS -I/usr/include -D_REENTRANT -D_THREAD_SAFE -MT showaccel2.o -MD -MP -MF ".deps/showaccel2.Tpo" -c -o showaccel2.o showaccel2.c; \
then mv -f ".deps/showaccel2.Tpo" ".deps/showaccel2.Po"; else rm -f ".deps/showaccel2.Tpo"; exit 1; fi
speed.c: In function `time_offset':
speed.c:57: error: `CLK_TCK' undeclared (first use in this function)
speed.c:57: error: (Each undeclared identifier is reported only once
speed.c:57: error: for each function it appears in.)
speed.c: In function `time_stop':
speed.c:70: error: `CLK_TCK' undeclared (first use in this function)
make[3]: *** [speed.o] Error 1
make[3]: *** Waiting for unfinished jobs....
make[3]: Leaving directory `/var/tmp/portage/libggi-2.1.1/work/libggi-2.1.1/programs/check'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory `/var/tmp/portage/libggi-2.1.1/work/libggi-2.1.1/programs'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/var/tmp/portage/libggi-2.1.1/work/libggi-2.1.1'
make: *** [all] Error 2


On glibc-2.3.6 from portage - it compilled without problems ... But I cant downgrade now :(
_________________
Trying to forget (Wasting my time)
We're falling right through
Lying to forget (Telling more lies)
We're raising our truth
Back to top
View user's profile Send private message
immudium
Guru
Guru


Joined: 12 Oct 2004
Posts: 332
Location: Utah

PostPosted: Tue Feb 14, 2006 6:11 pm    Post subject: Reply with quote

n0rbi666 wrote:
I've got another problem - can't compile libggi


Easily fixed. They're simply using a deprecated function call. Here's a patch.
Code:
--- programs/check/speed.c   2004-09-08 13:26:22.000000000 -0600
+++ programs/check/speed.c   2006-02-14 10:56:40.266332480 -0700
@@ -22,7 +22,7 @@
 #include <string.h>
 #include <unistd.h>
 #include <sys/times.h>
-#ifndef CLK_TCK
+#ifndef CLOCKS_PER_SEC
 #include <time.h>
 #endif
 
@@ -54,8 +54,8 @@
 {
    struct tms end;
    times(&end);
-   u_time = (double) -(end.tms_utime - timer.tms_utime) / CLK_TCK;
-   s_time = (double) -(end.tms_stime - timer.tms_stime) / CLK_TCK;
+   u_time = (double) -(end.tms_utime - timer.tms_utime) / CLOCKS_PER_SEC;
+   s_time = (double) -(end.tms_stime - timer.tms_stime) / CLOCKS_PER_SEC;
    time_start();
 }
 
@@ -67,8 +67,8 @@
 {
    struct tms end;
    times(&end);
-   u_time += (double) (end.tms_utime - timer.tms_utime) / CLK_TCK;
-   s_time += (double) (end.tms_stime - timer.tms_stime) / CLK_TCK;
+   u_time += (double) (end.tms_utime - timer.tms_utime) / CLOCKS_PER_SEC;
+   s_time += (double) (end.tms_stime - timer.tms_stime) / CLOCKS_PER_SEC;
 }
 
 /* The pixelvalue for the color white.


n0rbi666 wrote:
... But I cant downgrade now :(

Sure you can if you want to. It's a little bit of an extra pain, but you just have to copy a few files built against the older glibc like from a livecd and you're good to go.
Back to top
View user's profile Send private message
n0rbi666
l33t
l33t


Joined: 04 Mar 2005
Posts: 703
Location: Poland -> Krakow

PostPosted: Tue Feb 14, 2006 7:46 pm    Post subject: Reply with quote

immudium wrote:
n0rbi666 wrote:
I've got another problem - can't compile libggi


Easily fixed. They're simply using a deprecated function call. Here's a patch.

Uff it works :) big THX :)

Quote:

Sure you can if you want to. It's a little bit of an extra pain, but you just have to copy a few files built against the older glibc like from a livecd and you're good to go.

But I don't want to downgrade :)
_________________
Trying to forget (Wasting my time)
We're falling right through
Lying to forget (Telling more lies)
We're raising our truth
Back to top
View user's profile Send private message
PrakashP
Veteran
Veteran


Joined: 27 Oct 2003
Posts: 1246
Location: C.C.A.A., Germania

PostPosted: Tue Feb 14, 2006 9:18 pm    Post subject: Reply with quote

immudium wrote:
PrakashP wrote:
Well, I'll ask whether I can give out the sources for the solver, than you can try yourself. I'll also provide a static compile with goog glibc, so you can see the difference.


I'd be interested in giving that a try. I have three Pentium 4 machines all very close in hardware specs. Two of them have the patched 2.3.90 glibc and one has an unpatched 2.3.6-r2 glibc. The unpatched Pentium 4 kicks the other two machines upside down and sideways in nbench scores. I am currently working on getting one of the two 2.3.90 machines switched to nxsty's 2.3.6 patched glibc so that I can compare performance between gentoo's 2.3.6 glibc, nxsty's patched 2.3.6 and the patched 2.3.90 glibc. I'm hoping 2.3.90 may also explain why Java has been running so pathetically crappy for me as well.


Unfortunately the author doesn't want to release it just yet. (It is part of his diploma thesis.) But maybe the old solver will already exhibit the same behaviour and it was long released. (I can't test anymore as I ditched the 2.3.90 glibc versions.) I'll try to pack it together including static 32 bit and 64 compiles. The impatient may search for "max böhm sat"... ;)
Back to top
View user's profile Send private message
piwacet
Guru
Guru


Joined: 30 Dec 2004
Posts: 486

PostPosted: Wed Feb 15, 2006 7:13 am    Post subject: Reply with quote

dirtyepic wrote:
Code:
[06:31] <CIA-2> azarah * gentoo/src/patchsets/glibc/2.3.6/ (16 files):
[06:31] <CIA-2> Add -Bdirect and amd64 string/math patches among other things.
[06:31] <CIA-2> Many thanks to Simon Strandman for his work on maintaining the patchset for 2.3.90
[06:31] <CIA-2> I used to port these patches to 2.3.6.


:o


Err - pardon my ignorance, but what does this mean?

TIA
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Unsupported Software All times are GMT
Goto page Previous  1, 2, 3 ... 20, 21, 22, 23  Next
Page 21 of 23

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum