Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Emerge -e world on PII 233@291
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Documentation, Tips & Tricks
View previous topic :: View next topic  
Author Message
Roderik
n00b
n00b


Joined: 06 Jun 2002
Posts: 55
Location: Belgium

PostPosted: Tue Sep 03, 2002 9:15 am    Post subject: Emerge -e world on PII 233@291 Reply with quote

Ever wondered how long compiling would take on a PII 233@291 with 128MB ram?

Code:

Talt0s root # cat /proc/version
Linux version 2.4.19-gentoo-r7 (root@(none)) (gcc version 3.1) #1 Tue Jul 23 20:15:17 CEST 2002



Code:

Talt0s root # emerge -ep world

These are the packages that I would merge, in order.

Calculating world dependencies ...done!
[ebuild  N   ] sys-apps/cronbase-0.2.1 to /
[ebuild  N   ] sys-apps/groff-1.17.2-r2 to /
[ebuild  N   ] sys-apps/man-1.5k to /
[ebuild  N   ] sys-apps/man-pages-1.52 to /
[ebuild  N   ] sys-devel/libtool-1.4.1-r9 to /
[ebuild  N   ] sys-devel/gettext-0.11.1 to /
[ebuild  N   ] sys-devel/m4-1.4p to /
[ebuild  N   ] sys-libs/db-1.85-r1 to /
[ebuild  N   ] sys-libs/db-3.2.9-r1 to /
[ebuild  N   ] sys-libs/gdbm-1.8.0-r5 to /
[ebuild  N   ] sys-devel/perl-5.6.1-r6 to /
[ebuild  N   ] sys-devel/autoconf-2.53a to /
[ebuild  N   ] sys-libs/pwdb-0.61-r3 to /
[ebuild  N   ] dev-libs/glib-1.2.10-r4 to /
[ebuild  N   ] sys-devel/automake-1.6.1-r6 to /
[ebuild  N   ] sys-apps/miscfiles-1.3 to /
[ebuild  N   ] sys-libs/cracklib-2.7-r5 to /
[ebuild  N   ] sys-libs/pam-0.75-r7 to /
[ebuild  N   ] sys-apps/shadow-4.0.3 to /
[ebuild  N   ] sys-apps/pam-login-3.6-r2 to /
[ebuild  N   ] sys-apps/bzip2-1.0.2-r2 to /
[ebuild  N   ] sys-libs/ncurses-5.2.20020511-r1 to /
[ebuild  N   ] sys-apps/util-linux-2.11u to /
[ebuild  N   ] sys-apps/grep-2.5-r1 to /
[ebuild  N   ] sys-apps/reiserfsprogs-3.6.2 to /
[ebuild  N   ] sys-libs/zlib-1.1.4 to /
[ebuild  N   ] sys-apps/textutils-2.0.21 to /
[ebuild  N   ] sys-apps/kbd-1.06-r1 to /
[ebuild  N   ] sys-apps/netkit-base-0.17-r6 to /
[ebuild  N   ] dev-python/python-fchksum-1.6.1 to /
[ebuild  N   ] sys-apps/bash-2.05a-r3 to /
[ebuild  N   ] sys-libs/readline-4.2a to /
[ebuild  N   ] dev-libs/expat-1.95.4 to /
[ebuild  N   ] dev-lang/python-2.2.1-r2 to /
[ebuild  N   ] sys-apps/debianutils-1.16 to /
[ebuild  N   ] sys-apps/fileutils-4.1.8-r2 to /
[ebuild  N   ] sys-apps/portage-2.0.34 to /
[ebuild  N   ] sys-devel/binutils-2.12.90.0.7 to /
[ebuild  N   ] sys-devel/patch-2.5.4-r4 to /
[ebuild  N   ] sys-devel/flex-2.5.4a-r4 to /
[ebuild  N   ] sys-apps/tcp-wrappers-7.6-r4 to /
[ebuild  N   ] dev-libs/openssl-0.9.6g to /
[ebuild  N   ] net-misc/openssh-3.4_p1-r3 to /
[ebuild  N   ] sys-libs/lib-compat-1.0-r2 to /
[ebuild  N   ] sys-devel/ld.so-1.9.11-r2 to /
[ebuild  N   ] sys-apps/gawk-3.1.1 to /
[ebuild  N   ] sys-apps/gzip-1.3.2 to /
[ebuild  N   ] sys-devel/bin86-0.15.5 to /
[ebuild  N   ] sys-apps/sed-3.02.80-r4 to /
[ebuild  N   ] sys-apps/grub-0.90-r7 to /
[ebuild  N   ] sys-apps/setserial-2.17-r2 to /
[ebuild  N   ] sys-devel/make-3.79.1-r4 to /
[ebuild  N   ] sys-apps/sh-utils-2.0.11-r5 to /
[ebuild  N   ] sys-devel/bc-1.06-r3 to /
[ebuild  N   ] sys-devel/bison-1.35 to /
[ebuild  N   ] sys-apps/texinfo-4.2-r5 to /
[ebuild  N   ] sys-apps/file-3.37 to /
[ebuild  N   ] sys-apps/cpio-2.4.2-r4 to /
[ebuild  N   ] sys-apps/iptables-1.2.6a-r2 to /
[ebuild  N   ] sys-apps/psmisc-19-r3 to /
[ebuild  N   ] sys-devel/gcc-3.1.1-r1 to /
[ebuild  N   ] net-misc/dhcpcd-1.3.20_p0-r1 to /
[ebuild  N   ] net-misc/rsync-2.5.5-r1 to /
[ebuild  N   ] sys-apps/procps-2.0.7-r6 to /
[ebuild  N   ] app-arch/ncompress-4.2.4 to /
[ebuild  N   ] sys-apps/tar-1.13.25-r2 to /
[ebuild  N   ] net-ftp/ftp-0.17-r1 to /
[ebuild  N   ] net-misc/netkit-telnetd-0.17-r3 to /
[ebuild  N   ] app-shells/sash-3.4-r5 to /
[ebuild  N   ] sys-apps/hdparm-5.2-r2 to /
[ebuild  N   ] sys-apps/sharutils-4.2.1-r5 to /
[ebuild  N   ] sys-apps/modutils-2.4.19 to /
[ebuild  N   ] sys-kernel/linux-headers-2.4.18 to /
[ebuild  N   ] sys-apps/baselayout-1.8.2 to /
[ebuild  N   ] sys-libs/glibc-2.2.5-r5 to /
[ebuild  N   ] sys-apps/findutils-4.1.7-r1 to /
[ebuild  N   ] sys-apps/e2fsprogs-1.27 to /
[ebuild  N   ] sys-libs/slang-1.4.5-r2 to /
[ebuild  N   ] app-editors/nano-1.0.8-r1 to /
[ebuild  N   ] net-misc/wget-1.8.2 to /
[ebuild  N   ] sys-apps/which-2.14 to /
[ebuild  N   ] sys-apps/diffutils-2.7.7 to /
[ebuild  N   ] sys-apps/devfsd-1.3.25 to /
[ebuild  N   ] sys-apps/less-376 to /
[ebuild  N   ] sys-apps/ed-0.2-r3 to /
[ebuild  N   ] sys-apps/raidtools-0.90-r2 to /
[ebuild  N   ] sys-apps/net-tools-1.60-r4 to /
[ebuild  N   ] sys-apps/fbset-2.1 to /


Code:

Talt0s root # cat /proc/meminfo
        total:    used:    free:  shared: buffers:  cached:
Mem:  129519616 123195392  6324224        0 16826368 76566528
Swap: 271425536 21233664 250191872
MemTotal:       126484 kB
MemFree:          6176 kB
MemShared:           0 kB
Buffers:         16432 kB
Cached:          66904 kB
SwapCached:       7868 kB
Active:          67388 kB
Inact_dirty:     27732 kB
Inact_clean:     12180 kB
Inact_target:    21460 kB
HighTotal:           0 kB
HighFree:            0 kB
LowTotal:       126484 kB
LowFree:          6176 kB
SwapTotal:      265064 kB
SwapFree:       244328 kB
Committed_AS:   115664 kB


Code:

Talt0s root # cat /proc/cpuinfo
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 3
model name      : Pentium II (Klamath)
stepping        : 3
cpu MHz         : 291.590
cache size      : 512 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 2
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov mmx
bogomips        : 575.48



With these optimalisations

Code:

CHOST="i686-pc-linux-gnu"
CFLAGS="-march=i686 -mmmx -fomit-frame-pointer -ffast-math -fexpensive-optimizations -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -fforce-addr -falign-functions=4 -O3 -pipe"
CXXFLAGS="-march=i686 -mmmx -fomit-frame-pointer -ffast-math -fexpensive-optimizations -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -fforce-addr -falign-functions=4 -O3 -pipe"




Reason i did this:

1. i was bored and it was only my 3rd pc so wasn't really important
2. wanted to see how long a complete system recompilation took

Result:

only 27 hours!!

[img:4e0db5b9fc]http://roderik.student.khleuven.be/images/diversen/cpu-talt0s-day.png[/img:4e0db5b9fc]
_________________
"Never argue with an idiot, they'll drag you down to their level and beat you with experience."
Back to top
View user's profile Send private message
giant
Tux's lil' helper
Tux's lil' helper


Joined: 01 Aug 2002
Posts: 107

PostPosted: Tue Sep 03, 2002 9:19 am    Post subject: Reply with quote

interesting :-)

How did you make the nice graph ?
I could throw in my P200 with 64 MB for comparrisson.
Back to top
View user's profile Send private message
Roderik
n00b
n00b


Joined: 06 Jun 2002
Posts: 55
Location: Belgium

PostPosted: Tue Sep 03, 2002 9:24 am    Post subject: Reply with quote

MRTG and the cpuinfo.pl from the contrib dir but there are a lot of other scripts for cpu info
_________________
"Never argue with an idiot, they'll drag you down to their level and beat you with experience."
Back to top
View user's profile Send private message
kirill
Apprentice
Apprentice


Joined: 01 Aug 2002
Posts: 183
Location: Finland

PostPosted: Tue Sep 03, 2002 11:03 am    Post subject: Reply with quote

I'm planning to upgrade my server's default-1.0-gcc3 to default-x86-1.4 soon (gcc3.1.1->3.2). Then emerge -e world

It's a P166MMX, 32MB, 64MB swap.
The bootstrap took 24h link

let's compare then. brmmmmmm need to setup mrtg first ;)
_________________
--kirill
Back to top
View user's profile Send private message
DArtagnan
l33t
l33t


Joined: 30 Apr 2002
Posts: 942
Location: Israel, Jerusalem

PostPosted: Tue Sep 03, 2002 12:10 pm    Post subject: Reply with quote

Roderik wrote:
MRTG and the cpuinfo.pl from the contrib dir but there are a lot of other scripts for cpu info


I saw u've used a lot of "crazy" optimisations...is faster now?
_________________
All for one and one for All
--

MACPRO machine...
Back to top
View user's profile Send private message
Roderik
n00b
n00b


Joined: 06 Jun 2002
Posts: 55
Location: Belgium

PostPosted: Tue Sep 03, 2002 2:24 pm    Post subject: Reply with quote

hehe i don't really know, i only use that pc to do some devving on apache/mysql/php :) and i've never run any benchmarks
_________________
"Never argue with an idiot, they'll drag you down to their level and beat you with experience."
Back to top
View user's profile Send private message
kyron
Apprentice
Apprentice


Joined: 26 Aug 2002
Posts: 198
Location: Montreal, Qc.

PostPosted: Tue Sep 03, 2002 2:34 pm    Post subject: Reply with quote

Woah! I am defenately stating this next time I do an emerge world!! (I'll make a script to collect all the data :-) )

I presently have a P4 1.6G 256MB RAM...took 11hours to compile a system with KDE and qt (with all that this implies :) )
_________________
M$ Windows: When in doubt, REBOOT
Linux: When in doubt, RTFM ;-)
Back to top
View user's profile Send private message
debian
n00b
n00b


Joined: 24 Jun 2002
Posts: 22

PostPosted: Mon Sep 09, 2002 2:03 am    Post subject: Re: Emerge -e world on PII 233@291 Reply with quote

Roderik wrote:
With these optimalisations

Code:

CHOST="i686-pc-linux-gnu"
CFLAGS="-march=i686 -mmmx -fomit-frame-pointer -ffast-math -fexpensive-optimizations -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -fforce-addr -falign-functions=4 -O3 -pipe"
CXXFLAGS="-march=i686 -mmmx -fomit-frame-pointer -ffast-math -fexpensive-optimizations -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -fforce-addr -falign-functions=4 -O3 -pipe"

I've a laptop of even lesser power and I recommend the following changes to your list:
Code:
add    -mpreferred-stack-boundary=4
use    -fforce-mem instead of -fforce-addr
use    -march=pentium2 instead of -march=i686
remove -fexpensive-optimizations (no matter how sexy the name is ^_^)

In fact some flags are *not* improving the speed and actually drag down the compilation and overall performance of the binaries created. My compilation time improved very much by changing the flags.

These flags are not giving out of guessing, in fact I've a benchmark script that test each optimization flag bit by bit.

I'll post the script and my example C[XX]FLAGS when I off work today.
Back to top
View user's profile Send private message
debian
n00b
n00b


Joined: 24 Jun 2002
Posts: 22

PostPosted: Mon Sep 09, 2002 2:10 am    Post subject: Re: Emerge -e world on PII 233@291 Reply with quote

debian wrote:
I've a laptop of even lesser power and I recommend the following changes to your list:
Code:
add    -mpreferred-stack-boundary=4
use    -fforce-mem instead of -fforce-addr
use    -march=pentium2 instead of -march=i686
remove -fexpensive-optimizations (no matter how sexy the name is ^_^)

In fact some flags are *not* improving the speed and actually drag down the compilation and overall performance of the binaries created. My compilation time improved very much by changing the flags.

These flags are not giving out of guessing, in fact I've a benchmark script that test each optimization flag bit by bit.

I'll post the script and my example C[XX]FLAGS when I off work today.


Oops, one thing to note. The changes above is for Roderik and those who has simlar processor only, because I happen to have a laptop of similar power.

E.g. -fexpensive-optimizations might work for you, and you'd like to use -mpreferred-stack-boundary=2 and -fforce-addr instead of -mpreferred-stack-boundary=4 and -fforce-mem. It's all depending on your processor.

The best you could do to optimize your optimization flags is to use the script I'll post tonight to determine best optmization flags suits your system. Stay tune.
Back to top
View user's profile Send private message
debian
n00b
n00b


Joined: 24 Jun 2002
Posts: 22

PostPosted: Tue Sep 10, 2002 1:53 am    Post subject: Optimized CFLAGS/CXXFLAGS Reply with quote

Hi, I decided not posting my version of the said script as I worried that not all part of the orignal is release under GPL. You may take a look here:
Code:
http://www.rocklinux.org/sources/misc/bench/

Note that you need to modify it to cater for the latest change in gcc 3.2.

Here are the benchmark results:

My laptop:
Code:
vendor_id       : GenuineIntel
model name      : Mobile Pentium MMX
flags           : fpu vme de pse tsc msr mce cx8 mmx
gcc version 3.2 (i586-pc-linux-gnu)

Result: '-O3 -march=pentium-mmx -fomit-frame-pointer -finline-functions-fcse-follow-jumps -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -fno-cprop-registers -funroll-all-loops -maccumulate-outgoing-args -fschedule-insns'

Performance gain(compare to -O3 only) ~ 9.9%


My desktop 1
Code:

vendor_id       : GenuineIntel
model name      : Pentium III (Coppermine)
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 mmx fxsr sse
gcc version 3.2 (i686-pc-linux-gnu)

Result: '-O3 -march=pentium3 -fomit-frame-pointer -finline-functions -funroll-loops'

Performance gain(compare to -O3 only) ~ 13.7%


My desktop 2
Code:

vendor_id       : AuthenticAMD
model name      : AMD Athlon(TM) MP 2000+
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow
vendor_id       : AuthenticAMD
model name      : AMD Athlon(TM) MP 2000+
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow
gcc version 3.2 (i686-pc-linux-gnu)

Result: '-O3 -march=athlon-mp -fomit-frame-pointer -finline-functions -fforce-mem -s -funroll-loops -frerun-loop-opt -fdelete-null-pointer-checks -fprefetch-loop-arrays -ffast-math -maccumulate-outgoing-args -fschedule-insns'

Performance gain(compare to -O3 only) ~ 19.6%


Adding some necessary flags, the final CFLAGS of, say, desktop 2 is:

CFLAGS="-O3 -march=athlon-mp -fomit-frame-pointer -finline-functions -fforce-mem -s -funroll-loops -frerun-loop-opt -fdelete-null-pointer-checks -fprefetch-loop-arrays -ffast-math -maccumulate-outgoing-args -fschedule-insns -Wall -pipe -msse -m3dnow -fpmath=sse -mmmx"
CXXFLAGS=${CFLAG}

Note: the orignal benchmark programs are pretty old(1995) and do not handle graphics and multimedia functions. Nevertheless, it still suit my purpose. :)

Hope this help.

I've made some amendment to the original post as some people told me that flags like -fmove-all-movables, -frename-registers and -malign-double are too dangerous to use(well, you should really bear risk on using optimizing flags); -fdelayed-branch is not supported and -mpreferred-stack-boundary=4 is already the default. Thank you for all your suggestions.
Back to top
View user's profile Send private message
Dennis4b
n00b
n00b


Joined: 31 Aug 2002
Posts: 5
Location: Finland

PostPosted: Sat Sep 21, 2002 10:51 pm    Post subject: Hmmmm Reply with quote

I just installed 1.4 into a chrooted directory in my 1.2 system, was WAY afraid the conversion scripts would take forever, swapped the 1.4 and 1.2 systems, moved my home directory, and now have a fresh 1.4 setup.

System is a Dell Inspiron 5000e, 128mb RAM, Intel PIII Coppermine (Celeron 2?) 700MHz.

From the stage 3 tarball to emerging KDE (so that includes of course X and some other libraries, NOT including remerging the stuff contained on the stage 3 tarball) took 24 hours!

That is a lot longer than when I did the same for 1.2 :)

I guess thats what weekends are for!
Back to top
View user's profile Send private message
attilahooper
n00b
n00b


Joined: 23 Jan 2003
Posts: 1

PostPosted: Fri Jan 24, 2003 7:58 pm    Post subject: Emerge -e world on PII 233@291 Reply with quote

i have a 233 mmx intel and i thought the cflags cpu type was i586.
anyway, my emerge -u world consistently hangs. locks scroll lock caps and numlock. using a stage 3 install, broadband connect, 98 megs of ram,4 gig drive ext2 fs.



first time using gentoo, never had this kinda headache with redhat.
Back to top
View user's profile Send private message
julot
n00b
n00b


Joined: 27 Aug 2003
Posts: 55
Location: Mexicus

PostPosted: Tue Nov 25, 2003 5:51 am    Post subject: Time Normal and VMware Reply with quote

Curiously, The time to compile gentoo on a "emulated" enviroment is lesser than in a normal desktop. (I think that many system calls and IRQ's are completely drowning the performance of the "real" PC), and maybe is showing it's architechtural limits.

Example:
My server:

P3 Tualatin 1.26 Ghz 512k Cachè, 815 Server Board with 512 MB Doble density 133 mhz dimms, Gentoo Install 15 hrs. ( Stage 1 and so With Mozilla, Gnome And KDE).

The Opposite:
2.5 GSX Server (VMware).
On Pentium 4 2.4 Ghz, Northwood, 512k Cachè, 100 mhz Quad, ("400 mhz" bus), The system host memory is 1 GB Rambus Samsung PC-800-45 using 256 MB share memory on Gentoo.

The same configuration than the server.
Time 8 hrs.

The same flags on the two:

Code:

CFLAGS="-march=pentium3 -O2 -pipe -fomit-frame-pointer"


And you agree with me that the overall performance of a Tualatin CPU is stronger than the tricky, misunderstood and misused P4.

Maybe the future of the PC is a Emulated enviroment on a REAL efficient architechture. (PowerPC, IBM, or even a Crusoe).

:wink:
_________________
"Sine ira et studio" Tacitus. (c.56-c.177 AD).
(With neither anger nor partiality).
Back to top
View user's profile Send private message
sireyessire
Advocate
Advocate


Joined: 20 Mar 2003
Posts: 2991
Location: back in Paris, France

PostPosted: Tue Nov 25, 2003 12:11 pm    Post subject: Re: Emerge -e world on PII 233@291 Reply with quote

Roderik wrote:
Ever wondered how long compiling would take on a PII 233@291 with 128MB ram?



i am sure you can do better with distcc installed and at least 2 others computers like P4 or athlon to help you :wink: :twisted:
_________________
I never think of the future. It comes soon enough.
Albert Einstein

Try simpler first
Shockley
Back to top
View user's profile Send private message
Clete2
Guru
Guru


Joined: 09 Aug 2003
Posts: 530
Location: Bloomington, Illinois

PostPosted: Tue Nov 25, 2003 2:13 pm    Post subject: Reply with quote

Haha, nice one... Good thing I'm upgrading to a P4 2.8GHz with HyperThreading before I emere -e world for P4 optimizations (I am on a PIII 933MHz right now, runs everything fine, but I figure I ought to upgrade as my motherboard is 4x AGP...)

Problem is, I have to replace the retarded dell case, waste of money... And my parents can't get all $400 of that for Christmas... I have to save 200 of it (their total budget for each of the us is $400, so I would get nothing else and that sucks)...
Back to top
View user's profile Send private message
Boris27
Guru
Guru


Joined: 05 Nov 2003
Posts: 562
Location: Almelo, The Netherlands

PostPosted: Tue Nov 25, 2003 3:34 pm    Post subject: Reply with quote

4X AGP is fast enough. 2X even is. Its marketing, baby, bloody marketing!
Back to top
View user's profile Send private message
NorseHeathenX
n00b
n00b


Joined: 08 Jul 2003
Posts: 42
Location: Florida

PostPosted: Sun Sep 26, 2004 9:41 am    Post subject: Reply with quote

What opts would be goof for a pentium4? I want to make this thing blaze!!!
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Documentation, Tips & Tricks All times are GMT
Page 1 of 1

 
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