| View previous topic :: View next topic |
| Author |
Message |
Naib
Veteran
Joined: 21 May 2004
Posts: 3188
Location: UK - Birmingham
|
 Posted: Sun Mar 30, 2008 3:56 pm Post subject: [GENTOO] PackageManager showdowns |
 |
|
FOr quite some time the Paludis lot have been saying how broken portage is and how slow it is (and thus how fast paludis is) as well as a load of other "noise"
Well enough talk and more actual benchmarking
#1 some code mangling comes up with a poor easter-egg for April 1st
http://gentooexperimental.org/~patrick/weblog/archives/2008-03.html#e2008-03-29T21_36_39.txt
just an aside
#2 install times, footprints,
http://gentooexperimental.org/~patrick/weblog/archives/2008-03.html#e2008-03-30T00_31_03.txt
| Code: |
portage: Total files : 186
Total size : 1504.34 KiB
pkgcore: Total files : 409
Total size : 2121.04 KiB
paludis: Total files : 1887
Total size : 41662.27 KiB
|
And what abt general speed comparison?
| Code: |
emerge -p gnome && time emerge -p gnome 5 seconds
354 lines of output
pmerge -p gnome && time pmerge -p gnome 2 seconds
349 lines of output
paludis -ip gnome && time paludis -ip gnome 12 seconds
2071 lines of output plus ~100 on stderr |
mmmm nice and slow
Basically Paludis is HUGE w.r.t. the other two managers as well as being virtually complely a non drop-in replacement (w.r.t. commands, behaviour and such). Read artical for infomation on help comparision, general command differences,
finally
#3 bit more in-depth benchmark
http://gentooexperimental.org/~patrick/weblog/archives/2008-03.html#e2008-03-30T22_30_23.txt
| Code: |
Setup:
* A naked stage3, untarred and waiting to be used
* One emerge --sync to get a pristine ebuild tree
* An update gcc (4.1.2) to make the mangler happy
* emerge pkgcore
* emerge paludis
That setup is pretty much in the same state I left it yesterday.
So let's just try to resolve gnome with a cold cache. For that we run "echo 3 > /proc/sys/vm/drop_caches" and notice that the filesystem caches have shrunk to a minimum.
Edit: forgot the exact commands. emerge -p $1, pmerge -p $1, paludis -ip $1
Times
0m30.054s; 0m29.824s; 0m15.890s
All three are wall time as measured by "time"- can you guess who is who?
Same with KDE:
0m10.285s; 0m14.402s; 0m16.949s
And now xorg-x11 plus kde plus gnome:
0m18.182s; 0m23.251s; 0m34.159s |
oh yes. paludis is nice and fast 
any disagreements can be met with equal benchmarks no BS hotair from either side
_________________
| Quote: | | There is a forgotten, nay almost forbidden word, which means more to me than any other. That word is ENGLAND |
Weaver Projects
Last edited by Naib on Sun Mar 30, 2008 4:02 pm; edited 1 time in total |
|
| Back to top |
|
 |
ShiRhoHaiSun
n00b
Joined: 28 Mar 2008
Posts: 1
|
| Posted: Sun Mar 30, 2008 3:59 pm Post subject: What you say? |
|
|
Is fake? Paludis people always say is faster, u claim they not speak thruth always?
Very confusing! |
|
| Back to top |
|
|
Naib
Veteran
Joined: 21 May 2004
Posts: 3188
Location: UK - Birmingham
|
| Posted: Sun Mar 30, 2008 4:01 pm Post subject: |
|
|
yes 2+2=4 is also fake
_________________
| Quote: | | There is a forgotten, nay almost forbidden word, which means more to me than any other. That word is ENGLAND |
Weaver Projects |
|
| Back to top |
|
|
dleverton
Guru
Joined: 28 Aug 2006
Posts: 517
|
| Posted: Sun Mar 30, 2008 4:02 pm Post subject: |
|
|
| That blog is full of lies, hypocrisy and personal attacks. |
|
| Back to top |
|
|
ShiRhoHaiSun
n00b
Joined: 28 Mar 2008
Posts: 1
|
| Posted: Sun Mar 30, 2008 4:05 pm Post subject: hmm? |
|
|
| dleverton wrote: | | That blog is full of lies, hypocrisy and personal attacks. |
U can show errors in post? I'm wondering how it is fail. |
|
| Back to top |
|
|
Xaero
Apprentice
Joined: 12 Jun 2006
Posts: 180
Location: East Lansing, Michigan
|
| Posted: Sun Mar 30, 2008 4:05 pm Post subject: |
|
|
| Interesting. I remember putting paludis on my laptop before I killed my install several months back, and thinking that it was alot faster than portage. Guess that's not the case anymore. |
|
| Back to top |
|
|
Naib
Veteran
Joined: 21 May 2004
Posts: 3188
Location: UK - Birmingham
|
| Posted: Sun Mar 30, 2008 4:09 pm Post subject: |
|
|
| Xaero wrote: | | Interesting. I remember putting paludis on my laptop before I killed my install several months back, and thinking that it was alot faster than portage. Guess that's not the case anymore. |
Thats the thing, there has always been "Placebo" with everything, its just human psychology to think something is better because that is what is said so yr mind fills in the blanks. But at the end of the day it is hard numbers that do the talking - that is why drug trials have control patients that take a placebo drug to rule out the mind doing the work
and again any disagreement with these results can easily be met with test done by anyone
_________________
| Quote: | | There is a forgotten, nay almost forbidden word, which means more to me than any other. That word is ENGLAND |
Weaver Projects |
|
| Back to top |
|
|
Xaero
Apprentice
Joined: 12 Jun 2006
Posts: 180
Location: East Lansing, Michigan
|
| Posted: Sun Mar 30, 2008 4:47 pm Post subject: |
|
|
| Naib wrote: | | Xaero wrote: | | Interesting. I remember putting paludis on my laptop before I killed my install several months back, and thinking that it was alot faster than portage. Guess that's not the case anymore. |
Thats the thing, there has always been "Placebo" with everything, its just human psychology to think something is better because that is what is said so yr mind fills in the blanks. But at the end of the day it is hard numbers that do the talking - that is why drug trials have control patients that take a placebo drug to rule out the mind doing the work
and again any disagreement with these results can easily be met with test done by anyone |
That's true, it was several months ago though, maybe at that time paludis was faster than portage.
btw, did anyone else catch this bit in the article?
| Quote: | | At first I thought about methods of skewing the benchmarks if needed. |
That's nice to know.  I think if I get the time I will try and perform the same benchmark he did just to see how accurate his is, because now I'm curious. |
|
| Back to top |
|
|
Phenax
l33t
Joined: 10 Mar 2006
Posts: 965
|
| Posted: Sun Mar 30, 2008 4:52 pm Post subject: |
|
|
On my current system, with no KDE components installed.
| Code: |
emerge -p kde ; time emerge -p kde
real 0m3.031s
user 0m2.824s
sys 0m0.164s
paludis -ip kde ; time paludis -ip kde (using suboptimal portage config files)
real 0m2.704s
user 0m2.088s
sys 0m0.596s
pmerge -p kde ; time pmerge -p kde
real 0m0.593s
user 0m0.547s
sys 0m0.044s
|
|
|
| Back to top |
|
|
Zougloub
n00b
Joined: 14 Dec 2007
Posts: 23
|
| Posted: Sun Mar 30, 2008 5:09 pm Post subject: |
|
|
Doing pkgmgr -p stuff ; time pkgmgr -p stuff is not fair to Paludis.
That's not what users do, why benchmark this way... just use pkgmgr -p stuff and you'll get very different results.
Btw, don't expect to get objective benchmarks from Naib or patrick.
Let's imagine that paludis was indeed slower than emerge or pkgcore (which is not the case) ; I would be still using it.
It is fast and offers convenient features.
--
Someone who used portage for a long time, uses paludis everyday and expects pkgcore to get better (ie. work) some day. |
|
| Back to top |
|
|
zixnub
n00b
Joined: 27 Dec 2007
Posts: 32
Location: Ghent, Belgium
|
| Posted: Sun Mar 30, 2008 5:22 pm Post subject: |
|
|
You could convert portage with jythonc to java bytecode and have it run at java speeds 
_________________
Erik's Window Manager
http://code.google.com/p/ewm |
|
| Back to top |
|
|
Dirk.R.Gently
Guru
Joined: 28 Jan 2007
Posts: 546
Location: Titan
|
|
| Back to top |
|
|
jonnevers
Veteran
Joined: 02 Jan 2003
Posts: 1542
Location: Gentoo64 land
|
| Posted: Sun Mar 30, 2008 5:32 pm Post subject: |
|
|
| Quote: | Conclusion
At first I thought about methods of skewing the benchmarks if needed. |
garbage. regardless of any relative technical merit it may have. |
|
| Back to top |
|
|
Jokey_
Developer
Joined: 24 Jan 2006
Posts: 34
Location: Germany
|
|
| Back to top |
|
|
ShiRhoHaiSun
n00b
Joined: 28 Mar 2008
Posts: 1
|
| Posted: Sun Mar 30, 2008 5:44 pm Post subject: |
|
|
| Zougloub wrote: | Doing pkgmgr -p stuff ; time pkgmgr -p stuff is not fair to Paludis.
That's not what users do, why benchmark this way... just use pkgmgr -p stuff and you'll get very different results.
|
Is not default action to see useflags etc.etc. ? In compile no difference I say!
| Zougloub wrote: |
Btw, don't expect to get objective benchmarks from Naib or patrick.
|
You have numbers different? Would be nice to see what you from it get!
| Zougloub wrote: |
Let's imagine that paludis was indeed slower than emerge or pkgcore (which is not the case) ; I would be still using it.
It is fast and offers convenient features.
|
So why not use Ubuntu, has automatic updates. Use Windows, even easier! But looks like not much friendly being here is ...  |
|
| Back to top |
|
|
Think4UrS11
Administrator
Joined: 25 Jun 2003
Posts: 6118
Location: Murphy is my neighbor
|
| Posted: Sun Mar 30, 2008 5:53 pm Post subject: |
|
|
merged Dirk.R.Gently & jonnevers in here
_________________
Nothing is secure / Security is always a trade-off with usability / Do not assume anything / Trust no-one, nothing / Paranoia is your friend / Think for yourself |
|
| Back to top |
|
|
dleverton
Guru
Joined: 28 Aug 2006
Posts: 517
|
|
| Back to top |
|
|
ShiRhoHaiSun
n00b
Joined: 28 Mar 2008
Posts: 1
|
| Posted: Sun Mar 30, 2008 6:12 pm Post subject: |
|
|
Haha, u need to test on gentoo system, not sabayon. U silly ricer. -funroll-loops -funroll-all-over-floor !! |
|
| Back to top |
|
|
ShiRhoHaiSun
n00b
Joined: 28 Mar 2008
Posts: 1
|
| Posted: Sun Mar 30, 2008 6:13 pm Post subject: |
|
|
| jonnevers wrote: | | Quote: | Conclusion
At first I thought about methods of skewing the benchmarks if needed. |
garbage. regardless of any relative technical merit it may have. |
U not like facts? Why? |
|
| Back to top |
|
|
dleverton
Guru
Joined: 28 Aug 2006
Posts: 517
|
| Posted: Sun Mar 30, 2008 6:15 pm Post subject: |
|
|
| ShiRhoHaiSun wrote: | | Haha, u need to test on gentoo system, not sabayon. U silly ricer. -funroll-loops -funroll-all-over-floor !! |
No, I'm quite sure that I'm using Gentoo. I'm also quite sure that I'm not using any ridiculous CFLAGS. I'm also quite sure that that's a personal attack, although I won't bother reporting it because apparently attacking Paludis users and devs is perfectly acceptable on this forum. |
|
| Back to top |
|
|
ShiRhoHaiSun
n00b
Joined: 28 Mar 2008
Posts: 1
|
| Posted: Sun Mar 30, 2008 6:21 pm Post subject: |
|
|
| dleverton wrote: | | No, I'm quite sure that I'm using Gentoo. I'm also quite sure that I'm not using any ridiculous CFLAGS. I'm also quite sure that that's a personal attack, although I won't bother reporting it because apparently attacking Paludis users and devs is perfectly acceptable on this forum. |
U started attack. Saying everything a lie ... now you can't tolerate opinion? Very not nice. U should not hit others if u start crying when someone looks back ...
And it not gentoo system if not portage runnable. Next u say Ubuntu is gentoo too and we all laugh and drink beer? |
|
| Back to top |
|
|
dleverton
Guru
Joined: 28 Aug 2006
Posts: 517
|
| Posted: Sun Mar 30, 2008 6:26 pm Post subject: |
|
|
| ShiRhoHaiSun wrote: | | U started attack. Saying everything a lie |
I didn't attack anyone, I stated a fact.
| ShiRhoHaiSun wrote: | | ... now you can't tolerate opinion? |
Whether or not I use Sabayon is not a matter of opinion.
| ShiRhoHaiSun wrote: | | Very not nice. U should not hit others if u start crying when someone looks back ... |
See above.
| ShiRhoHaiSun wrote: | | And it not gentoo system if not portage runnable. |
1) A Gentoo system is one primarily built with ebuilds from Gentoo's repository. (I say "primarily" because many people use personal or third-party overlays, and I don't think we want to exclude those.)
2) If you would actually read my link, you would see that Portage worked fine. It was pkgcore that died spectacularly.
| ShiRhoHaiSun wrote: | | Next u say Ubuntu is gentoo too and we all laugh and drink beer? |
Only if you can tell me how to build an Ubuntu system using ebuilds from Gentoo's repository. |
|
| Back to top |
|
|
UberPinguin
Guru
Joined: 20 Nov 2005
Posts: 503
Location: 2416.94 Miles From Home
|
| Posted: Sun Mar 30, 2008 6:34 pm Post subject: |
|
|
It's pretty clear from the error messages spat out by portage that paludis has mangled a number of important files, which might go a long way to explaining why pmerge doesn't work on your system. How about you try again, this time with each package manager in an independent chroot or VM?
_________________
Dream with the feathers of angels stuffed beneath your head. |
|
| Back to top |
|
|
tylerwylie
Guru
Joined: 19 Sep 2004
Posts: 445
Location: /US/Illinois/Champaign
|
| Posted: Sun Mar 30, 2008 6:36 pm Post subject: |
|
|
Lol Gentoo.
_________________
| dmitchell wrote: | | Live a virtuous and moral life, become an example to others, and be radical in your opposition to the state. |
|
|
| Back to top |
|
|
dleverton
Guru
Joined: 28 Aug 2006
Posts: 517
|
| Posted: Sun Mar 30, 2008 6:39 pm Post subject: |
|
|
| UberPinguin wrote: | | It's pretty clear from the error messages spat out by portage that paludis has mangled a number of important files |
Those are warnings, not errors, and Paludis has done no such thing.
| UberPinguin wrote: | | which might go a long way to explaining why pmerge doesn't work on your system. |
No, pmerge doesn't work because it cries like a little baby at the first thing it sees that it doesn't recognise. If you want see for yourself with something that is supported by portage, you might like to put a -foo/bar line in an overlay profiles/package.mask, but I didn't do that because adding it just to make pkgcore look bad would be cheating.
| UberPinguin wrote: | | How about you try again, this time with each package manager in an independent chroot or VM? |
Why? |
|
| Back to top |
|
|
|
Off the Wall
