ui slowness on newly installed ubuntu 13.04 64bit
Hi - I've recently installed Pharo 1.4 & 2.0 on a laptop I upgraded to Ubuntu 13.04 and using both Eliot's vm for Pharo 1.4 and the pharovm from the Ubuntu PPA. The UI is incredibly slow with it sometimes taking 1 second to switch windows when they are clicked or > 8 seconds to find the senders/implementors of a method. Is there something I can check to see what the problem may be? cat /proc/cpuinfo shows that the CPU is running at the high frequency. Thanks Paul
Paul, On 05 Jun 2013, at 23:52, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi -
I've recently installed Pharo 1.4 & 2.0 on a laptop I upgraded to Ubuntu 13.04 and using both Eliot's vm for Pharo 1.4 and the pharovm from the Ubuntu PPA. The UI is incredibly slow with it sometimes taking 1 second to switch windows when they are clicked or > 8 seconds to find the senders/implementors of a method. Is there something I can check to see what the problem may be?
cat /proc/cpuinfo shows that the CPU is running at the high frequency.
Thanks
Paul
It is probably best to open an issue, I couldn't find one. Can you find out if it is VM, image or OS related, by elimination ? There were/are other problems with Ubuntu 13.04, due to an upgraded libc if I am not mistaken. Sorry that I can't be of any more assistance. Sven
On Wed, Jun 5, 2013 at 11:52 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
I've recently installed Pharo 1.4 & 2.0 on a laptop I upgraded to Ubuntu 13.04 and using both Eliot's vm for Pharo 1.4 and the pharovm from the Ubuntu PPA. The UI is incredibly slow with it sometimes taking 1 second to switch windows when they are clicked or > 8 seconds to find the senders/implementors of a method. Is there something I can check to see what the problem may be?
cat /proc/cpuinfo shows that the CPU is running at the high frequency.
I'm also using 13.04 (but mine is 32bits). I can *not* reproduce your problem with the VM on the PPA. Just to be sure, I sent new versions of the packages to the https://launchpad.net/~pharo/+archive/stable/ PPA this morning. They will be ready in a few hours. Is there anything special with your image? -- Damien Cassou http://damiencassou.seasidehosting.st "Success is the ability to go from one failure to another without losing enthusiasm." Winston Churchill
Hi Damien, Because of the attached Message Tally outputs I don't think the problem is specific to changes I've made to an image. I also don't think that there is much differnece between Eliots vm or the pharo ppa vm. The Message Tally outputs are from profiling the UI while finding the senders of #setUp using various VM & image combinations. On squeak finding the senders of #setUp seems instantaneous. On Pharo it takes multiple seconds to find and display the senders. It may be my expectations are different than they should be. Is the result from the pharo ppa vm + pharo 2 image similar to what you're seeing on the 32bit Ubuntu? Thanks for looking into this. For what its worth I'm using a 5 yr old Dell D630 laptop with an 3GB of ram and a 120 GB SSD. I've also attached the output from /proc/cpuinfo Please let me know if there are other things I should check. Paul On 06/06/2013 12:31 AM, Damien Cassou wrote:
On Wed, Jun 5, 2013 at 11:52 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
I've recently installed Pharo 1.4 & 2.0 on a laptop I upgraded to Ubuntu 13.04 and using both Eliot's vm for Pharo 1.4 and the pharovm from the Ubuntu PPA. The UI is incredibly slow with it sometimes taking 1 second to switch windows when they are clicked or > 8 seconds to find the senders/implementors of a method. Is there something I can check to see what the problem may be?
cat /proc/cpuinfo shows that the CPU is running at the high frequency.
I'm also using 13.04 (but mine is 32bits). I can *not* reproduce your problem with the VM on the PPA. Just to be sure, I sent new versions of the packages to the https://launchpad.net/~pharo/+archive/stable/ PPA this morning. They will be ready in a few hours.
Is there anything special with your image?
-- Damien Cassou http://damiencassou.seasidehosting.st
"Success is the ability to go from one failure to another without losing enthusiasm." Winston Churchill
Hi Damien,
Because of the attached Message Tally outputs I don't think the problem is specific to changes I've made to an image. I also don't think that there is much differnece between Eliots vm or the pharo ppa vm. The Message Tally outputs are from profiling the UI while finding the senders of #setUp using various VM & image combinations. On squeak finding the senders of #setUp seems instantaneous. On Pharo it takes multiple seconds to find and display the senders.
ouch! So may be we have a regression in an iteration.
It may be my expectations are different than they should be.
No I do not see why sender would take ages.
Is the result from the pharo ppa vm + pharo 2 image similar to what you're seeing on the 32bit Ubuntu?
Thanks for looking into this. For what its worth I'm using a 5 yr old Dell D630 laptop with an 3GB of ram and a 120 GB SSD. I've also attached the output from /proc/cpuinfo
Please let me know if there are other things I should check.
Paul
On 06/06/2013 12:31 AM, Damien Cassou wrote:
On Wed, Jun 5, 2013 at 11:52 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
I've recently installed Pharo 1.4 & 2.0 on a laptop I upgraded to Ubuntu 13.04 and using both Eliot's vm for Pharo 1.4 and the pharovm from the Ubuntu PPA. The UI is incredibly slow with it sometimes taking 1 second to switch windows when they are clicked or > 8 seconds to find the senders/implementors of a method. Is there something I can check to see what the problem may be?
cat /proc/cpuinfo shows that the CPU is running at the high frequency.
I'm also using 13.04 (but mine is 32bits). I can *not* reproduce your problem with the VM on the PPA. Just to be sure, I sent new versions of the packages to the https://launchpad.net/~pharo/+archive/stable/ PPA this morning. They will be ready in a few hours.
Is there anything special with your image?
-- Damien Cassou http://damiencassou.seasidehosting.st
"Success is the ability to go from one failure to another without losing enthusiasm." Winston Churchill
<cpuinfo.txt><pharo-ppa-vm-pharo-1.4><squeak-4.4-all-in-one><eliots-vm-my-1.4-image><pharo-ppa-vm-pharo-2.txt><pharo-ppa-vm-my-1.4-image>
On Jun 6, 2013, at 5:12 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi Damien,
Because of the attached Message Tally outputs I don't think the problem is specific to changes I've made to an image. I also don't think that there is much differnece between Eliots vm or the pharo ppa vm. The Message Tally outputs are from profiling the UI while finding the senders of #setUp using various VM & image combinations. On squeak finding the senders of #setUp seems instantaneous. On Pharo it takes multiple seconds to find and display the senders.
This is really strange⦠here even doing "senders of" for all the >6000 senders of ifTrue: is quite fast for 2.0 and even faster in 3.0. So I really wonder what the problem is... Marcus
could we get the image so that we can try? Stef On Jun 7, 2013, at 8:27 AM, Marcus Denker <marcus.denker@inria.fr> wrote:
On Jun 6, 2013, at 5:12 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi Damien,
Because of the attached Message Tally outputs I don't think the problem is specific to changes I've made to an image. I also don't think that there is much differnece between Eliots vm or the pharo ppa vm. The Message Tally outputs are from profiling the UI while finding the senders of #setUp using various VM & image combinations. On squeak finding the senders of #setUp seems instantaneous. On Pharo it takes multiple seconds to find and display the senders.
This is really strange⦠here even doing "senders of" for all the >6000 senders of ifTrue: is quite fast for 2.0 and even faster in 3.0.
So I really wonder what the problem is...
Marcus
I thought the profiles showed it wasn't specific to my image but happened on both Pharo 1.4 & 2 using a cog vm on Ubuntu 13.04 64 bit. On Jun 7, 2013, at 12:44 AM, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
could we get the image so that we can try?
Stef
On Jun 7, 2013, at 8:27 AM, Marcus Denker <marcus.denker@inria.fr> wrote:
On Jun 6, 2013, at 5:12 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi Damien,
Because of the attached Message Tally outputs I don't think the problem is specific to changes I've made to an image. I also don't think that there is much differnece between Eliots vm or the pharo ppa vm. The Message Tally outputs are from profiling the UI while finding the senders of #setUp using various VM & image combinations. On squeak finding the senders of #setUp seems instantaneous. On Pharo it takes multiple seconds to find and display the senders.
This is really strange⦠here even doing "senders of" for all the >6000 senders of ifTrue: is quite fast for 2.0 and even faster in 3.0.
So I really wonder what the problem is...
Marcus
On 2013-06-07, at 08:27, Marcus Denker <marcus.denker@inria.fr> wrote:
On Jun 6, 2013, at 5:12 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi Damien,
Because of the attached Message Tally outputs I don't think the problem is specific to changes I've made to an image. I also don't think that there is much differnece between Eliots vm or the pharo ppa vm. The Message Tally outputs are from profiling the UI while finding the senders of #setUp using various VM & image combinations. On squeak finding the senders of #setUp seems instantaneous. On Pharo it takes multiple seconds to find and display the senders.
This is really strange⦠here even doing "senders of" for all the >6000 senders of ifTrue: is quite fast for 2.0 and even faster in 3.0.
So I really wonder what the problem is...
the provided profiling evidence shows the same. Pharo 1.4 / 2.0 on cog doesn't show any anomalies :/. But I must say that I had an image as well that behaved similarly recently...
For the attached profile I downloaded a new pharo 20606 image and ran the search for senders of #ifTrue: twice. It shows that it took 17 seconds on Ubuntu 13.04 64bit. How long does it take on your machine? On 06/06/2013 11:27 PM, Marcus Denker wrote:
On Jun 6, 2013, at 5:12 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi Damien,
Because of the attached Message Tally outputs I don't think the problem is specific to changes I've made to an image. I also don't think that there is much differnece between Eliots vm or the pharo ppa vm. The Message Tally outputs are from profiling the UI while finding the senders of #setUp using various VM & image combinations. On squeak finding the senders of #setUp seems instantaneous. On Pharo it takes multiple seconds to find and display the senders.
This is really strange⦠here even doing "senders of" for all the >6000 senders of ifTrue: is quite fast for 2.0 and even faster in 3.0.
So I really wonder what the problem is...
Marcus
On 07 Jun 2013, at 15:44, Paul DeBruicker <pdebruic@gmail.com> wrote:
For the attached profile I downloaded a new pharo 20606 image and ran the search for senders of #ifTrue: twice. It shows that it took 17 seconds on Ubuntu 13.04 64bit. How long does it take on your machine?
[ SystemNavigation new allSendersOf: #ifTrue: ] timeToRun =>> 258 "milliseconds" I feel for you, Paul, this is really annoying...
On 06/06/2013 11:27 PM, Marcus Denker wrote:
On Jun 6, 2013, at 5:12 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi Damien,
Because of the attached Message Tally outputs I don't think the problem is specific to changes I've made to an image. I also don't think that there is much differnece between Eliots vm or the pharo ppa vm. The Message Tally outputs are from profiling the UI while finding the senders of #setUp using various VM & image combinations. On squeak finding the senders of #setUp seems instantaneous. On Pharo it takes multiple seconds to find and display the senders.
This is really strange⦠here even doing "senders of" for all the >6000 senders of ifTrue: is quite fast for 2.0 and even faster in 3.0.
So I really wonder what the problem is...
Marcus
<pharo-vm-pharo-20606-ifTrue-twice.txt>
Le 07/06/2013 16:00, Sven Van Caekenberghe a écrit :
On 07 Jun 2013, at 15:44, Paul DeBruicker <pdebruic@gmail.com> wrote:
For the attached profile I downloaded a new pharo 20606 image and ran the search for senders of #ifTrue: twice. It shows that it took 17 seconds on Ubuntu 13.04 64bit. How long does it take on your machine?
[ SystemNavigation new allSendersOf: #ifTrue: ] timeToRun =>> 258 "milliseconds"
20606, Ubuntu 13.04 64bits, Core i3-2350M CPU @ 2.30GHz [ SystemNavigation new allSendersOf: #ifTrue: ] timeToRun =>> [292, 330] ms Thierry -- Thierry Goubier CEA list Laboratoire des Fondations des Systèmes Temps Réel Embarqués 91191 Gif sur Yvette Cedex France Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95
Sounds like I need some new hardware then? On 06/07/2013 07:16 AM, Goubier Thierry wrote:
Le 07/06/2013 16:00, Sven Van Caekenberghe a écrit :
On 07 Jun 2013, at 15:44, Paul DeBruicker <pdebruic@gmail.com> wrote:
For the attached profile I downloaded a new pharo 20606 image and ran the search for senders of #ifTrue: twice. It shows that it took 17 seconds on Ubuntu 13.04 64bit. How long does it take on your machine?
[ SystemNavigation new allSendersOf: #ifTrue: ] timeToRun =>> 258 "milliseconds"
20606, Ubuntu 13.04 64bits, Core i3-2350M CPU @ 2.30GHz [ SystemNavigation new allSendersOf: #ifTrue: ] timeToRun =>> [292, 330] ms
Thierry
I'll check with some older hardware; I have something slower than yours, but I'll have the results later tonight. Thierry Le 07/06/2013 17:04, Paul DeBruicker a écrit :
Sounds like I need some new hardware then?
On 06/07/2013 07:16 AM, Goubier Thierry wrote:
Le 07/06/2013 16:00, Sven Van Caekenberghe a écrit :
On 07 Jun 2013, at 15:44, Paul DeBruicker <pdebruic@gmail.com> wrote:
For the attached profile I downloaded a new pharo 20606 image and ran the search for senders of #ifTrue: twice. It shows that it took 17 seconds on Ubuntu 13.04 64bit. How long does it take on your machine?
[ SystemNavigation new allSendersOf: #ifTrue: ] timeToRun =>> 258 "milliseconds"
20606, Ubuntu 13.04 64bits, Core i3-2350M CPU @ 2.30GHz [ SystemNavigation new allSendersOf: #ifTrue: ] timeToRun =>> [292, 330] ms
Thierry
-- Thierry Goubier CEA list Laboratoire des Fondations des Systèmes Temps Réel Embarqués 91191 Gif sur Yvette Cedex France Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95
OK. I'm going to check on a 12.04 vm and windows xp vm on this hardware right now. I'll report back in a bit On 06/07/2013 08:13 AM, Goubier Thierry wrote:
I'll check with some older hardware; I have something slower than yours, but I'll have the results later tonight.
Thierry
Le 07/06/2013 17:04, Paul DeBruicker a écrit :
Sounds like I need some new hardware then?
On 06/07/2013 07:16 AM, Goubier Thierry wrote:
Le 07/06/2013 16:00, Sven Van Caekenberghe a écrit :
On 07 Jun 2013, at 15:44, Paul DeBruicker <pdebruic@gmail.com> wrote:
For the attached profile I downloaded a new pharo 20606 image and ran the search for senders of #ifTrue: twice. It shows that it took 17 seconds on Ubuntu 13.04 64bit. How long does it take on your machine?
[ SystemNavigation new allSendersOf: #ifTrue: ] timeToRun =>> 258 "milliseconds"
20606, Ubuntu 13.04 64bits, Core i3-2350M CPU @ 2.30GHz [ SystemNavigation new allSendersOf: #ifTrue: ] timeToRun =>> [292, 330] ms
Thierry
On Jun 7, 2013, at 3:45 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
For the attached profile I downloaded a new pharo 20606 image and ran the search for senders of #ifTrue: twice. It shows that it took 17 seconds on Ubuntu 13.04 64bit. How long does it take on your machine?
[SystemNavigation new browseMessageList: (SystemNavigation new allCallsOn: #ifTrue:) name: 'Senders or Users of ', #ifTrue: autoSelect: #ifTrue: ] timeToRun ->> 743 0.7 seconds. out of that, [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun takes ca. 0.11 seconds.
On 2013-06-07, at 16:03, Marcus Denker <marcus.denker@inria.fr> wrote:
On Jun 7, 2013, at 3:45 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
For the attached profile I downloaded a new pharo 20606 image and ran the search for senders of #ifTrue: twice. It shows that it took 17 seconds on Ubuntu 13.04 64bit. How long does it take on your machine?
On my virtualbox ubuntu 13.04 [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun == 1052 :/
My laptop is: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 1518 On 06/07/2013 07:07 AM, Camillo Bruni wrote:
On 2013-06-07, at 16:03, Marcus Denker <marcus.denker@inria.fr> wrote:
On Jun 7, 2013, at 3:45 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
For the attached profile I downloaded a new pharo 20606 image and ran the search for senders of #ifTrue: twice. It shows that it took 17 seconds on Ubuntu 13.04 64bit. How long does it take on your machine?
On my virtualbox ubuntu 13.04
[SystemNavigation new allCallsOn: #ifTrue: ] timeToRun == 1052
:/
On Jun 7, 2013, at 5:04 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
My laptop is:
[SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 1518
(3.0): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 117 I have a slow machine: 1.8Ghz Core i5 in a MacBook Air. It auto-overclocks to 2.8 Ghz when not using two cores. It's supposed to be not a fast system, as far as I understood. Marcus
Can you try on the command line only? curl get.pharo.org/30+vm | bash ./pharo Pharo.image eval "[SystemNavigation new allCallsOn: #ifTrue: ] timeToRun" On 2013-06-07, at 17:12, Marcus Denker <marcus.denker@inria.fr> wrote:
On Jun 7, 2013, at 5:04 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
My laptop is:
[SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 1518
(3.0): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 117
I have a slow machine: 1.8Ghz Core i5 in a MacBook Air. It auto-overclocks to 2.8 Ghz when not using two cores.
It's supposed to be not a fast system, as far as I understood.
Marcus
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms On 06/07/2013 08:15 AM, Camillo Bruni wrote:
Can you try on the command line only?
curl get.pharo.org/30+vm | bash ./pharo Pharo.image eval "[SystemNavigation new allCallsOn: #ifTrue: ] timeToRun"
On 2013-06-07, at 17:12, Marcus Denker <marcus.denker@inria.fr> wrote:
On Jun 7, 2013, at 5:04 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
My laptop is:
[SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 1518
(3.0): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 117
I have a slow machine: 1.8Ghz Core i5 in a MacBook Air. It auto-overclocks to 2.8 Ghz when not using two cores.
It's supposed to be not a fast system, as far as I understood.
Marcus
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit. On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
I think that we do have a problem, and I think it is VM related. This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image. Using the vm from get.pharo.org/30+vm [ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101 [ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420 Using the vm from ppa:pharo/stable [ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518 [ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156 That is a 10x slowdown ! Here are the VM details: $ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/] $ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/] One version seems older, but I don't know what changed inbetween them Sven On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
thanks for the insight. I opened an issue: https://pharo.fogbugz.com/default.asp?10879 can you try get.pharo.org/30+vmLatest to try out the "unstable" VM? On 2013-06-09, at 19:04, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I think that we do have a problem, and I think it is VM related.
This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image.
Using the vm from get.pharo.org/30+vm
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420
Using the vm from ppa:pharo/stable
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156
That is a 10x slowdown !
Here are the VM details:
$ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
$ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/]
One version seems older, but I don't know what changed inbetween them
Sven
On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
I am sorry, some 10x slowdown with vmLatest. $ ./pharo-ui --version 3.9-7 #1 Fri May 31 13:43:31 CEST 2013 gcc 4.6.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux pharo-linux 3.2.0-31-generic-pae #50-Ubuntu SMP Fri Sep 7 16:39:45 UTC 2012 i686 i686 i386 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/] On Mon, Jun 10, 2013 at 9:55 AM, Camillo Bruni <camillobruni@gmail.com>wrote:
thanks for the insight. I opened an issue: https://pharo.fogbugz.com/default.asp?10879
can you try get.pharo.org/30+vmLatest to try out the "unstable" VM?
On 2013-06-09, at 19:04, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I think that we do have a problem, and I think it is VM related.
This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image.
Using the vm from get.pharo.org/30+vm
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420
Using the vm from ppa:pharo/stable
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156
That is a 10x slowdown !
Here are the VM details:
$ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
$ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/]
One version seems older, but I don't know what changed inbetween them
Sven
On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
Hmm, unless there is a difference in compilation, then it is maybe the check we added with Igor to handle correctly objects as methods. I'll put this in my todo for the weekend :/. On Thu, Jun 13, 2013 at 9:57 PM, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I am sorry, some 10x slowdown with vmLatest.
$ ./pharo-ui --version 3.9-7 #1 Fri May 31 13:43:31 CEST 2013 gcc 4.6.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013
git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux pharo-linux 3.2.0-31-generic-pae #50-Ubuntu SMP Fri Sep 7 16:39:45 UTC 2012 i686 i686 i386 GNU/Linux
plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
On Mon, Jun 10, 2013 at 9:55 AM, Camillo Bruni <camillobruni@gmail.com>wrote:
thanks for the insight. I opened an issue: https://pharo.fogbugz.com/default.asp?10879
can you try get.pharo.org/30+vmLatest to try out the "unstable" VM?
On 2013-06-09, at 19:04, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I think that we do have a problem, and I think it is VM related.
This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image.
Using the vm from get.pharo.org/30+vm
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420
Using the vm from ppa:pharo/stable
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156
That is a 10x slowdown !
Here are the VM details:
$ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
$ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/]
One version seems older, but I don't know what changed inbetween them
Sven
On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
Wait, ppa:pharo/stable is version NBCoInterpreter NativeBoost-CogPlugin- GuillermoPolito.19? Sven, from your email I understood that the vm versions are in the same order as the measures, is that so? Guille On Thu, Jun 13, 2013 at 10:05 PM, Guillermo Polito < guillermopolito@gmail.com> wrote:
Hmm, unless there is a difference in compilation, then it is maybe the check we added with Igor to handle correctly objects as methods. I'll put this in my todo for the weekend :/.
On Thu, Jun 13, 2013 at 9:57 PM, Sven Van Caekenberghe <sven@stfx.eu>wrote:
I am sorry, some 10x slowdown with vmLatest.
$ ./pharo-ui --version 3.9-7 #1 Fri May 31 13:43:31 CEST 2013 gcc 4.6.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013
git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux pharo-linux 3.2.0-31-generic-pae #50-Ubuntu SMP Fri Sep 7 16:39:45 UTC 2012 i686 i686 i386 GNU/Linux
plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
On Mon, Jun 10, 2013 at 9:55 AM, Camillo Bruni <camillobruni@gmail.com>wrote:
thanks for the insight. I opened an issue: https://pharo.fogbugz.com/default.asp?10879
can you try get.pharo.org/30+vmLatest to try out the "unstable" VM?
On 2013-06-09, at 19:04, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I think that we do have a problem, and I think it is VM related.
This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image.
Using the vm from get.pharo.org/30+vm
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420
Using the vm from ppa:pharo/stable
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156
That is a 10x slowdown !
Here are the VM details:
$ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
$ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/]
One version seems older, but I don't know what changed inbetween them
Sven
On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
On 13 Jun 2013, at 22:07, Guillermo Polito <guillermopolito@gmail.com> wrote:
Wait, ppa:pharo/stable is version NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19?
Sven, from your email I understood that the vm versions are in the same order as the measures, is that so?
Yes they are: ./pharo-ui from get.pharo.org/vm is fast (and older) pharo-vm-x from pap:pharo/stable is slow (and newer) ./pharo-ui from get.pharo.org/vmLatest is also slow (and newer)
Guille
On Thu, Jun 13, 2013 at 10:05 PM, Guillermo Polito <guillermopolito@gmail.com> wrote: Hmm, unless there is a difference in compilation, then it is maybe the check we added with Igor to handle correctly objects as methods. I'll put this in my todo for the weekend :/.
On Thu, Jun 13, 2013 at 9:57 PM, Sven Van Caekenberghe <sven@stfx.eu> wrote: I am sorry, some 10x slowdown with vmLatest.
$ ./pharo-ui --version 3.9-7 #1 Fri May 31 13:43:31 CEST 2013 gcc 4.6.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013
git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux pharo-linux 3.2.0-31-generic-pae #50-Ubuntu SMP Fri Sep 7 16:39:45 UTC 2012 i686 i686 i386 GNU/Linux
plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
On Mon, Jun 10, 2013 at 9:55 AM, Camillo Bruni <camillobruni@gmail.com> wrote: thanks for the insight. I opened an issue: https://pharo.fogbugz.com/default.asp?10879
can you try get.pharo.org/30+vmLatest to try out the "unstable" VM?
On 2013-06-09, at 19:04, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I think that we do have a problem, and I think it is VM related.
This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image.
Using the vm from get.pharo.org/30+vm
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420
Using the vm from ppa:pharo/stable
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156
That is a 10x slowdown !
Here are the VM details:
$ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
$ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/]
One version seems older, but I don't know what changed inbetween them
Sven
On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
Ok, so besides the problem, It is strange to me that te ppa:stable yields to the latest unstable :). On Thu, Jun 13, 2013 at 10:11 PM, Sven Van Caekenberghe <sven@stfx.eu>wrote:
On 13 Jun 2013, at 22:07, Guillermo Polito <guillermopolito@gmail.com> wrote:
Wait, ppa:pharo/stable is version NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19?
Sven, from your email I understood that the vm versions are in the same order as the measures, is that so?
Yes they are:
./pharo-ui from get.pharo.org/vm is fast (and older)
pharo-vm-x from pap:pharo/stable is slow (and newer)
./pharo-ui from get.pharo.org/vmLatest is also slow (and newer)
Guille
On Thu, Jun 13, 2013 at 10:05 PM, Guillermo Polito < guillermopolito@gmail.com> wrote: Hmm, unless there is a difference in compilation, then it is maybe the check we added with Igor to handle correctly objects as methods. I'll put this in my todo for the weekend :/.
On Thu, Jun 13, 2013 at 9:57 PM, Sven Van Caekenberghe <sven@stfx.eu> wrote: I am sorry, some 10x slowdown with vmLatest.
$ ./pharo-ui --version 3.9-7 #1 Fri May 31 13:43:31 CEST 2013 gcc 4.6.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013
git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux pharo-linux 3.2.0-31-generic-pae #50-Ubuntu SMP Fri Sep 7 16:39:45 UTC 2012 i686 i686 i386 GNU/Linux
plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
On Mon, Jun 10, 2013 at 9:55 AM, Camillo Bruni <camillobruni@gmail.com> wrote: thanks for the insight. I opened an issue: https://pharo.fogbugz.com/default.asp?10879
can you try get.pharo.org/30+vmLatest to try out the "unstable" VM?
On 2013-06-09, at 19:04, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I think that we do have a problem, and I think it is VM related.
This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image.
Using the vm from get.pharo.org/30+vm
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420
Using the vm from ppa:pharo/stable
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156
That is a 10x slowdown !
Here are the VM details:
$ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
$ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/]
One version seems older, but I don't know what changed inbetween them
Sven
On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
On 13 Jun 2013, at 22:05, Guillermo Polito <guillermopolito@gmail.com> wrote:
Hmm, unless there is a difference in compilation, then it is maybe the check we added with Igor to handle correctly objects as methods. I'll put this in my todo for the weekend :/.
Thx!
On Thu, Jun 13, 2013 at 9:57 PM, Sven Van Caekenberghe <sven@stfx.eu> wrote: I am sorry, some 10x slowdown with vmLatest.
$ ./pharo-ui --version 3.9-7 #1 Fri May 31 13:43:31 CEST 2013 gcc 4.6.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013
git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux pharo-linux 3.2.0-31-generic-pae #50-Ubuntu SMP Fri Sep 7 16:39:45 UTC 2012 i686 i686 i386 GNU/Linux
plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
On Mon, Jun 10, 2013 at 9:55 AM, Camillo Bruni <camillobruni@gmail.com> wrote: thanks for the insight. I opened an issue: https://pharo.fogbugz.com/default.asp?10879
can you try get.pharo.org/30+vmLatest to try out the "unstable" VM?
On 2013-06-09, at 19:04, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I think that we do have a problem, and I think it is VM related.
This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image.
Using the vm from get.pharo.org/30+vm
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420
Using the vm from ppa:pharo/stable
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156
That is a 10x slowdown !
Here are the VM details:
$ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
$ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/]
One version seems older, but I don't know what changed inbetween them
Sven
On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
On 13 June 2013 22:05, Guillermo Polito <guillermopolito@gmail.com> wrote:
Hmm, unless there is a difference in compilation, then it is maybe the check we added with Igor to handle correctly objects as methods. I'll put this in my todo for the weekend :/.
10 times slowdown because of a single flag check?
On Thu, Jun 13, 2013 at 9:57 PM, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I am sorry, some 10x slowdown with vmLatest.
$ ./pharo-ui --version 3.9-7 #1 Fri May 31 13:43:31 CEST 2013 gcc 4.6.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013
git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux pharo-linux 3.2.0-31-generic-pae #50-Ubuntu SMP Fri Sep 7 16:39:45 UTC 2012 i686 i686 i386 GNU/Linux
plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
On Mon, Jun 10, 2013 at 9:55 AM, Camillo Bruni <camillobruni@gmail.com> wrote:
thanks for the insight. I opened an issue: https://pharo.fogbugz.com/default.asp?10879
can you try get.pharo.org/30+vmLatest to try out the "unstable" VM?
On 2013-06-09, at 19:04, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I think that we do have a problem, and I think it is VM related.
This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image.
Using the vm from get.pharo.org/30+vm
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420
Using the vm from ppa:pharo/stable
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156
That is a 10x slowdown !
Here are the VM details:
$ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
$ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/]
One version seems older, but I don't know what changed inbetween them
Sven
On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote:
And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
-- Best regards, Igor Stasenko.
On Jun 9, 2013, at 7:03 PM, Sven Van Caekenberghe <sven@stfx.eu> wrote:
I think that we do have a problem, and I think it is VM related.
This is what I see on my Ubuntu 13.04 64-bit machine (AMD Athlon(tm) II X4 640, 4GB, AMD Radeon HD 7750) using a Pharo 3.0 #30189 image.
Using the vm from get.pharo.org/30+vm
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 99 101
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 1455 1420
Using the vm from ppa:pharo/stable
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun. 1495 1518
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun. 11535 11156
That is a 10x slowdown !
ouch!
Here are the VM details:
$ ./pharo-ui -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/sven/Pharo/pharo-vm/ [default: /home/sven/Pharo/pharo-vm/]
$ pharo-vm-x --version 3.9-7 #1 Thu Jun 6 11:05:55 UTC 2013 gcc 4.7.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 6 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux samarium 2.6.24-32-xen #1 SMP Thu Jul 12 14:30:40 UTC 2012 i686 athlon i686 GNU/Linux plugin path: /usr/lib/pharo-vm/ [default: /usr/lib/pharo-vm/]
One version seems older, but I don't know what changed inbetween them
Sven
On Fri, Jun 7, 2013 at 6:49 PM, Paul DeBruicker <pdebruic@gmail.com> wrote: And the pharo2 linux all-in-one runs in 116ms on my hardware on Ubuntu 13.04 64bit.
On 06/07/2013 09:35 AM, Markus Fritsche wrote:
On 07.06.2013 17:55, Paul DeBruicker wrote:
From the below it seems to be something with running on linux 64bits with a UI slows it down by a factor of 10x.
Command line per Camillo below on ubuntu 13.04 64bit on the hardware was: 134ms
Using Pharo-2 windows all in one on a windows xp virtualbox VM was: 163ms
Using Pharo-2 linux all in one on a ubuntu 12,04 64bit virtualbox VM was: 1918ms
Using Pharo-2 linux all in one on a debian 6 32bit virtualbox VM was: 152 ms Further research: Pharo PPA VM, Ubuntu 13.04 64Bit: [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 2230 CogVM from Eliot directly ( http://www.mirandabanda.org/files/Cog/VM/VM.r2732/): [SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 164
Kind regards, Markus
On 07 Jun 2013, at 17:15, Camillo Bruni <camillobruni@gmail.com> wrote:
Can you try on the command line only?
This is getting _very_ weird: This is on Ubuntu 12.10 x64 running on a non-virtualised Intel(R) Core(TM) i5-3470 CPU @ 3.20GHz, 8GB RAM. $ ./pharo Pharo.image printVersion [version] 3.0 #30199 -1- $ curl get.pharo.org/vm| bash $ ./pharo -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/bt/pharo/pharo-vm/ [default: /home/bt/pharo/pharo-vm/] $ ./pharo Pharo.image eval '1 tinyBenchmarks' '1276807980 bytecodes/sec; 172715872 sends/sec' $ ./pharo Pharo.image eval '[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun' 59 $ ./pharo Pharo.image eval '[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun' 59 -2- $ curl get.pharo.org/vmLatest | bash $ ./pharo --version 3.9-7 #1 Fri May 31 13:43:31 CEST 2013 gcc 4.6.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux pharo-linux 3.2.0-31-generic-pae #50-Ubuntu SMP Fri Sep 7 16:39:45 UTC 2012 i686 i686 i386 GNU/Linux plugin path: /home/bt/pharo/pharo-vm/ [default: /home/bt/pharo/pharo-vm/] $ ./pharo Pharo.image printVersion [version] 3.0 #30199 $ ./pharo Pharo.image eval '1 tinyBenchmarks' '1335071707 bytecodes/sec; 179533604 sends/sec' $ ./pharo Pharo.image eval '[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun' 705 $ ./pharo Pharo.image eval '[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun' 697 So #tinyBenchmarks reports approx. the same result, but #allCallsOn: differs by a factor 10. Why ? Sven PS: We really need to start tracking benchmarks across platforms and versions - so much to do. -- Sven Van Caekenberghe http://stfx.eu Smalltalk is the Red Pill
yes I discovered the same behavior yesterday with esteban. thanks very much for the details! On 2013-06-14, at 23:53, Sven Van Caekenberghe <sven@stfx.eu> wrote:
On 07 Jun 2013, at 17:15, Camillo Bruni <camillobruni@gmail.com> wrote:
Can you try on the command line only?
This is getting _very_ weird:
This is on Ubuntu 12.10 x64 running on a non-virtualised Intel(R) Core(TM) i5-3470 CPU @ 3.20GHz, 8GB RAM.
$ ./pharo Pharo.image printVersion [version] 3.0 #30199
-1-
$ curl get.pharo.org/vm| bash
$ ./pharo -version 3.9-7 #1 Wed Mar 13 18:22:44 CET 2013 gcc 4.4.3 NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013 git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535 Linux linux-ubuntu-10 2.6.32-38-server #83-Ubuntu SMP Wed Jan 4 11:26:59 UTC 2012 x86_64 GNU/Linux plugin path: /home/bt/pharo/pharo-vm/ [default: /home/bt/pharo/pharo-vm/]
$ ./pharo Pharo.image eval '1 tinyBenchmarks' '1276807980 bytecodes/sec; 172715872 sends/sec'
$ ./pharo Pharo.image eval '[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun' 59
$ ./pharo Pharo.image eval '[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun' 59
-2-
$ curl get.pharo.org/vmLatest | bash
$ ./pharo --version 3.9-7 #1 Fri May 31 13:43:31 CEST 2013 gcc 4.6.3 NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013 git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595 Linux pharo-linux 3.2.0-31-generic-pae #50-Ubuntu SMP Fri Sep 7 16:39:45 UTC 2012 i686 i686 i386 GNU/Linux plugin path: /home/bt/pharo/pharo-vm/ [default: /home/bt/pharo/pharo-vm/]
$ ./pharo Pharo.image printVersion [version] 3.0 #30199
$ ./pharo Pharo.image eval '1 tinyBenchmarks' '1335071707 bytecodes/sec; 179533604 sends/sec'
$ ./pharo Pharo.image eval '[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun' 705
$ ./pharo Pharo.image eval '[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun' 697
So #tinyBenchmarks reports approx. the same result, but #allCallsOn: differs by a factor 10. Why ?
Sven
PS: We really need to start tracking benchmarks across platforms and versions - so much to do.
actually https://ci.inria.fr/pharo/view/VM/job/PharoVM-tests/ succesfully detected the degeneration and thus failed. And yes, we will add real benchmarks at some point. But for that we first need dedicated machines, as running them on a virtualized environment doesn't make much sense :/.
actually https://ci.inria.fr/pharo/view/VM/job/PharoVM-tests/ succesfully detected the degeneration and thus failed. And yes, we will add real benchmarks at some point. But for that we first need dedicated machines, as running them on a virtualized environment doesn't make much sense :/.
What do we need to start? Stef
On 2013-06-15, at 13:16, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
actually https://ci.inria.fr/pharo/view/VM/job/PharoVM-tests/ succesfully detected the degeneration and thus failed. And yes, we will add real benchmarks at some point. But for that we first need dedicated machines, as running them on a virtualized environment doesn't make much sense :/.
What do we need to start?
some machines :) as long as they are ours it doesn't really matter what size as long as they only run benchmarks. the current mac slave is not the best choice as it will be under load from jenkins... Anne told me that there is sort of a service from inria where they provide certain machines for benchmarking
On Jun 15, 2013, at 3:54 PM, Camillo Bruni <camillobruni@gmail.com> wrote:
On 2013-06-15, at 13:16, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
actually https://ci.inria.fr/pharo/view/VM/job/PharoVM-tests/ succesfully detected the degeneration and thus failed. And yes, we will add real benchmarks at some point. But for that we first need dedicated machines, as running them on a virtualized environment doesn't make much sense :/.
What do we need to start?
some machines :) as long as they are ours it doesn't really matter what size as long as they only run benchmarks. the current mac slave is not the best choice as it will be under load from jenkinsâ¦
Is a another mac ok?
Anne told me that there is sort of a service from inria where they provide certain machines for benchmarking
I do not know but we can ask.
On 2013-06-15, at 18:09, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
On Jun 15, 2013, at 3:54 PM, Camillo Bruni <camillobruni@gmail.com> wrote:
On 2013-06-15, at 13:16, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
actually https://ci.inria.fr/pharo/view/VM/job/PharoVM-tests/ succesfully detected the degeneration and thus failed. And yes, we will add real benchmarks at some point. But for that we first need dedicated machines, as running them on a virtualized environment doesn't make much sense :/.
What do we need to start?
some machines :) as long as they are ours it doesn't really matter what size as long as they only run benchmarks. the current mac slave is not the best choice as it will be under load from jenkinsâ¦
Is a another mac ok?
sure :) whatever dedicated machine will do, I think the only requirement is, that at least OSX 10.6 runs on it. (or we can always install a linux on top :P)
Anne told me that there is sort of a service from inria where they provide certain machines for benchmarking
I do not know but we can ask.
I will check on monday...
Ok, i checked this problem with three vms and the results are pretty similar between them (while they are different from Sven's)... - self compiled latest pharo vm - latest vm from get.pharo.org - stable vm from get.pharo.org "PharoVM self compiled" Smalltalk vm version 'NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 16 2013, NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 16 2013, git@gitorious.org:cogvm/blessed.git Commit: 0505d0c93504af2ad6dbd4cc580e6732120363d7 Date: 2013-06-12 16:26:39 +0200 By: Esteban Lorenzano <estebanlm@gmail.com> ' [ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 124 129 [ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun 4977 4688 "From our servers latest" Smalltalk vm version 'NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013, NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013, git:// gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595' [ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 123 124 [ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun 5187 5578 "From our servers stable" Smalltalk vm version 'NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013, NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013, git:// gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535' [ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 123 122 [ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun 5429 5036 ideas? Guille On Sun, Jun 16, 2013 at 2:22 PM, Camillo Bruni <camillobruni@gmail.com>wrote:
On 2013-06-15, at 18:09, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
On Jun 15, 2013, at 3:54 PM, Camillo Bruni <camillobruni@gmail.com>
wrote:
On 2013-06-15, at 13:16, Stéphane Ducasse <stephane.ducasse@inria.fr>
wrote:
actually https://ci.inria.fr/pharo/view/VM/job/PharoVM-tests/succesfully detected the degeneration and thus failed. And yes, we will add real benchmarks at some point. But for that we
first need dedicated machines, as running them on a virtualized environment doesn't make much sense :/.
What do we need to start?
some machines :) as long as they are ours it doesn't really matter what size as long as they only run benchmarks. the current mac slave is not the best choice as it will be under load from jenkinsâ¦
Is a another mac ok?
sure :) whatever dedicated machine will do, I think the only requirement is, that at least OSX 10.6 runs on it. (or we can always install a linux on top :P)
Anne told me that there is sort of a service from inria where they provide certain machines for benchmarking
I do not know but we can ask.
I will check on monday...
On 2013-06-16, at 21:42, Guillermo Polito <guillermopolito@gmail.com> wrote:
ideas? Guille
I checked a bit the gitorious sources and I couldn't immediately see something that would cause this. So I assume that it is on VM maker side... good look browsing diffs :( So what was changed in the Pharo VMMaker recently?
Did you try the one from the Ubuntu ppa or were those vms from somewhere else? On Jun 16, 2013, at 12:42 PM, Guillermo Polito <guillermopolito@gmail.com> wrote:
Ok, i checked this problem with three vms and the results are pretty similar between them (while they are different from Sven's)... - self compiled latest pharo vm - latest vm from get.pharo.org - stable vm from get.pharo.org
"PharoVM self compiled" Smalltalk vm version 'NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 16 2013, NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Jun 16 2013, git@gitorious.org:cogvm/blessed.git Commit: 0505d0c93504af2ad6dbd4cc580e6732120363d7 Date: 2013-06-12 16:26:39 +0200 By: Esteban Lorenzano <estebanlm@gmail.com> '
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 124 129
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun 4977 4688
"From our servers latest" Smalltalk vm version 'NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013, NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 May 31 2013, git://gitorious.org/cogvm/blessed.git Commit: 0f562e0afe14e533470278943b5d65d42ea133c2 Date: 2013-05-30 16:38:08 +0200 By: Igor Stasenko <siguctua@gmail.com> Jenkins build #14595'
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 123 124
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun 5187 5578
"From our servers stable" Smalltalk vm version 'NBCoInterpreter NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013, NBCogit NativeBoost-CogPlugin-EstebanLorenzano.18 uuid: a53445f9-c0c0-4015-97a3-be7db8d9ed6b Mar 13 2013, git://gitorious.org/cogvm/blessed.git Commit: 412abef33cbed05cf1d75329e451d71c0c6aa5a7 Date: 2013-03-13 17:48:50 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14535'
[ SystemNavigation new allCallsOn: #ifTrue: ] timeToRun 123 122
[ SystemNavigation new browseAllCallsOn: #ifTrue: ] timeToRun 5429 5036
ideas? Guille
On Sun, Jun 16, 2013 at 2:22 PM, Camillo Bruni <camillobruni@gmail.com> wrote:
On 2013-06-15, at 18:09, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
On Jun 15, 2013, at 3:54 PM, Camillo Bruni <camillobruni@gmail.com> wrote:
On 2013-06-15, at 13:16, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
actually https://ci.inria.fr/pharo/view/VM/job/PharoVM-tests/ succesfully detected the degeneration and thus failed. And yes, we will add real benchmarks at some point. But for that we first need dedicated machines, as running them on a virtualized environment doesn't make much sense :/.
What do we need to start?
some machines :) as long as they are ours it doesn't really matter what size as long as they only run benchmarks. the current mac slave is not the best choice as it will be under load from jenkinsâ¦
Is a another mac ok?
sure :) whatever dedicated machine will do, I think the only requirement is, that at least OSX 10.6 runs on it. (or we can always install a linux on top :P)
Anne told me that there is sort of a service from inria where they provide certain machines for benchmarking
I do not know but we can ask.
I will check on monday...
participants (11)
-
Camillo Bruni -
Damien Cassou -
Goubier Thierry -
Guillermo Polito -
Igor Stasenko -
Marcus Denker -
Markus Fritsche -
Paul DeBruicker -
Stéphane Ducasse -
Sven Van Caekenberghe -
Sven Van Caekenberghe