Welcome to MilkyWay@home

Posts by [AF>Le_Pommier] Jerome_C2005

1) Message boards : News : New Poll Regarding GPU Application of N-Body (Message 71000)
Posted 22 Jul 2021 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
I'm afraid the poll is a bit light : as said above we need more details to be able to compare,

- would 1 GPU task do the same *amount of work* than 1 CPU thread task ? (since you already say the process time will be equivalent)
- would 1 GPU task also need a % of a CPU thread ? (as it is often the case with GPU tasks, especially with OpenCL, I can only run AMD OpenCL tasks on my iMac) and then do you know what % ? and again, also to do the same amount of of work than the regular 1 thread CPU task ? (if yes it would then be counter productive and not equivalent !)
2) Message boards : News : Nbody Release 1.54 (Message 64192)
Posted 18 Dec 2015 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Thanks for the explanation captiosus.
3) Message boards : News : New Release- Nbody version 1.52 (Message 64191)
Posted 18 Dec 2015 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
OK so you cannot say "I don't want this specific app but I'm Ok with any other new one not in the list yet" ?
4) Message boards : News : Nbody Release 1.54 (Message 64165)
Posted 11 Dec 2015 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Please read in the other topic for announcement of n-body 1.52 what I wrote (one month ago and now again), nobody ever answered me but they are not working on my Mac : running on 1 core out of 8, blocking the other boinc apps to run (pretending it is "mt"), and in 1.52 they would not even complete normally (AND they started running again now when my setup does not allow nbody)...

http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=3819&postid=64164


EDIT : there is a change now, after 15mn of running they actually start to run on all 8 cores and really using 100% of my CPU (instead of 1/8th).


EDIT2 : they don't completely run 100%, it's more like 90% of available CPU, letting a 10% idle.


EDIT3 : OK so the task did finish on estimated time and was sent successfully back to the servers, so things look better with this 1.54, even though not perfect (running 15mn with 1/8 of CPU and not running a full 100% for the remaining time).
5) Message boards : News : New Release- Nbody version 1.52 (Message 64164)
Posted 11 Dec 2015 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Guess what, even though it is still marked as not selected in my account :



I got AGAIN those n-body simulation (version is now 1.54) PRETENDING to be "mt" that are only using ONE CORE instead of 8 and blocking all other apps...

One has completed 10mn with 25%, it says 38mn left to complete, I'll wait this to see if it's or if they also have the bad behavior...


EDIT : things went better now, read the topic about v1.54.
Even though my account setup is not to run them...
6) Message boards : News : New Release- Nbody version 1.52 (Message 64089)
Posted 12 Nov 2015 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Am I the only one ?
7) Message boards : News : New Release- Nbody version 1.52 (Message 64064)
Posted 6 Nov 2015 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Hi

Nbody is really not working well on my iMac : it behaves like mt app (ie stops all other apps running on my i7 8 cores) BUT it only takes 1/8th of the CPU power.




(99% means "of one core", not all 8 cores, look at the CPU charge at the bottom)

Besides sometimes it doesn't stop to run after 100 % completion, and just continues forever using 0% CPU, blocking boinc, I have to cancel it ! (if I can see it, I don't spend all day in front of my computer...)

Actually the current one is not 100% yet but already stopped using CPU for a while (and lasted already much more than the 9 min of the previous one, that ran well)





Or it will crash by itself (there were others).

I'm removing nBody from my preferences for the moment, I hope this will be solved soon.

Thanks for your help.
8) Message boards : News : New Nbody version (Message 62639)
Posted 27 Oct 2014 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Well I just came home and saw your answer, and I just found out that the WU finished only a few minutes ago ! What a coincidence, or was it fear ? (from the WU, that I would cancel it) :D

Anyway here is the beast : 2 days 18 hours of calculation for a total of 18 days 11 hours of computing time on my i7 !!!

http://milkyway.cs.rpi.edu/milkyway/result.php?resultid=859028388

I definitively hope that someone else is going to calculate it... and that I get "some kind of credit" for this :)

Guess what, now I have another mt WU working now (since 6mn) but it claims only 1h30 min left as an estimate... what can be SO different between these two ?

http://milkyway.cs.rpi.edu/milkyway/result.php?resultid=859028387
9) Message boards : News : New Nbody version (Message 62630)
Posted 26 Oct 2014 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Hi

I juste realized I have a MilkyWay@Home N-Body Simulation 1.44 (mt) running on my iMac (i7 late 2009) since more than a day

Nom ps_nbody_08_05_orphan_sim_3_1413455402_39907_0
Application MilkyWay@Home N-Body Simulation 1.44 (mt)
Nom unité ps_nbody_08_05_orphan_sim_3_1413455402_39907
Etat Calcul en cours
Reçue 24/10/2014 23:00:02
Date limite d'envoi 05/11/2014 22:00:02
Vitesse estimée 13,03 GFLOPs/sec
Taille de la tâche estimée 159 243 GFLOPs
Ressources 8 CPUs
Temps CPU au dernier point de reprise 09d,07:26:27
Temps CPU 09d,07:30:09
Temps de calcul 01d,09:48:49
Estimation temps restant 01d,04:35:09
Pourcentage effectué 54,189%
Taille mémoire virtuelle 2 472,48 MB
Réglage du volume de travail 30,50 MB
Répertoire slots/9
Processus ID 86875


The % is moving slowly, but it's moving. The WU is eating avg 600/650 % CPU on the mac (out of 800% max, Mac OS X measures 100% per core), so it's working.

So the calculation estimate is more than another day, it will be over 2 days of calculation if it terminates based on this estimate.

Is this OK ? should I let it run ?
10) Message boards : Number crunching : Home Separation (Modified Fit) v1.28 taking ages (Message 60789)
Posted 22 Jan 2014 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
My question is not about credit, my question is how on earth can 816 secs of calculation produce the same scientific result than 45 hours done to calculate the same task on my Mac ?? is the application doing some random loops or something ? being a developer in my early IT days this seems quite astounding...

Regarding credit, as I said, I actually don't care.
11) Message boards : Number crunching : Home Separation (Modified Fit) v1.28 taking ages (Message 60782)
Posted 21 Jan 2014 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
... and the first guy took 816 secs to crunch the SAME WU ??

Isn't there something very wrong in there ? how can they all 3 produce the same output ???
12) Message boards : Number crunching : Home Separation (Modified Fit) v1.28 taking ages (Message 60773)
Posted 19 Jan 2014 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Well, both finally terminated, very oddly only one can be seen in my WU list now... luckily I have the full history in my BoincTasks, the last one finished today after 45 hours of calculation, the first one finished 2 days ago after 47 hours, why is only one visible anymore ?

Anyway I see I got 320 credits for a 45 hours WU, this is... I mean... wow, thank you so much. (hey, don't take this badly, I'm just kidding, if I would only participate to projects for the credits, I wouldn't he here and wouldn't do most of the projects I do ! ;) )

Regarding the fact I crunch many projects at the same time : I've always done this and I think I'll always do, this is the way I like it and only during raids or special events organized by my team do I switch to "mono project" or a reduced list during a limited time (max 2 weeks). When it's over I'm always very happy to go back to my multi-project habit.

So if the project application is not able to crunch together with others app, well, it's just missing boinc spirit, if boinc was supposed to be only mono projet it wouldn't exist at all (seti, WCG, etc, would still have their own system). Boinc is about sharing and letting options.


Anyway thank you all for your time and explanations, I always like to discuss and share on projects forums, and very happy when I do have feedback. Dead project's forums are just dead project :)
13) Message boards : Number crunching : Home Separation (Modified Fit) v1.28 taking ages (Message 60759)
Posted 17 Jan 2014 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Well I'm running many different projects on my Mac, most of the long running WU of any project are, at some point, suspended by boinc (then other WU/project start, or restart), and then after some time they are resumed (and eventually they can do this several times, if the WU is very long, ie CPDN...), unless they are running in high priority mode (due to lack of time before deadline), then they won't stop.

I have increased the boinc parameter that defines the time before switching WU (I put 1440 mn instead of the standard low value, which I can't remember) and I have the "keep in memory while suspended" parameter ON, because I have lot's of RAM, and for those projects/applications that don't implement checkpoints, the WU crunching done will not be lost.
14) Message boards : Number crunching : Home Separation (Modified Fit) v1.28 taking ages (Message 60751)
Posted 16 Jan 2014 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Gee I don't get it, now I come home and I see boinc says one has crunched 33 hours remaining 7 hours (67%) and the other 33 hours remaining 14 hours (59%)... they are the same WUs than before, I can't see anything in boinc messages appart that the WU did resume at some point, on my account page I only see the same 2 WUs sent on the 09/01 !!

WTF ?
15) Message boards : Number crunching : Home Separation (Modified Fit) v1.28 taking ages (Message 60744)
Posted 15 Jan 2014 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Hi

I have a Home Separation (Modified Fit) v1.28 Mac OS X WU that has been crunching on CPU (no MW GPU for my Mac, I'd need OpenCL !) for more than 70 hours, with less than 54% completion, showing no remaining time ("-") with a deadline below 140 hours.

Should I worry ? :)

(this one has been running for 26 hours with a 59% completion, with a displayed remaining time of 21 hours)

Thanks for you help.
16) Message boards : Number crunching : N-Body + Mac OS X + i7 = no good ? (Message 50255)
Posted 19 Jul 2011 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
... then at some point in the afternoon, it decided to run for 10 mn without stopping (wow) and finished, and requested a new WU, that also ran for a complete 10 mn (nbody also) and also finished... and then it requested another WU and got a MW@home 0.82 "classical non MT WU", which is obviously behaving like the other project with one single core among other projects...

So I don't know what to think about those nbody unit, that tend to have a very strange behavior, on Mac OS X at least.
17) Message boards : Number crunching : N-Body + Mac OS X + i7 = no good ? (Message 50190)
Posted 18 Jul 2011 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Definitively something wrong, here are all the milkyway project entries for today (after I left home this morning) in the boinc message list :

Lun 18 jul 12:25:56 2011 | Milkyway@home | Resuming task ps_nbody_test3_498129_1 using milkyway_nbody version 60

Lun 18 jul 14:28:28 2011 | Milkyway@home | Restarting task ps_nbody_test3_498129_1 using milkyway_nbody version 60

Lun 18 jul 15:36:37 2011 | Milkyway@home | Restarting task ps_nbody_test3_498129_1 using milkyway_nbody version 60

Lun 18 jul 19:13:14 2011 | Milkyway@home | Restarting task ps_nbody_test3_498129_1 using milkyway_nbody version 60

And each time, 10 secs after, all the other projects are restarting (successfully, unlike the initial problem I had back in june). This WU is cumulating... 30 secs of computing so far... more than 24 hours remaining, good luck !!!
18) Message boards : Number crunching : N-Body + Mac OS X + i7 = no good ? (Message 50156)
Posted 18 Jul 2011 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Of course it is not normal that boinc would only run 3 cores, or 5 cores, when it's normally using 8, and this because a nbody has just ran !

I'm using 6.12.33 now, I had stopped milky until now and i decided to give it another try : when nbody starts, I receive some errors from a few other project (not the 8 running before)

Lun 18 jul 04:34:29 2011 | ibercivis | Task Mr.Wilson_10_01_50_52_824277135_2133303569_2: no shared memory segment
Lun 18 jul 04:34:29 2011 | ibercivis | Task Mr.Wilson_10_01_50_52_824277135_2133303569_2 exited with zero status but no 'finished' file
Lun 18 jul 04:34:29 2011 | ibercivis | If this happens repeatedly you may need to reset the project.

I have the same for CPDN, but not the other 6 running projects.

Then it finishes after only 10 mn, and boinc is able to restart 8 projects just after that, including the ibercivis with that "error", apparently not affecting it (it finished in valid status).

For CPDN I don't know cause it has not restarted it yet.

So I'd say : things have improved, though a bit weird.


Another thing is that I can see it ran 2 tasks during the night,

Lun 18 jul 00:18:04 2011 | Milkyway@home | Sending scheduler request: To fetch work.
Lun 18 jul 00:18:04 2011 | Milkyway@home | Requesting new tasks for CPU
Lun 18 jul 00:18:06 2011 | Milkyway@home | Scheduler request completed: got 1 new tasks
Lun 18 jul 00:18:08 2011 | Milkyway@home | Started download of milkyway_nbody_0.60_x86_64-apple-darwin__mt
Lun 18 jul 00:18:11 2011 | Milkyway@home | Finished download of milkyway_nbody_0.60_x86_64-apple-darwin__mt

Lun 18 jul 01:13:14 2011 | Milkyway@home | Starting task ps_nbody_test3_480599_1 using milkyway_nbody version 60
Lun 18 jul 01:22:42 2011 | Milkyway@home | Computation for task ps_nbody_test3_480599_1 finished
Lun 18 jul 01:22:43 2011 | Milkyway@home | Sending scheduler request: To fetch work.
Lun 18 jul 01:22:43 2011 | Milkyway@home | Reporting 1 completed tasks, requesting new tasks for CPU
Lun 18 jul 01:22:44 2011 | Milkyway@home | Scheduler request completed: got 1 new tasks

Lun 18 jul 04:34:17 2011 | Milkyway@home | Starting task ps_nbody_test3_504091_0 using milkyway_nbody version 60
Lun 18 jul 04:47:11 2011 | Milkyway@home | Computation for task ps_nbody_test3_504091_0 finished
Lun 18 jul 04:47:11 2011 | Milkyway@home | Sending scheduler request: To fetch work.
Lun 18 jul 04:47:11 2011 | Milkyway@home | Reporting 1 completed tasks, requesting new tasks for CPU
Lun 18 jul 04:47:13 2011 | Milkyway@home | Scheduler request completed: got 1 new tasks

but as you can there is not the traditional download sequence after it says "got 1 new tasks" after the first and second time it finishes, and also on my account on the website I can only see one pending WU, no trace of the other two...


edit : I can tell you that the behavior is very strange : now I can see that another nbody started

Lun 18 jul 08:59:01 2011 | Milkyway@home | Starting task ps_nbody_test3_498129_1 using milkyway_nbody version 60

and 15 secs later all the other WU did start again :

Lun 18 jul 08:59:16 2011 | World Community Grid | Restarting task CMD2_2051-2IAE_B.clustersOccur-3D1M_A.clustersOccur_6_1 using hcmd2 version 640
Lun 18 jul 08:59:16 2011 | Einstein@Home | Restarting task h1_0303.55_S6GC1__1179_S6BucketA_0 using einstein_S6Bucket version 101
Lun 18 jul 08:59:16 2011 | ibercivis | Restarting task Mr.Wilson_18_04_44_25_372159899_2071232659_1 using wilson version 6
Lun 18 jul 08:59:16 2011 | Poem@Home | Restarting task poempp_gvpj_1310904715_747262093_0 using poempp version 6
Lun 18 jul 08:59:16 2011 | Test4Theory@Home | Restarting task uc_1310460186_4776_0 using cernvm version 601
Lun 18 jul 08:59:16 2011 | malariacontrol.net | Restarting task wu_1167_35_10675_0_1310959567_0 using openMalariaB version 657
Lun 18 jul 08:59:16 2011 | Leiden Classical | Restarting task wu_898976128_1309873376_21394_0 using classical version 556
Lun 18 jul 08:59:16 2011 | NFS@Home | Starting task S2m1061a_508041_0 using lasievef version 108

with 15 secs of run time and then going back to sleep, I really don't think it's a normal behavior...

Also, interesting fact maybe, it remained in memory, using no CPU (the setup of my boinc is that WUs don't remain when suspended), and it's the only one like that in almost all my projects (appart from enigma where this is not working well neither).
19) Message boards : Number crunching : N-Body + Mac OS X + i7 = no good ? (Message 49345)
Posted 15 Jun 2011 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Hi,

I'm running boinc 6.10.58 on an iMac i7 10.6.7 and I realized yesterday that N-Body 0.40 multithreaded was causing a real problem : the WU (this one) would run normally, keeping 8 cores busy, then suddenly (after a too short time compared to my 240 mn param before switching app) it would stop and now the weirdest part : boinc wasn't able anymore to allocate the 8 cores to the other projects running (I have many projects running at the same time), a variable number of WUs (from 3 to 5) would be running... then after some time same story, N-body running with 8 cores, stop, another different number of WU of other projects restarting...

I tried to stop / restart the client, kill manually the processes, reboot the mac : same stuff.

Then I suspended milkyway and all went back to normal since then.

I see there is no parameter to select project inside milkyway so I cannot block N-body, so I'll keep milkyway suspended for the moment...

Did anybody ever experience this on Mac OS X ?

Thanks.
20) Message boards : Cafe MilkyWay : Race on MilkyWay (Message 3272)
Posted 21 Apr 2008 by Profile[AF>Le_Pommier] Jerome_C2005
Post:
Hi all,

our RAID is over and I think it gave a significant boost to MilkyWay overall production, and brought it to a new level.

Thanks a lot for all those who participated, and especially to the project admin(s) who allowed us to do such a raid on their project by providing a stable environment to receive such new level of production - with less than 2 days offline, which is not bad considering the huge increase of activity.

Cheers.
Jerome




©2021 Astroinformatics Group