1) Forums : General Topics : Why do you participate in Cosmology@Home? Who is the \"most surprising\" participant? (Message 6298)
Posted 8 Jun 2008 by Profile Mr.Pernod
Post:
why did I join?
well, I read fantasy (Terry Pratchett),
I play fantasy (GuildWars),
I listen to fantasy (Ayreon),
so why not crunch fantasy ;-)
2) Forums : Technical Support : Pending Because Work Unit Remains "Unsent" (Message 3000)
Posted 30 Sep 2007 by Profile Mr.Pernod
Post:
Windows / AMDAthlonXP
Windows / AMDAthlonMP
Windows / IntelCore2
Windows / IntelXeon
Windows / Intel

these all seem to get "Message from server: (there was work but it was committed to other platforms)", so they should all have wingmen by now.


PS. Scott, could you please look into the AthlonMP? it still seems to be classified as AMDAthlonMP instead of AMDAthlonXP.
3) Forums : Technical Support : Odd Result (Message 2950)
Posted 26 Sep 2007 by Profile Mr.Pernod
Post:
did you by any chance restart BOINC or the computer during this particular result?
4) Forums : Technical Support : proposal new "AMD"section in hr.C (Message 2947)
Posted 26 Sep 2007 by Profile Mr.Pernod
Post:
    if (strcasestr(host.p_vendor, "AMD")) { 

        if (strcasestr(host.p_model, "Opteron")) return AMDOpteron;
        if (strcasestr(host.p_model, "Athlon")) { 
            if (strcasestr(host.p_model, "XP"))  return AMDAthlonXP; 
            // for Athlon MP, uncomment one of the following lines,
            // 
            // if (strcasestr(host.p_model, "MP"))  return AMDAthlonXP; 
            // if (strcasestr(host.p_model, "MP"))  return AMDAthlonMP; 
            if (strcasestr(host.p_model, "64"))  return AMDAthlon64; 
            return AMDAthlon;
        }

        if (strcasestr(host.p_model, "Family 6 ")) {
            if (strcasestr(host.p_model, "Family 6 Model 6")) return AMDAthlonXP;
            if (strcasestr(host.p_model, "Family 6 Model 7")) return AMDAthlonXP;
            if (strcasestr(host.p_model, "Family 6 Model 8")) return AMDAthlonXP;
            if (strcasestr(host.p_model, "Family 6 Model 10")) return AMDAthlonXP;
            return AMDAthlon;
        }

// for Durons not caught by "Family 6 " uncomment one of the following lines
//      if (strcasestr(host.p_model, "Duron")) return AMDDuron;
//      if (strcasestr(host.p_model, "Duron")) return AMDAthlon;
//      if (strcasestr(host.p_model, "Duron")) return AMDAthlonXP;

        if (strcasestr(host.p_model, "Family 15")) {
            if (strcasestr(host.p_model, "Family 15 Model 33")) return AMDOpteron;
            if (strcasestr(host.p_model, "Family 15 Model 55")) return AMDAthlon64;
            if (strcasestr(host.p_model, "Family 15 Model 65")) return AMDOpteron;
            if (strcasestr(host.p_model, "Family 15 Model 5")) return AMDOpteron;
            return AMDAthlon64;
        }

        if (strcasestr(host.p_model, "Family 16 Model 2")) return AMDOpteron;
        if (strcasestr(host.p_model, "Family 16")) return AMDAthlon64;
//      if (strcasestr(host.p_model, "Family 17")) return AMDAthlon64;
       

// for Semprons not caught by "Family 6" or "Family 15" uncomment one of the following lines
// no advise here, it's the projects choice.
//      if (strcasestr(host.p_model, "Sempron")) return AMDSempron;
//      if (strcasestr(host.p_model, "Sempron")) return AMDAthlonXP;
//      if (strcasestr(host.p_model, "Sempron")) return AMDAthlon64;

// for Turions not caught by "Family 15" uncomment one of the following lines,
// no advise here, it's the projects choice.
//      if (strcasestr(host.p_model, "Turion")) return AMDTurion;
//      if (strcasestr(host.p_model, "Turion")) return AMDAthlon64;

        return AMD;
    }



  • make sure it's AMD.
  • force-filter Opteron by name. This select puts all Opterons together, while the following "Family 15" only seperates socket 940 and socket F Opterons from Athlon64.
  • force-filter Athlon/AthlonXP/AthlonMP/64 by name (pick a classification for the AthlonMP in this section).
  • filter "Family 6 " for the Athlon or AthlonXP based "Unknown CPU Type", this will sort them into AMDAthlon or AMDAthlonXP depending on their model. Only AthlonXP-based models are filtered out to AMDAthlonXP, the remaining family-members are AMDAthlon.
  • force-filter Durons by name in case they fell through the "Family 6 " selection, pick one. If you want to force a classification on the Duron, put this before the "Family 6 " selection.
  • filter "Family 15". Only the exceptions for the Opteron 2xx, 8xx, 2xxx and 8xxxx (socket 940 and socket F) have been made. Opteron 1xx and 1xxx can't be distinguished from their Athlon64 counterparts as they use the same model number (and sit in the same socket 939 or AM2). This selection will put Athlon64, Turion and Athlon64-based Semrpons in the AMDAthlon64 group.
  • "Family 16 Model 2" is the new Barcelona based quad-core Opteron, further models are unknown (to me) at this time.
  • a bit of futureproofing, Barcelona and Phenom based Opterons and Athlon64's are classified as AMDAthlon64.
  • even more futureproofing, "Family 17 Model 0" and "Family 17 Model 3" have been spotted with the "AMD64" suffix, so guess what...use at your own discretion ;-)
  • force-filter Semprons by name in case they fell through the "Family 6 " and "Family 15" selections, pick one. If you want to force a classification on the Semrpon, put this before the "Family 6 " selection.
  • force-filter Turions by name in case they fell through the "Family 15" selection, pick one. If you want to force a classification on the Duron, put this before the "Family 15" selection.



Hope this is of use (and I didn't forget any { or} or ; or....)


[EDIT]
lovely.... the code-tag removes leading spaces and the pre-tag adds extra lines
[/EDIT]

5) Forums : Technical Support : proposal new "AMD"section in hr.C (Message 2946)
Posted 26 Sep 2007 by Profile Mr.Pernod
Post:
Since I have some spare time to burn lately I have been looking into the AMD section of the hr.C
After a lot of research I have come to the following idea.

There are currently 2 mayor processor families for AMD in the BOINC projects,
namely "Family 6" for the Athlon, AthlonXP and derivatives
and "Family 15" for the K8 (Athlon64 and it's derivatives)
some "Family 16" have been spotted on BOINC projects and in the near future more of these will start to appear as these are the new Barcelona / Phenom based

Opterons and Athlon64's

Athlon, Duron series (Thunderbird, Spitfire cores)
"Family 6 Model 0"     Athlon
"Family 6 Model 1"     Athlon
"Family 6 Model 2"     Athlon
"Family 6 Model 3"     Duron
"Family 6 Model 4"     Athlon


AthlonXP, AthlonMP, Duron, Sempron series
(Palomino, Thouroughbred, Barton, Applebred, Thorton, Morgan, Camaro cores)
"Family 6 Model 6"     Athlon XP, Athlon MP
"Family 6 Model 7"     Duron
"Family 6 Model 8"     Athlon XP, Athlon MP, Sempron
"Family 6 Model 10"    Athlon XP, Athlon MP, Sempron


Athlon64, Opteron, Turion, Sempron series
"Family 15 Model 3"       Athlon64
"Family 15 Model 4"       Athlon64,     Turion
"Family 15 Model 5"       Opteron 2xx,  Opteron 8xx
"Family 15 Model 7"       Athlon64,     Sempron
"Family 15 Model 12"      Athlon64,     Mobile Athlon64
"Family 15 Model 15"      Athlon64,     Mobile Sempron
"Family 15 Model 28"      Sempron,      Mobile Sempron
"Family 15 Model 31"      Athlon64,     Sempron
"Family 15 Model 33"      Opteron 2xx,  Opteron 8xx
"Family 15 Model 35"      Athlon64,     Opteron 1xx      (Socket 939)
"Family 15 Model 36"      Turion,       Mobile Athlon64
"Family 15 Model 39"      Athlon64,     Opteron 1xx      (Socket 939)
"Family 15 Model 43"      Athlon64
"Family 15 Model 44"      Athlon64,     Sempron
"Family 15 Model 47"      Athlon64,     Sempron
"Family 15 Model 55"      Athlon64
"Family 15 Model 63"      Athlon64
"Family 15 Model 65"      Opteron 2xxx, Opteron 8xxx
"Family 15 Model 67"      Athlon64,     Opteron 1xxx     (Socket AM2)
"Family 15 Model 72"      Turion
"Family 15 Model 75"      Athlon64
"Family 15 Model 76"      Turion,       Sempron
"Family 15 Model 79"      Athlon64,     Semrpon
"Family 15 Model 95"      Athlon64,     Sempron
"Family 15 Model 104"     Turion
"Family 15 Model 107"     Athlon64


These family lists show how closely the current HR-classes the AMD cpu's are related,
so I have come up with a more detailed selection set, similar to the x86 section for the Intel CPU's.
6) Forums : Announcements : New Website Design, Part Deux (Message 2841)
Posted 19 Sep 2007 by Profile Mr.Pernod
Post:
Just curious but the wallpapers with the 3 spheres on it, do they represent anything? I see one looks like a universe and the other is Cosmo but what does the green one represent?

P.S. Sorry I do not use them only because my wife thinks I am wierd for doing BOINC projects. Women. (They can't all be like you Kathryn) :)

the green one looks like a representation of the Cosmic Microwave Background radiation.

PS. if there was a wallpaper like a dark gray or black background and a graphic like the centre or righthand side of the banner on http://cosmos.astro.uiuc.edu/ and the C@H logo I would probably use it.
7) Forums : Technical Support : Valid WU got invalidated by project (Message 2698)
Posted 13 Sep 2007 by Profile Mr.Pernod
Post:
judging from the errors-status on the wu-page, a bunch of workunits got cancelled by Scott.

application [b]CAMB[/b]
..
..
errors [b]Cancelled[/b]

8) Forums : Technical Support : Pending Because Work Unit Remains "Unsent" (Message 2374)
Posted 26 Aug 2007 by Profile Mr.Pernod
Post:
I will see if I can convince a few people to run some Linux.

I added two single-CPU Xeon Linux VM's, I already had them running on my Xeons but for a different (non-BOINC) purpose.
9) Forums : Technical Support : Download Errors ... !!! (Message 2281)
Posted 24 Aug 2007 by Profile Mr.Pernod
Post:
download error on wuid=402166

Cosmology@Home	8/24/2007 4:42:26 PM	[error] MD5 check failed for params_082107_111732_1.ini
Cosmology@Home	8/24/2007 4:42:26 PM	[error] expected 951cd732c4bd91f20644260687e519b9, got bfd2597af56292ff93429160108eaf5a
Cosmology@Home	8/24/2007 4:42:26 PM	[error] Checksum or signature error for params_082107_111732_1.ini


download error on wuid=402869

Cosmology@Home	8/24/2007 4:42:29 PM	[error] MD5 check failed for params_082107_113028_0.ini
Cosmology@Home	8/24/2007 4:42:29 PM	[error] expected e71125368156c4670100709d439a8019, got e750e76ae44f5203da0444ba908e3892
Cosmology@Home	8/24/2007 4:42:29 PM	[error] Checksum or signature error for params_082107_113028_0.ini


download error on wuid=402520

Cosmology@Home	8/24/2007 4:42:31 PM	[error] MD5 check failed for params_082107_112333_0.ini
Cosmology@Home	8/24/2007 4:42:31 PM	[error] expected 6c53286e7442e9a214d2b0faad54488b, got f8615aa2d136f5a1d8aec6ea4d71f132
Cosmology@Home	8/24/2007 4:42:31 PM	[error] Checksum or signature error for params_082107_112333_0.ini


it looks like these were all created between 21 Aug 2007 16:17:32 UTC and 21 Aug 2007 16:30:28 UTC
10) Forums : General Topics : Team Recruitment Thread (Message 1755)
Posted 5 Aug 2007 by Profile Mr.Pernod
Post:
For those who know that your PC is missing out if there's only one core inside.
How else could so few crunch so much?
Duallers current, past and aspirational are always welcome to join Team 2CPU.com


11) Forums : General Topics : More Wu's Please ... !!! (Message 1690)
Posted 1 Aug 2007 by Profile Mr.Pernod
Post:
ohiomike,
the matter at hand is not the 1000 WU/Day/CPU limit, but the limitation set by the project that hosts can not have more than 25 results in their local cache at any one time, independent of the WU/Day/CPU limit.
12) Forums : General Topics : More Wu's Please ... !!! (Message 1684)
Posted 1 Aug 2007 by Profile Mr.Pernod
Post:
wouldn't it be possible to link the number of results a host can cache at one time to the number of cpu's on that host, like the "Maximum daily WU quota per CPU"?
13) Forums : Technical Support : cam 1.24 Intel P4 Performance (Message 1559)
Posted 25 Jul 2007 by Profile Mr.Pernod
Post:
time /T displays only in the windows command-prompt, no input required
14) Forums : General Topics : Are we running production already? (Message 1555)
Posted 25 Jul 2007 by Profile Mr.Pernod
Post:
I don't really feel like finding the thread in which I already discussed it, but to answer your question, no, the results so far have mostly been for testing. One of the reasons is that the runs have too low accuracy for us to use, hence our desire to make longer, higher accuracy WUs with checkpointing.

The results are *real* results, it's just that we haven't yet achieved the accuracy that we need.

at which level of accuracy are we at the moment with 1.24 and the "long" workunits?
15) Forums : Technical Support : homogenous redundancy and invalid results (Message 1545)
Posted 25 Jul 2007 by Profile Mr.Pernod
Post:
How about this one then, Mr. Pernod? http://cosmos.astro.uiuc.edu/cosmohome/show_host_detail.php?hostid=78

That's the Xeon counterpart of the Q6600. So I guess it should also be put in the Core 2 category?

In my opinion it should, as the Xeon X3220 is based on the same architecture as the Q6600.
Even though it is a Xeon, the basic architecture of this CPU is as different from my Xeons as the Core2 desktop cpu's are from the Pentium4's, which could potentially lead to different results on the same workunit.
16) Forums : Technical Support : homogenous redundancy and invalid results (Message 1541)
Posted 25 Jul 2007 by Profile Mr.Pernod
Post:
well, actualy, what I mean is not subdividing the Xeons, but matching them to their desktop or mobile counterparts.

for example, the two Xeon system I have listed are based on the P4 Northwood desktop cpu, which places them in the

if (strcasestr(host.p_vendor, "Intel")) {
..
   if (strcasestr(host.p_model, "x86")) {
..
..
      if (strcasestr(host.p_model, "Family 15 Model 2")) return IntelPentium4;
..
..
   }
..
}


while host 509 for example is the SMP-capable version of host 446, which should place host 509 into the same group as host 446.

if (strcasestr(host.p_vendor, "Intel")) {
..
  if (strcasestr(host.p_model, "Core")) return IntelCore2;
..
}

but due to the setup of the filter, that will not happen


arf

ok

I think this might be a job for the BOINC-developers to sort out, as it may impact more (especially small) projects that use the cpu_fine section of the hr.C
17) Forums : Technical Support : Didn't need (Message 1512)
Posted 25 Jul 2007 by Profile Mr.Pernod
Post:
if you check the first few lines on the workunit page, you will see that the workunit has been cancelled.

during the last weekend, there was a batch of faulty work released.
While trying to cancel those workunits Scott made a typo and cancelled all work from the 21st (wu_072107_*) instead of the faulty work dated the 20st.
for more info, check the Error on longer WU thread
18) Forums : Technical Support : Lots of compute errors. (Message 1503)
Posted 25 Jul 2007 by Profile Mr.Pernod
Post:
Error on longer WU
19) Forums : Technical Support : homogenous redundancy and invalid results (Message 1500)
Posted 25 Jul 2007 by Profile Mr.Pernod
Post:
Scott,

I just had my first invalid result where a P4-Xeon (x86 Family 15 Model 2 Stepping 9) was teamed up with a core 2 based Xeon (x86 Family 6 Model 15 Stepping 7).
When looking at the source of the hr.C posted in the BOINC trac, I notice that all cpu's that have "Xeon" in their p_model record are put into one category (like in the example workunit I linked to), while the P2, P3, P4 and Core based Intels are all split into different categories.
Will this generalization of the different Xeon architectures be causing issues when the project starts using higher precision in the work later on?
20) Forums : Technical Support : Error on longer WU (Message 1440)
Posted 23 Jul 2007 by Profile Mr.Pernod
Post:
Scott, I think you may have made a slight typo cancelling work

it looks like you cancelled wu_0721_1* instead of wu_0720_13 / 14 / 15 etcetera

Sigh.

It's back to remedial reading comprehension classes for me. Sorry about that. The correct ones have now been canceled.


lol

oh well, look at it this way, at least now we know for sure the <send_result_abort> works for serverside cancelled results ;-)


Next 20