TAS-member thread for August 2008: Milkyway@Home

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Rudy Toody

Diamond Member
Sep 30, 2006
4,267
421
126
Error WUs

I think it is a bad batch of WUs that are only attracted to me! Actually, they error out on others, too. Unless they are doing some kind of homogeneous distribution, I don't see how the bad ones can find just me.
 

biodoc

Diamond Member
Dec 29, 2005
6,284
2,238
136
Originally posted by: Rudy Toody
Error WUs

I think it is a bad batch of WUs that are only attracted to me! Actually, they error out on others, too. Unless they are doing some kind of homogeneous distribution, I don't see how the bad ones can find just me.

May result from I/O read errors. After some searching it's not that uncommon. I saw a couple of different projects suggest to "reset project" which would clean out all milkyway files/WUs etc. Have you tried that yet? Or perhaps a clean install?

Other than that, who knows? Odd that it happened on all 3 machines though. Maybe you're right, all three got a bad set of WUs.
 

Bok7575

Member
Mar 6, 2004
89
9
66
Originally posted by: biodoc

Bok says this:

"It's not resetting for some reason during the parse. So when it comes across a team in the xml where there is no <description>xxx</description> it's reeuing the last value of the variable..

I'll get it fixed today."

If you can get TA_Andy (captain of team at milkyway) to write a description of the TeAm like "proof that intelligent...", then that would appear in it's place.

:beer:

*should* be fixed after the next update to each project throughout all the stats.. Yeah, it's something I'd been collecting for a while as it's in the xml. I just never displayed it. Put it in this week and it showed up this bug. I'll probably tweak it some more, but I think it looks quite nice, especially if a good team description has been put in on the project sites (ours needs a lot or work too.. )

Bok

 

petrusbroder

Elite Member
Nov 28, 2004
13,346
1,146
126
Thanks, Bok, for looking in.
We appreciate your being here and telling us.
We also appreciate your stats-site a lot! :D
 

petrusbroder

Elite Member
Nov 28, 2004
13,346
1,146
126
Here is a stats update:

___ Date _____ Position _____ Credits ______ Production ____ Change
________________________ in August ______ in August ___ in position
July 31, 2008 ____ 57 ______ 341 090 _____ _______ 0 ________ 0
Aug 01, 2008 ____ 51 ______ 422 983 ________ 81 893 ________ 6
Aug 02, 2008 ____ 45 ______ 546 170 _______ 205 080 _______ 12
Aug 03, 2008 ____ 41 ______ 670 075 _______ 247 092 _______ 16
Aug 04, 2008 ____ 35 ______ 814 123 _______ 473 033 _______ 22
Aug 05, 2008 ____ 33 ______ 947 125 _______ 606 035 _______ 24
Aug 06, 2008 ____ 31 __ __ 1 085 319 _______ 744 229 _______ 26
Aug 07, 2008 ____ 29 __ __ 1 260 296 _______ 919 206 _______ 28
Aug 08, 2008 ____ 27 __ __ 1 439 031 _____ 1 097 941 _______ 30
Aug 09, 2008 ____ 26 __ __ 1 644 136 _____ 1 303 046 _______ 31
Aug 10, 2008 ____ 22 __ __ 1 816 417 _____ 1 475 327 _______ 35

Looks just terrific, doesn't it? :thumbsup:

Daily production:
Aug 06: ___ 138 194 credits _____ +5 192 (compared to previous day)
Aug 07: ___ 174 977 credits ____ +36 783
Aug 08: ___ 178 735 credits _____ +3 758
Aug 09: ___ 205 105 credits ____ +26 370 ---> best day!
Aug 10: ___ 172 281 credits ____ - 32 824

We will pass 5 more teams in the next five days (we passed 4 today), and seven additional teams (a total of 12 teams) within the next 20 days.

So, lets crunch some more! :D
 

caferace

Golden Member
May 31, 2005
1,472
6
76
Originally posted by: petrusbroderJim, can you not add 10 more cores? Or doesn't MilkyWay play nice with the imbeddded Win?

Unfortunately, the company changed the mobo package in our embedded product from PIII based to EPIA Centaur Hauls based. In essence, the boxes now have about 30% the processing power of the PIII's although they are silent and power friendly. Great for our customers but not so good for me/us. That, and two of my fleet are still AWOL, though at least one will be reincarnated next weekend.

-jim

 

biodoc

Diamond Member
Dec 29, 2005
6,284
2,238
136
Congrats to Alyx!! 500,000 points in 10 days!!:beer:

:thumbsup::beer:
 

Rudy Toody

Diamond Member
Sep 30, 2006
4,267
421
126
I have tried again today to get some good WUs. Errored in 1 sec each. I appears that others have problems with the same WUs. I posted a message on the Milkyway forum.

There is no co-relation to Intel/AMD, Windows/Linux, 64Bit/32Bit, and BOINC manager/client versions.

Edit: I have upgraded all quads to BOINC 6.2.14. Everything but Milkyway works!
 

petrusbroder

Elite Member
Nov 28, 2004
13,346
1,146
126
Here is a stats update:

___ Date _____ Position _____ Credits ______ Production ____ Change
________________________ in August ______ in August ___ in position
July 31, 2008 ____ 57 ______ 341 090 _____ _______ 0 ________ 0
Aug 01, 2008 ____ 51 ______ 422 983 ________ 81 893 ________ 6
Aug 02, 2008 ____ 45 ______ 546 170 _______ 205 080 _______ 12
Aug 03, 2008 ____ 41 ______ 670 075 _______ 247 092 _______ 16
Aug 04, 2008 ____ 35 ______ 814 123 _______ 473 033 _______ 22
Aug 05, 2008 ____ 33 ______ 947 125 _______ 606 035 _______ 24
Aug 06, 2008 ____ 31 __ __ 1 085 319 _______ 744 229 _______ 26
Aug 07, 2008 ____ 29 __ __ 1 260 296 _______ 919 206 _______ 28
Aug 08, 2008 ____ 27 __ __ 1 439 031 __ ___ 1 097 941 _______ 30
Aug 09, 2008 ____ 26 __ __ 1 644 136 __ ___ 1 303 046 _______ 31
Aug 10, 2008 ____ 22 __ __ 1 816 417 __ ___ 1 475 327 _______ 35
Aug 11, 2008 ____ 20 __ __ 1 984 315 __ ___ 1 643 225 _______ 37

Looks just terrific, doesn't it? :thumbsup:

Daily production:
Aug 06: ___ 138 194 credits _____ +5 192 (compared to previous day)
Aug 07: ___ 174 977 credits ____ +36 783
Aug 08: ___ 178 735 credits _____ +3 758
Aug 09: ___ 205 105 credits ____ +26 370 ---> best day!
Aug 10: ___ 172 281 credits __ __ -32 824
Aug 11: ___ 167 898 credits __ ___ -4 383

We will pass 3 more teams in the next five days (we passed 2 today), and seven additional teams (a total of 10 more teams) within the next 18 days. If all turns out allright we will end up in position 11 or 10 ...

So, lets crunch some more! :D
 

Rudy Toody

Diamond Member
Sep 30, 2006
4,267
421
126
I was just told be a Milkyway cruncher who was trying to help me get going again, that all the devs have abandoned the project.

I'm trying to find a second source for that rumor. I'm thinking it might be only for the summer vacation---it is run by a university.

--Fred

Edit: I hope we haven't been crunching junk!
 

Philippart

Golden Member
Jul 9, 2006
1,290
0
0
two things are weird:
1st: the admins rarely reply to your messages and the answers are often vague
2nd: the frontpage states that the project is used to test the power of DC, the institute by which it is run has tons of computing power...
 

petrusbroder

Elite Member
Nov 28, 2004
13,346
1,146
126
Originally posted by: Philippart
two things are weird:
1st: the admins rarely reply to your messages and the answers are often vague
2nd: the frontpage states that the project is used to test the power of DC, the institute by which it is run has tons of computing power...

OTOH: (My emphasis):

from the project's home page:
Milkyway@home is a research project that uses Internet-connected computers to do research in modeling and determining the evolution of the Milkyway galaxy. You can participate by downloading and running a free program on your computer.

Milkyway@home is based at the Rensselaer Computer Science Department. This particular project is being developed to better understand the power of volunteer computer resources.

As I read the description, we are looking at the three goals of the project:
First aim: ... research in modeling ...
Second aim: ... determining the evolution of the Milkyway galaxy ...
Third aim: ... being developed to better understand the power of volunteer computer resource ...

Three different goals using the same sets of WUs. I think the first goal has the puropse to make the second goal possible, while the third goal is more related to infrastructure than to understanding the milky way's evolution. If I had written the description above I would have added the word "also" just before the third bold part of the sentence ... ;)

Probably the first two goals are the ones of the astronomy/physics department and the third one is the goal of the computer deprartment.

Anyhow: still very good reasons to crunch the project, and, as I see it, not a bogus ...
 

biodoc

Diamond Member
Dec 29, 2005
6,284
2,238
136
I looked at it too. It all seems legit to me.

Congrats on the 250K milestone Petrusbroder!!:beer:
 

Philippart

Golden Member
Jul 9, 2006
1,290
0
0
Originally posted by: petrusbroder
Originally posted by: Philippart
two things are weird:
1st: the admins rarely reply to your messages and the answers are often vague
2nd: the frontpage states that the project is used to test the power of DC, the institute by which it is run has tons of computing power...

OTOH: (My emphasis):

from the project's home page:
Milkyway@home is a research project that uses Internet-connected computers to do research in modeling and determining the evolution of the Milkyway galaxy. You can participate by downloading and running a free program on your computer.

Milkyway@home is based at the Rensselaer Computer Science Department. This particular project is being developed to better understand the power of volunteer computer resources.

As I read the description, we are looking at the three goals of the project:
First aim: ... research in modeling ...
Second aim: ... determining the evolution of the Milkyway galaxy ...
Third aim: ... being developed to better understand the power of volunteer computer resource ...

Three different goals using the same sets of WUs. I think the first goal has the puropse to make the second goal possible, while the third goal is more related to infrastructure than to understanding the milky way's evolution. If I had written the description above I would have added the word "also" just before the third bold part of the sentence ... ;)

Probably the first two goals are the ones of the astronomy/physics department and the third one is the goal of the computer deprartment.

Anyhow: still very good reasons to crunch the project, and, as I see it, not a bogus ...

I've tried to read it that way too, but the "This particular project" in the last sentence made me suspicious
 

petrusbroder

Elite Member
Nov 28, 2004
13,346
1,146
126
Thanks biodoc!

Oh, BTW, here is a stats update:

___ Date _____ Position _____ Credits ______ Production ____ Change
________________________ in August ______ in August ___ in position
July 31, 2008 ____ 57 ______ 341 090 _____ _______ 0 ________ 0
Aug 01, 2008 ____ 51 ______ 422 983 ________ 81 893 ________ 6
Aug 02, 2008 ____ 45 ______ 546 170 _______ 205 080 _______ 12
Aug 03, 2008 ____ 41 ______ 670 075 _______ 247 092 _______ 16
Aug 04, 2008 ____ 35 ______ 814 123 _______ 473 033 _______ 22
Aug 05, 2008 ____ 33 ______ 947 125 _______ 606 035 _______ 24
Aug 06, 2008 ____ 31 __ __ 1 085 319 _______ 744 229 _______ 26
Aug 07, 2008 ____ 29 __ __ 1 260 296 _______ 919 206 _______ 28
Aug 08, 2008 ____ 27 __ __ 1 439 031 __ ___ 1 097 941 _______ 30
Aug 09, 2008 ____ 26 __ __ 1 644 136 __ ___ 1 303 046 _______ 31
Aug 10, 2008 ____ 22 __ __ 1 816 417 __ ___ 1 475 327 _______ 35
Aug 11, 2008 ____ 20 __ __ 1 984 315 __ ___ 1 643 225 _______ 37
Aug 12, 2008 ____ 19 __ __ 2 125 612 __ ___ 1 784 522 _______ 38

Daily production:
Aug 06: ___ 138 194 credits _____ +5 192 (compared to previous day)
Aug 07: ___ 174 977 credits ____ +36 783
Aug 08: ___ 178 735 credits _____ +3 758
Aug 09: ___ 205 105 credits ____ +26 370 ---> best day!
Aug 10: ___ 172 281 credits __ __ -32 824
Aug 11: ___ 167 898 credits __ ___ -4 383
Aug 12: ___ 141 297 credits __ __ -26 601

We will pass 2 more teams in the next five days (we passed 1 today), and seven additional teams (a total of 10 more teams) within the next 17 days. If all turns out allright we will end up in position 11 or 10 ...

So, lets crunch some more! :D
 

petrusbroder

Elite Member
Nov 28, 2004
13,346
1,146
126
Originally posted by: Philippart
I've tried to read it that way too, but the "This particular project" in the last sentence made me suspicious

Yeah - but I - being a man who believes the best from my fellow humans - read this as being ... of all the projects the computer department runs (in contrast to the astronomy/physics department) "This particulat project is being developed to better understand the power of volunteer computer resources".

I think we could discuss this for ever. I agree that the sentence could make me think twice about the project.

But: If you look at this paper, you will see that the scientists are doing several things at one time: developing the maths for their project, developing the computational infrastructure for their project, and from the applicational point of view, also looking at the origin and structure of the Milky Way Galaxy as well as searching for missing baryons (see chapters 3.1 and 3.2 in this paper).
Although this paper does not present the results of the WU-crunching done by the MilkyWay@Home project (way too early for that) it is sufficiently motivating for me ...

If I understand the paper correctly, the MilkyWay-project is crunched by DC, the missing baryon project is being crunched by dedicted computers. Correct my if I am wrong, though...

 

Rudy Toody

Diamond Member
Sep 30, 2006
4,267
421
126
I have been discussing my problem over on the BOINC developers' forum and they told me that I am jinxed.

I told them I was going to crunch a non-BOINC project next month.
 

biodoc

Diamond Member
Dec 29, 2005
6,284
2,238
136
Originally posted by: petrusbroder
But: If you look at this paper, you will see that the scientists are doing several things at one time: developing the maths for their project, developing the computational infrastructure for their project, and from the applicational point of view, also looking at the origin and structure of the Milky Way Galaxy as well as searching for missing baryons (see chapters 3.1 and 3.2 in this paper).
Although this paper does not present the results of the WU-crunching done by the MilkyWay@Home project (way too early for that) it is sufficiently motivating for me ...

If I understand the paper correctly, the MilkyWay-project is crunched by DC, the missing baryon project is being crunched by dedicted computers. Correct my if I am wrong, though...
I agree 100%. These young graduate students are practicing the scientific paradigm:

1) develop a mathematical model to address a scientific question
2) validate the model with existing data
3) use the model to develop new hypotheses

They have published and presented the mathematical model/results in the scientific community and seem to be moving forward in a rational way.

The only issue I have with these 2 graduate students is they seem to be on vacation!!

Nathan (astronomer) 19 days since last post
Travis (computer science) 45 days since last post

Give me a break...you should not be taking a vacation when you are in graduate school!!:laugh:

If you do...dial in and answer the questions on the message board in the morning and then head to the beach for the rest of the day!!

:beer:
 

Fardringle

Diamond Member
Oct 23, 2000
9,197
763
126
Originally posted by: Fardringle
Originally posted by: biodoc
Maybe I can catch Fardringle?!?;)

Probably. Our daily production is VERY close right now. A few points either way could make the difference. :)

I lost two machines today (both dual P4s) so my points will probably drop quite a bit until the replacement machines arrive. I guess this is your chance to pass me in the rankings. ;)
 

Wolfsraider

Diamond Member
Jan 27, 2002
8,305
0
76
Hmmm time to ramp up a bit in production since you are building a new one lol (cruncher)

I look forward to the race :)

Mik

Oh and ty I am only running 2 computers shouldnt be that hard to catch me :)
 

petrusbroder

Elite Member
Nov 28, 2004
13,346
1,146
126
Originally posted by: biodoc

(quote edited by petrusbroder)

The only issue I have with these 2 graduate students is they seem to be on vacation!!

Nathan (astronomer) 19 days since last post
Travis (computer science) 45 days since last post

Give me a break...you should not be taking a vacation when you are in graduate school!!:laugh:

If you do...dial in and answer the questions on the message board in the morning and then head to the beach for the rest of the day!!

:beer:

I agree. I have just now three graduate students. None of them is allowed vacations - only study time off site :shocked: .
If on the beach or in the mountains, check the e-mail at least once every second day. :roll:
If on the beach or after climbing the mountains, read at least one full (scientific) paper a day. :p
Report (at least 400 words) in my e-mail every second day (when they check in ...). That way I do not have to read those papers. :evil:

OTOH: All my graduate students (6 up to date) have got their degrees ahead of time; three of them are professors themselves now, all six are self supportive ... :thumbsup:





Just so you all know: :D a vacation is a period of time you are not allowed to do the work you enjoy. :laugh: