Multiple machines working on the same WU...wtf??

[timvek]

I have a meager 3 machines in my small patch, and just today, all three seem to have gotten the exact same WU.

Anyone know how to rectify this? Can I dump it off the 2 slowest and get new WUs with them?

Thanks,
Kev

 

[roftranspo]

Even if you do get the same run/clone/gen - the same exact work unit in all aspects - you are still helping them because you are verifing that the computations were 100% accurate. Stanford has checks in place to make sure that things do get computed correctly, that's why you hear stories of people overclocking and getting early unit ends, it's ending early because checks were in place to make sure that nothing major got corrupted. Though I'm sure that little errors in memory/cpu do sometimes happen, and these might be to small to get caught, and so they do need to send the same exact wu out a few times to check and make sure everythign matches up. So rest assured you are still helping the cause with whatever you fold.

Fold on with all your dual cpu goodness!

He says it all...multipule wu is just a check of the systems...you are doing just fine

 

[timvek]

Roger that. Good to know, I guess.

K

 

[MN Scout]

Yes, each protein is broken into parts - multiple sub wus. I tried to find what Stanford said on this issue, and I found in their FAQ this:

I just finished a WU and now I got another for the same protein. Is there something wrong?

No, most likely everything is fine. We're studying the dynamics of a few proteins, so you're likely to get the same protein to work on multiple times. Each WU gives us additional information about the dynamics of that protein, so it is important to us. Indeed, if we did only 1 WU per protein, we would not learn very much.

I've never read through there FAQ before. They have a lot of good information there. This quote comes from the section titled "Running". If you all haven't read it you should take sometime and do so, read both the "Topics" and "Other FAQ" links, very informative.

 

[gnewbury]

Since I'm bored:
If you are trunning Windoze.
What is an "exact same WU"
i.e.
Did you get ones with the exact same:
Project number?
like:
Protein: p1152_L939_K12M_ext_from638

or Project, Run, Clone, Gen?
like:
Protein: p1152_L939_K12M_ext_from638
Run: 176 (Clone 53, Gen 8)
if it's the latter double check your FAHlog.txt files files to make sure you have different user id's or at least different machine id's - look for the lines like these:
[22:19:32] - User ID: 557F1C627B1DC86
[22:19:32] - Machine ID: 6
no two user id's and machine number combo should be the same.
In this instance I'm running the number 6 instance form a machine which has the user id of 557F1C627B1DC86
If they are the same you may be duping work.

 

[timvek]

Turned out to be different Run, Clone Gen. Look exactly the same on the surface in EM3, tho. It fooled me, I'm just a lowly noob.

K

 

[Killer[MoB]]

Turned out to be different Run, Clone Gen. Look exactly the same on the surface in EM3, tho. It fooled me, I'm just a lowly noob.

That's good to hear. If it ever is the same R/C/G, that's a problem.
Welcome to the team.

 

[MN Scout]

That's good to hear. If it ever is the same R/C/G, that's a problem.
Welcome to the team.

That's not really a problem, it is possible to get the same R/C/G, just not likely. As I stated before they do have to verify that everything was computed correctly.