Posts by WendyR
|
1)
Message boards :
Number crunching :
Project down?
(Message 34408)
Posted 9 Jan 2007 by WendyR Post: Having a load of problems down loading. Whole load failwd with checksum error and error 400/500. I'm seeing problems with downloads too, but I am getting "http error"s rather that your checksum errors. I'm also seeing errors on uploads to SETI, but I think that's the same problem that they have been having off and on for a while now.... [edit]As soon as I hit the OK button, my Rosetta downloads went through.[/edit] Wendy |
|
2)
Message boards :
Number crunching :
Restarting Results ad infinitum II
(Message 29517)
Posted 17 Oct 2006 by WendyR Post: I have seen those "exited with zero status but no 'finished' file" errors quite a few times too. I found there were a couple of events that "triggered" them in my case. I found that I got one of those each time I closed the cover on my laptop to move it somewhere else. That triggers a "go to sleep" mode in my laptop, and exactly which order that things happen during the "sleep" and the corresponding "wakeup" process are probably causing this. I also get that message when I try to do something with the "client_state.xml" file at the same time that the BOINC manager wants it. In my case, I was opening "BOINC Debt Viewer" when BOINC decided to switch between tasks. I know that other people are doing stuff with grep to monitor things in that file. Are you running something else that is attempting to look at that file? Is your virus checking software examining that file a lot because it is getting changed regularly? Do you run some automated backup or indexing software that is hitting that file regularly? Just some things to think about... |
|
3)
Message boards :
Number crunching :
More seeds or more decoys per seed?
(Message 28725)
Posted 30 Sep 2006 by WendyR Post: At what point are decoys on a new WU more valuable than additional decoys on an original WU? Realistically, only ONE structure really matters -- that global minimum energy structure. To say the rest don't matter is perhaps a little too strong a statement, but its often not too far from the truth. Think of it as a sort of a lottery. There are typically millions of different combinations out there, but only one winning combination of numbers. Sometimes, the ones that get some of the numbers right are even worth something, but not as much as that winning ticket. So, how many lottery tickets have to be bought to "guarantee" a winner? (all possible combinations) Can there be a winner if there is only one ticket sold? (yes -- not likely, but still possible) How does one decide "how many tickets to buy"? Well, you start with a "seat of the pants" type of guess. You look at the number of variables involved. In the Rosetta case it would be related to the size and flexibility of the protein in question. Also, remember that a lot of the Rosetta work has been done on proteins of known structure (they know the "right answer"). Doing work like this will help them to refine the number of structures to request for proteins of unknown structure. As with all scientific models, the more data that you incoroporate from diverse areas, the better your model, and subsequent predictions, become. I also think that different people in the group are working on different proteins, and there is value in keeping all of them busy. |
|
4)
Message boards :
Number crunching :
Definition of "cpu time" in Boinc Manager
(Message 25895)
Posted 2 Sep 2006 by WendyR Post: You should get hold of Process Explorer by sysinternals (now owned by Microsoft) I would also recommend TCPView and pagedfrg by the same folks. TCPView will show which processes have TCP/UDP ports open. (I like to watch this as BOINC uploads and downloads workunits.) It should also give you some indication if your machine has been compromised and is being used to connect to foreign hosts. Pagedfrg will degragment some critical system files. The pagedfrg will only work before those system files get opened (i.e. before you fully boot). You can set it up to defragment those files once, or every time you reboot your machine. |
|
5)
Message boards :
Number crunching :
Putting a Bell on the Cat
(Message 25681)
Posted 31 Aug 2006 by WendyR Post: Or just post in a different color, to make moderator text jump out Whatever you decide to do, just make sure that you all do it consistently. Most people will "figure it out" in a reasonable amount of time, but if you keep changing the rules, it, very quickly, loses effectiveness. |
|
6)
Message boards :
Number crunching :
Ok... I have a challenge for all of you...
(Message 22436)
Posted 14 Aug 2006 by WendyR Post: Please tell me in your own words: I have this strange desire to go back to the original question with this one. Yes, this is an oversimplification of "reality". I know that. If you really want that kind of detail, I would point you to several million pages of text books and scientific journals out there. How do you find a cure for a disease? How do you make a drug that cures a disease? In both cases, historically, we have "tried things until we find something that works". Again, historically, most of the "breakthroughs" in these areas were "blind luck", a combination of the right person noticing some detail at the right time. One example is the discovery of penicillin -- a scientist noticed that one particular mold culture was killing bacteria in petri dishes.... Today, we have found a different path that seems to work better. We try to understand the disease. What happens in the body when a foreign invader (bacteria, virus, cancer, etc.) arrives? How does a foreign invader's biochemistry differ from the human host's? Has some normal process "broken down" or "sped up" in a sick person's biochemistry? All of these processes involve chemical reactions in the body, and all of these chemical reactions happen because specific proteins are present to "help them along". They help the reactions in several different ways. They may help get the right "pieces" to the right places, or they may help a chemical reaction along by holding the two pieces that are supposed to react at just the right distance and angle for the reaction to happen. But all of the different pieces need to be in the right place, at the right time, under the right conditions. Most commonly, when we try to design a new drug, we try to inhibit one of the reactions in the body. The hope is that it is a reaction that is causing the disease condition we are trying to correct. So, if we can find, and stop, a chemical reaction that a bacteria needs to survive, but a human does not, the bacteria will die, and the human will [hopefully] remain healthy. If we can find a drug to stop the chemical reactions that make blood clot, we can prevent the blood clots that become strokes and heart attacks. If we inhibit the right reaction too strongly, or inhibit the wrong chemical reaction, or inhibit reactions other than the one we are targeting, the person may have other problems [side effects]. One of the critical pieces of this bigger process is the binding of the "small molecule" in the reaction to the protein that is going to "help" it along. This process is often compared to a "lock and key". While not particularly accurate, it is a useful model. The protein is the "lock" part, and the small molecule that is reacting is the "key" part. The goal now becomes to design a new "key" (a drug) to fit into the right "lock" (the protein), and hopefully stay around for a while. With the "locks" all filled with the drug molecules, the reaction in question is slowed down, and hopefully the person feels better. So, how do you design a "key". One way to start is to get a picture of the "lock" in question. There are a couple of different ways that a protein's structure can be determined. The two I am most familiar with are x-ray crystallography and NMR. Both have their drawbacks, and not all proteins can have their structures determined using these methods. Both also can take several months of hard work to determine the structure. Both also require several pieces of expensive equipment. Plus, there are way more proteins out there than we are able to determine structures for. Here, is where software like Rosetta can enter the picture! The ability to accurately, quickly [and cheaply] predict the structure of a protein could speed up the drug discovery processes tremendously! Since the protein structure determination is so very basic and fundamental to almost all diseases and all drug discovery, its value would be incredible! Again, this is incredibly simplified, but I hope also understandable. |
|
7)
Message boards :
Number crunching :
New credit system now being tested at RALPH@home
(Message 22141)
Posted 9 Aug 2006 by WendyR Post: The biggest problem I see with "zeroing out" everyone existing credits is that is becomes a "standard". What happens on the next new version? What if someone does some great optimization on some existing code and it changes the relative credit/hour on different processors? Basically, any PERCEIVED change in someone's credits/hour is going to result in cries of "zero the credits out again". Remember, the Rosetta software is constantly changing and evolving by design. You have to take that into account too. Some people may have had an advantage running one particular version at some given time frame, but that particular advantage may disappear with the next release. |
|
8)
Message boards :
Number crunching :
Type of Protein??
(Message 22049)
Posted 8 Aug 2006 by WendyR Post: Oh thanks for that. I can never remember (because I don't fully undertand) if it's 1 or 0... either way 1.0 is pretty good ;) The way I remember it....RMSD=Root Mean Square Deviation (or sometimes Distance or Difference)... The "smallest" possible difference (or distance) is always going to be 0 -- or "the same". I found a pretty good explaination of it here |
|
9)
Message boards :
Number crunching :
Deleters stuck?
(Message 11462)
Posted 27 Feb 2006 by WendyR Post: I have several workunits that were completed on February 14 and 15. I would have thought that these workunits would have been deleted by now (February 28). Isn't the "normal" behavior to delete workunits about a week after completion? See here for mine. I did some poking around, and found that I am not the only one with workunits from those dates still hanging around.... |
|
10)
Message boards :
Number crunching :
Progress updating every 2%
(Message 10699)
Posted 12 Feb 2006 by WendyR Post: Cool. Instead of updating only every 10%, my rosetta workunits are now updating their percent complete every 2% |
©2024 University of Washington
https://www.bakerlab.org