Page 1 of 2 12 LastLast
Results 1 to 40 of 73

Thread: Client test version available

  1. #1

    Exclamation Client test version available

    The client test version is now available fro the ftp site: ftp.mshri.on.ca/pub/distribfold/download. The versions available are just for Windows and Linux, respectively named foldtrajlite.zip and foldtrajlite.gz

    Get ftp://ftp.mshri.on.ca/pub/distribfol...ldtrajlite.zip Windows version or ftp://ftp.mshri.on.ca/pub/distribfol...oldtrajlite.gz Linux version (RedHat 7.1 and up).

    Simply replace the executable in the existing installation of the client (make sure to keep a backup version).

    Changes were made to the algorithm, primarily that generations are now 100 structures each, which may lead to better structures. We would appreciate detailed feedback about any bugs. Thanks to all those who volunteer their time to test!
    Last edited by Brian the Fist; 12-13-2003 at 06:09 PM.
    Elena Garderman

  2. #2
    Team Anandtech DF!

  3. #3
    Ancient Programmer Paratima's Avatar
    Join Date
    Dec 2001
    Location
    West Central Florida
    Posts
    3,296
    OK. Let's post to this thread (unless the Admins want to set up a special one), to keep things separate from the mainstream.
    HOME: A physical construct for keeping rain off your computers.

  4. #4
    got it installed, no problems as of yet.
    i will let it go and check on it when i get back home after work and report anything strange if it happens.
    Driving home the sky accelerates and the clouds all form a geometric shape.

  5. #5
    Member
    Join Date
    Apr 2003
    Location
    Germany
    Posts
    59
    just downloading it..

    i asked in an other thread, whether the handshaking while uploading was optimized?..
    would be nice to now the facts

  6. #6
    Boinc'ing away
    Join Date
    Aug 2002
    Location
    London, UK
    Posts
    982
    Running the Linux version here - reset the processing to gen 0 and already getting energy scores in the mid-20s...whether that translates to good RMS is another matter...

    I did notice that even though it's now 100 strucs/gen it didn't seem any slower - the tight spot processing seemed a lot faster than I've noticed before...

  7. #7
    running fine here. got a 5.55 energy at gen 5 already. hope that corresponds to a low RMSD.
    Team Anandtech DF!

  8. #8
    OCworkbench Stats Ho
    Join Date
    Jan 2003
    Posts
    519
    Got her running on my Dual 2400 MP, so far so good
    I am not a Stats Ho, it is just more satisfying to see that my numbers are better than yours.

  9. #9
    ./foldtrajlite: error while loading shared libraries: libstdc++.so.5: cannot open shared object file: No such file or directory

    I'm assuming this error is because of my old copy of mandrake. Any chance of getting a test copy that does not have this problem?

  10. #10
    25/25Mbit is nearly enough :p pointwood's Avatar
    Join Date
    Dec 2001
    Location
    Denmark
    Posts
    831
    Been running it for 3 hours or something so far and it seems to do pretty okay here too
    Pointwood
    Jabber ID: pointwood@jabber.shd.dk
    irc.arstechnica.com, #distributed

  11. #11
    Downsized Chinasaur's Avatar
    Join Date
    Dec 2001
    Location
    WA Wine Country
    Posts
    1,847
    Will there be an OS X version to test or just Linux?

    TIA.
    Agent Smith was right!: "I hate this place. This zoo. This prison. This reality, whatever you want to call it, I can't stand it any longer. It's the smell! If there is such a thing. I feel saturated by it. I can taste your stink and every time I do, I fear that I've somehow been infected by it."

  12. #12
    I updated 2 boxes today with the new test executable. Since then both boxes are reporting the same "Best Energy" for each generation. I am using dfGUI 3.2 to obtain these readings. Both boxes are runn ing Win XP and clients are installed as a service.

    The first box is a an Athlon XP 2600+ 512MB ram. I changes to the test version at Generation 208. Since then every generation has a "best energy" of 74.990. It is up to generation 220 atm.

    The 2nd box is a pentium III 733Mhz 512MB Ram. and it show the same best energy readings. IT was on generation 26 when i switched. It's now on generation 32. All energy at 74.990.

    Anyone have any ideas why this would occur?


    edit:
    Since i noticed this problem, i have been watching it more closely and have noticed the following things: (Related to the Athlon XP Box)

    The time per generation is always the same 17 minutes 30 seconds
    the "Best gen" is usually discovered in the first 10 structures and is 74.620
    then changes to 74.990 usually between the 40 and 50th structure (I suspect the 50th)

    It looks as thought it is repeating the same generation over and over but the gen # is changing each time normally. It's currently up to gen 230 so ive decided to let it finish to see what happens. Seems to be scoring normally too.

    I restarted the Pentium III box with a newly installed client and the test exe and it seem to be performing normally.


    Thanks

    Googlybear
    Last edited by Googlybear; 12-04-2003 at 10:00 PM.

  13. #13
    has been eaten by a grue.
    Join Date
    Jun 2003
    Location
    Detroit, MI
    Posts
    384
    For any other Linux newbies who are doing this too tired to think (like I was)...

    After downloading the new foldtrajlite.gz to your /distributed folding directory:

    1) go to your df directory, and stop the program (rm *.lock)

    2) cp foldtrajlite foldtrajlite.bak

    3) gzip -d foldtrajlite.gz (allow it to overwrite)

    4) chmod 777 foldtrajlite (so the dang thing will execute)

    5) start as you normally do (probably ./foldit)

  14. #14
    Has been working fine for a few hours.

    System: Opteron 146 / 1GB ECC Corsair PC3200 / WinXP pro

    ========================[ Dec 4, 2003 10:03 PM ]========================
    Starting foldtrajlite built Dec 4 2003
    Thu Dec 04 22:03:18 2003 ERROR: [001.001] {trajtools.c, line 3641} Unable to open trajectory distribution file pbxj5kr9_protein_251.trj
    Thu Dec 04 22:03:18 2003 FATAL ERROR: [002.003] {foldtrajlite2.c, line 5609} Unable to read trajectory distribution pbxj5kr9_protein_251, please create a new one

  15. #15
    TazAmdmb, gen251 is an extremely rare error that's been around for a while. If you can save a copy of your directory, I'm sure Howard would like a look at it.

  16. #16
    I started the Test Client on an XP2100+ at stock this morning (now 15 hours ago) and it seems to be both faster, despite more structures per gen, and getting lower energy results.

    I've now copied it to another 3 boxen, XP2000, XP2400 and Celeron 2.2 to see if the results are coming through on all fronts.

    So far so good. If it proves to be stable and as good as it looks to be, I'll move the rest of my boxen over to the new client/back to DF
    Crunching for OCAU

  17. #17
    I forgot to mention, please delete 'filelist.txt' and start over at generation zero with the new executable. If you try to continue a simulation in mid-run, the results may be 'unpredictable'. Anyhow, please do post any errors you find in the new version here (if you started from gen. o wqith the new executable). Remember you can upload borked directories to ftp://ftp.mshri.on.ca/pub/distribfold/incoming and send an e-mail to trades@mshri.on.ca to tell us the file you uploaded and what the problem was (such as the gen. 251 thing above).

    For all the questions I've ignored, the answer is 'no'.

    Thanks to those of you testing this out, hopefully it will avoid problems when we switch everyone to it.
    Howard Feldman

  18. #18
    Originally posted by Brian the Fist

    For all the questions I've ignored, the answer is 'no'.
    okay I will assume that is for the other clients. I hope that we won't have problems with these other clients when you actually go live. or are you indicating that we will not have these other clients when you go live?

  19. #19
    Ancient Programmer Paratima's Avatar
    Join Date
    Dec 2001
    Location
    West Central Florida
    Posts
    3,296
    Running fine on two Win2K boxes, but strange energy values. One is on gen 76 with a best energy of 79.620 and the other is on gen 65 with energy flat-lined at 24.990.

    But hey! No run-time errors.
    HOME: A physical construct for keeping rain off your computers.

  20. #20
    Boinc'ing away
    Join Date
    Aug 2002
    Location
    London, UK
    Posts
    982
    new client seems fine here - it's around 50% faster and I've dropped from a 11.22 RMS to a 9.96RMS in the day it's been running...

  21. #21
    Senior Member
    Join Date
    Jul 2003
    Location
    Hamburg/Germany
    Posts
    386
    same here, working fine and I already had a 8.323 in gen 10.
    especially gen 0 seemed to be faster...


    Greets Thor

  22. #22
    Member
    Join Date
    Apr 2003
    Location
    Germany
    Posts
    59



    picture schould be self-explaining.. ist that normal

  23. #23
    Senior Member
    Join Date
    Jan 2003
    Location
    North Carolina
    Posts
    184
    I tried getting the test version to run on one of my machines. I had been running the Redhat 7.1 Intel compiler verion of the client. The new executable (also supposedly for Redhat 7.1) immediatly dies with the error:

    ./foldtrajlite: /lib/i686/libpthread.so.0: version `GLIBC_2.3.2' not found (required by ./foldtrajlite)
    ./foldtrajlite: /lib/i686/libc.so.6: version `GLIBC_2.3' not found (required by ./foldtrajlite)



    This happened on both a Mandrake 9.0 and a SuSE 8.1 machine.

  24. #24
    I've got it to work on Mandrake 9.2 Beta RC.
    THE SPANKIN' PARTY
    Crunchin for
    Barbarian Frogs

  25. #25
    Senior Member
    Join Date
    Mar 2002
    Location
    MI, U.S.
    Posts
    697
    Originally posted by AMD_is_logical
    ./foldtrajlite: /lib/i686/libpthread.so.0: version `GLIBC_2.3.2' not found (required by ./foldtrajlite)
    ./foldtrajlite: /lib/i686/libc.so.6: version `GLIBC_2.3' not found (required by ./foldtrajlite)
    Getting the same error here, except without the /i686 part (my glibc was installed to just plain /lib).

    I have glibc 2.2.5, and I'm almost positive that's the problem. This new executable was most definitely not compiled on a RedHat 7.1 box, at least not a standard one...

    ()
    "If you fail to adjust your notion of fairness to the reality of the Universe, you will probably not be happy."

    -- Originally posted by Paratima

  26. #26
    Ol' retired IT geezer
    Join Date
    Feb 2003
    Location
    Scarborough
    Posts
    92

    Talking Linux Test Version works on Gentoo 1.4

    As the above subject says... Linux test version works fine on my Gentoo Linux generated using the 1.4 release. No fixes applied from release either.

    Ned

  27. #27
    Originally posted by AMD_is_logical
    I tried getting the test version to run on one of my machines. I had been running the Redhat 7.1 Intel compiler verion of the client. The new executable (also supposedly for Redhat 7.1) immediatly dies with the error:

    ./foldtrajlite: /lib/i686/libpthread.so.0: version `GLIBC_2.3.2' not found (required by ./foldtrajlite)
    ./foldtrajlite: /lib/i686/libc.so.6: version `GLIBC_2.3' not found (required by ./foldtrajlite)



    This happened on both a Mandrake 9.0 and a SuSE 8.1 machine.
    Same problem here on Mandrake 9.0. I sure it will also occur on my 8.2 blades, so I am not going to try.

    Most disturbing is that Mandrake hasn't release an RPM of this Glibc version for anything except Mandrake 9.2 . They have released 2.3.1 for Mandrake 9.1 , so 2.3.2 will probably work with Mandrake 9.1.

    But on 9.0, Mandrake only seems to want to go to GLibc 2.2.5. This sends up a red flag.

    So does anyone want to link me up with ( the warm fuzzies ) an rpm for glibc 2.3.2 for Mandrake 9.0?? I am worried that Mandrake avoided this glibc release for a reason.

  28. #28

    Linux beta cleint is a sucess.

    The beta client is going very well on all my linux boxen except earlier glibc versions where it wont run.
    The memory leak has gone away.
    The performance is way better than the last client, I am curious to know how you got such an increse on the same protein? Also any idea when other platforms will get the beta eg FreeBSD and MacOS X?

  29. #29
    Administrator PCZ's Avatar
    Join Date
    Jun 2003
    Location
    Chertsey Surrey UK
    Posts
    2,428
    The linux beta is going well for me as well
    I had already upgraded to glibc 2.3 because of other DC clients requiring it.

  30. #30
    Registered User
    Join Date
    Apr 2003
    Location
    Tigard, OR
    Posts
    2
    So far so good

    I'm running the beta on 1 W2K box, 2 XP boxes, and 3 linux boxes. I'll probably have another two or three more XP boxes converted over tomorrow.

    So far RMSD numbers are typically less than half of what I've seen with the current release. I assume that's a good thing... Best of all it's about twice as fast too

  31. #31
    Senior Member
    Join Date
    Mar 2002
    Location
    MI, U.S.
    Posts
    697
    Originally posted by PCZ
    I had already upgraded to glibc 2.3
    I would just do that, except that everything else that I have installed will break. I tried it once (with 2.3.1 when it was current), and it was not pretty. Luckily I had saved the old libc (and dynamic linker/loader, and libm, and everything else), and the cp binary didn't try doing anything complicated, so I could still go back to 2.2.5.

    2.3 is not at all binary compatible with 2.2 is the conclusion I was forced to draw...

    How did you get it to work?
    "If you fail to adjust your notion of fairness to the reality of the Universe, you will probably not be happy."

    -- Originally posted by Paratima

  32. #32
    Administrator PCZ's Avatar
    Join Date
    Jun 2003
    Location
    Chertsey Surrey UK
    Posts
    2,428
    bwkaz

    I use netboot clients with a RH 9 server.
    The server has glibc 2.3 as standard.
    The netboot image [LTSP] is based on RH 7.1 and had glibc 2.2.
    I updated the files in the LTSP tree with the libs and binaries from the RH server.

  33. #33
    Downsized Chinasaur's Avatar
    Join Date
    Dec 2001
    Location
    WA Wine Country
    Posts
    1,847

    Beta fine - Knoppix 2.4.22

    Hit the file with chmod 755 after downloading.

    Placed in a copied df folder, started right up, no errors.
    Agent Smith was right!: "I hate this place. This zoo. This prison. This reality, whatever you want to call it, I can't stand it any longer. It's the smell! If there is such a thing. I feel saturated by it. I can taste your stink and every time I do, I fear that I've somehow been infected by it."

  34. #34
    Senior Member
    Join Date
    Mar 2002
    Location
    MI, U.S.
    Posts
    697
    So basically, PCZ, you're running RH 9, not LTSP.

    Since you replaced all the libraries and binaries, there's nothing left from LTSP...
    "If you fail to adjust your notion of fairness to the reality of the Universe, you will probably not be happy."

    -- Originally posted by Paratima

  35. #35
    Administrator PCZ's Avatar
    Join Date
    Jun 2003
    Location
    Chertsey Surrey UK
    Posts
    2,428
    bwkaz

    Why don't you install a more up to date distro ?
    I am sure that the latest DF exe is not the only program that wont work for you.

  36. #36
    OCworkbench Stats Ho
    Join Date
    Jan 2003
    Posts
    519
    He is probably running it on a ZX81, and there is no newer Distro for that PC
    I am not a Stats Ho, it is just more satisfying to see that my numbers are better than yours.

  37. #37
    Ol' retired IT geezer
    Join Date
    Feb 2003
    Location
    Scarborough
    Posts
    92

    Question Energy Plateau

    I've bee running the new version on a W2K machine for a couple of days now. The dfGUI plot for Energy values has some interesting features. It seems that the reported best energy is the same value for many generations in a row. In fact, it will move off to a different values and then come back to that same value several times. The effect is a display of several "plateau"'s during the course of the system run. Just curious as to what causes that effect.

    Ned

  38. #38
    Originally posted by bwkaz
    Getting the same error here, except without the /i686 part (my glibc was installed to just plain /lib).

    I have glibc 2.2.5, and I'm almost positive that's the problem. This new executable was most definitely not compiled on a RedHat 7.1 box, at least not a standard one...

    ()
    Ooops, actually built on RedHat 8.0...
    Anyways not to fear, the 'real' versions will be compiled on the same machines as always. If you can't use this test version, you'll just have to wait
    Howard Feldman

  39. #39

    Re: Energy Plateau

    Originally posted by Ned
    I've bee running the new version on a W2K machine for a couple of days now. The dfGUI plot for Energy values has some interesting features. It seems that the reported best energy is the same value for many generations in a row. In fact, it will move off to a different values and then come back to that same value several times. The effect is a display of several "plateau"'s during the course of the system run. Just curious as to what causes that effect.

    Ned
    I would appreciate if you could post a pic of this graph. It is probably caused by 'truncation' of the energy function (which ensures it stays between 0 and 100).

    Also, just to clarify, when people mention in this thread 'RMSD' - please indicate where you are getting this number from - I suspect some (many) of you are confusing RMSD with pseudo-energy, and the two are quite different. Right now the only place you can see your RMSD, I believe, is if you log in at the web site and view your personal stats.
    Howard Feldman

  40. #40

    Re: Re: Energy Plateau

    Originally posted by Brian the Fist
    I would appreciate if you could post a pic of this graph. It is probably caused by 'truncation' of the energy function (which ensures it stays between 0 and 100).

Page 1 of 2 12 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •