Beta test of new algorithm now available - see inside

[Brian the Fist]

We are ready to begin beta-testing of the new sampling algorithm. If you wish to help beta-test, it is VERY important that you read this message carefully and through to the end.

For testing purposes, binaries have been built for Intel Linux and Windows only. We are not testing the screensaver yet. You may test it as a Windows service now if you wish but would prefer you test it on the command line first.

Things to look for:

Try different switch combinations, and try things while you are and are not conencted to the internet.

You can now quit at ANY time. Verify this works for you, and that it continues where it left off. Try quitting with 'Q', by deleting the lock file, or any other proper way.

Uploading now occurs only once every generation. Watch stats on the web site to see if they change as expected.

Please do NOT make suggestion as to improvements, etc. yet, we first want to verify everything is working now as is. We will entertain suggestions and improvements at a later time.

When reporting problems, be sure to specify your exact OS, whether running as a service or normal ASCII mode, and list ALL the switches in foldit(.bat) or dfGUI that you have set when you observe the problem. Post all errors/problems TO THIS THREAD. Please do not e-mail us. This ensures each bug will only get reported once (hopefully).

The protein will get 'stuck' a lot more now after the first generation. After about 3 minutes it will restart the structure if it is still stuck. This behaviour is normal.

For the beta test the first generation has 500 structures and all others have 20 structures (after 50 generations, it should restart at gen. 0 again). When the beta is through though, the 'real' program will have 5000 for gen. 0 and 200 for all others (10 times as much that is).

The server for beta purposes can be accessed at http://beta.distributedfolding.org/ This web site is a copy of the main website and contains stats for the beta test only. It was copied over at some arbitrary time in the past so most registered users shoudl be there but if you signed up recently, you may need to sign up again. This is a physically different machine and anything you do/see here will have zero effect on www.distributedfolding.org. Likewise, all 'work done' will be credited to the beta server and will NOT be carried back to the real server ever. You are beta testing to help us help you, to ensure the client works well, not to get extra stats. Please do not beta-test if you are only interested in getting stats for your testing.
Work is also uploaded to beta.distributedfolding.org on port 80, so anyone needing to poke holes in firewalls or what not, that's all you should need.

Also note that the beta test will not be affected by protein changes/program updates in the 'real' program.

I think that's all I had to say for now. Now the important part - download and installation instructions:

Step Zero - beta software is available ONLY for Windows and Linux-i386

Step one - download a fresh copy of the client from the normal web site for one of the two OSes named above, and install to a clean directory

Step two - get one of the following:

ftp://ftp.mshri.on.ca/pub/distribfol...ux-i386.tar.gz

or

ftp://ftp.mshri.on.ca/pub/distribfol...beta-win9x.zip

Inside are four files: readme.txt, native.val, protein.trj and foldtrajlite(.exe)

overwrite the files of the same names in your clean install from step one with the files in this archive.

Step three - run foldit or foldit.bat as normal. Your handle, login, password, etc. are all the same as on the 'real' distributed folding site.

Thanks for helping to beta test this software and we are excited about getting this going!

[FoBoT]

w00t!

i am in

i will run one linux (RH 7.3) and one winders (Win98) on two identical pc's
Compaq Evo D300v Celeron 1.1 Ghz 128 MB RAM

[matrix_fan]

Count me in too..Windows 98 AmD k6-2

[MarcyDarcy]

Don't be afraid this is gonna be a long post
I've written down everything i've done and the client shows while running. I've made screenshots of all the situations

-------------------
AMD XP 1800+ (not overclocked)
512 mb RAM
Windows XP Professional (Dutch Language)


1)
- installed
- started foldit.bat
- running smooth
- Changed the windows size to very small (by using the "mouse-drag option") and then changed it back to the normal size. afterwards some ASCII art residu stays behind
- pressed Q
- client closed very quick. Faster than normal i would say
- error.log looks normal
- filelist.txt contains some information. It looks diffentent than normal

fold_0_[handletext]_0_protein.log.bz2
[handletext]4_0_protein_0000210.val
CurrentStruc 0 241 123 0 0 210 14.482

2)
- eddited foldit.bat with the -df function
- started foldit.bat
- looks ok
- pressed Q
- started foldit.bat
- client continues at the point it stopped

3)
- eddited foldit.bat with the -rt function
- started foldit.bat
- client runs much faster as it should be doing
- letting it run to complete 500 structures
- client shows message: "minimizing energy of best structure found". this takes a little while, around 1 minute
- client shows message: "making Gen. 1 trajectory distribution". this takes a little while, around 20 seconds
- client shows again: "minimizing energy of best structure found". this takes a little while, around 1 minute
- client shows message: "making Gen. 2 trajectory distribution". this takes a little while, around 20 seconds
- client continues with the structures
- only 20 structures and no generation buffered.
- client begins again with displaying the messages with making Gen. 3 and Gen. 4 trajectory distributions
- pressed Q
- error.log shows 4 times a start of the client. I only closed the client 3 times.
- filelist.txt looks different then before

.\fold_0_[handletext]_0_[handletext]_protein_2.log.bz2
.\[handletext]_0_[handletext]_protein_2_0000013.val
.\fold_0_[handletext]_0_[handletext]_protein_3.log.bz2
.\[handletext]_0_[handletext]_protein_3_0000011.val
CurrentStruc 0 14 123 3 1 11 11.776

4)
- started foldit.bat
- let it calculate the 20 residues. While showing message "minimizing energy..." pressed on Q and waited. Client closed after 3 seconds.
- started foldit.bat again. Client resumed with the message 'minimizing energy..."
- deleted foldit.lock
- pressed Q
- restarted foldit.bat. Client again begins with the fist part of 'minimizing energy..." Client doesn't remember at what point in this minimizing it stopped.

5)
- eddited foldit.bat with the -i f function
- started foldit.bat
- client resumes with displaying the messages.
- next the client begins to fold. I noticed that the structures are larger (in ASCII art) then before.
- at structure 15 it hangs and filled the complete screen. It runs like this for a second of 10.
- this screenfilling contunues with the structres 16 till 20.
- client shows the messages
- client has 3 generations buffered
- client now begins to 'hick' at structure 11. Calculating residue switches between 47 and some other all the time. By structure 16 this is residue 69 and ...
- pressed Q when client agains shows the messages.

Thought the client didn't upload but looking at the stats it did anyway.

6)
- eddited foldit.bat and removed the -i f and -df function
- client shows uploading filesets as normal while uploading
- pressed Q

7)
- placed a small tracking program in the directory (made by one of my teammembers almost 1/2 year ago).
- started foldit.bat
- client shows the messages for a minute or 2
- client started and hicked on structure 1
- the small tracking program doesn't work anymore with this beta client
- client hicks on structure 2 and filled the complete screen with the ASCII art for a while
- closed small tracking program
- client hicks on structure 3 etc. etc. etc. It calculates the residue for a second of 10 to 25
- pressed Q

In the client directory there are the following files:
[handletext]_0_[handletext]_protein_5_0000012_min.val.bz2
[handletext]_0_[handletext]_protein_6_0000004_min.val
[handletext]_0_[handletext]_protein_6_0000004.val
[handletext]_0_[handletext]_protein_7_0000002.val
fold_0_[handletext]_0_[handletext]_protein_6.log.bz2
fold_0_[handletext]_0_[handletext]_protein_7.log.bz2
error.log and filelist.txt are looking ok

And at this point i have to stop testing. But i'll be back

[Pascal]

Ok, I installed the actual client into a new folder under SuSE Linux 8.1 (KDE 3.1) - the the betaclient.

1. with internet access and quiet mode, -rt -df

-> the client crashed three or four times. Why? I don't know - no error message.

2. with internet access, no quiet mode, -rt -df

-> the client works well, but seems to produce rather huge structures (perhaps 1 structure per minute)

System:

Athlon TB-C, 1.2 GC/s, 512 MB DDR-SDRAM

[dano]

AMD XP1800+
256 mB RAM
Mandrake Linux 9.0

I followed your instructions then just typed ./foldit in a virtual console.

[dan@mdk9 BataDF]$ ./foldit
./foldtrajlite: error while loading shared libraries: libncurses.so.4: cannot open shared object file: No such file or directory
[dan@mdk9 BataDF]$

[Pascal]

Originally posted by dano
..

[dan@mdk9 BataDF]$ ./foldit
./foldtrajlite: error while loading shared libraries: libncurses.so.4: cannot open shared object file: No such file or directory
[dan@mdk9 BataDF]$ [/B]

You should have a look for the ncurses and the libs - and install them.

[Mpemba Effect]

I'm running the with -qf -dt -it -rt switches using the Linux client.

- The first thing I notice is that it's generating some huge asciiness for structures after generation 0. I normally run my clients with quiet mode so haven't noticed before, but these are very common now. It takes a long time building those structures

 ------------------------------------------------------------
Distributed Folding Linux Benchmark Script V1.0

Sample Size: 15 structures over 1003 seconds.

Structures Per Second:                  0.01
Structures Per Minute:                  0.90
Structures Per Hour:                    53.84
Structures Per Day:                     1292.16

Linux OS - Running Kernel Version 2.4.20-ck3
Intel(R) Pentium(R) 4 CPU 1.80GHz @ 2438mhz (512 KB cache)
------------------------------------------------------------

Is this normal?

- Uploading seem to fine so far.

- Resume also seem to be working when I stop the client via "Q", deleting the lock file, and even "killing" it (although the logs do show a "ERROR: [001.001] {foldtrajlite2.c, line 1262} Caught sig 15")

- Stopping the client halfway though "Minimizing energy of best structure found" causes it to start "minimizing" from the beginning. Same goes for if it's stopped halfway though "Making Gen. X Trajectory Distribution". Although the second time I tried stopping it at the "Minimizing energy of best structure found" stage, upon resume it started refolding structure 20 and then crashed. The error log give this message

========================[ Feb 19, 2003  9:29 PM ]========================
ERROR: [001.015] {randwalk.c, line 3745} Avoiding endless loop

========================[ Feb 19, 2003  9:31 PM ]========================
FATAL ERROR: [000.000] {foldtrajlite2.c, line 2573} Unable to find file ammf8ug
9_0_ammf8ug9_protein_3_0000018.val; cannot continue - replace file and start ag
ain, or manually delete filelist.txt

thats all for now

[Pascal]

The benchmarking script for Linux does not work with more generations. The number of the structures is always reset.

To the other problem - delete the filelist.txt and restart the client.

[Brian the Fist]

Just so you know, I said DO expect the generations 1 and up to take much longer to make, and look big on ASCII art mode. Alos the filelist.txt format and file naming is significantly different as you may notice.

If you quit during minimization or generating trajectory distirbution, when you restart, it will restart at the beginning of that step and not resume EXACTLY where it left off. This is intentional. It is too complicated to resume these tasks partway through, and since they typically take a minute or less compared with hours between generations (remember they'll be 10 times bigger eventually) then it should not be a big deal ion terms of 'lost work' to redo this little bit.

For that curses error, I should point out this is the linux-gcc version, not icc version. Redhat 7.x (but not 8.x) and equivalents is missing the curses-4 library. If you have Redhat 7.x (or are missing it) make a symbolic link from curses-5 to curses-4 and run 'ldconfig -v' as root. If you don't know what I'm talking about, perhaps someone else will go into more detail.

As for the 'unable to find file' error, that is a very important one, which I thought I had fixed. If others get this error at some point, please try to establish the exact conditions required in order to reproduce it. You will likely have to do as it says and delete filelist.txt to get it going again as well.

Thanks for the quick responses, I look forward (sort of) to more bug reports over the next few days.

[dano]

Originally posted by dano
AMD XP1800+
256 mB RAM
Mandrake Linux 9.0

I followed your instructions then just typed ./foldit in a virtual console.

[dan@mdk9 BataDF]$ ./foldit
./foldtrajlite: error while loading shared libraries: libncurses.so.4: cannot open shared object file: No such file or directory
[dan@mdk9 BataDF]$

Mandrake 9.0 and debian stable both use libncurses.so.5.2 , so you need create a symbolic link in /lib for debian or /usr/lib for mandrake

ln -s libncurses.so.5.2 libncurses.so.4

to get it to run

[Hydrox]

Thunderbird 900MHz
256MB RAM
Windows XP SP1
-rt -qt -g 1

ERROR: [001.015] {randwalk.c, line 3745} Avoiding endless loop
ERROR: [001.015] {randwalk.c, line 3745} Avoiding endless loop
FATAL ERROR: [012.002] {trajtools.c, line 2377} Attempt to insert duplicate residue number into database

[Bionic_Redneck]

to get the client running without libncurses.so.4 in mandrake 9 just make a symbolic link from libncurses.so.5.2 and rename it libncurses.so.4

[Brian the Fist]

I have verified that it will crash under Linux with the '-qt' option. So please run it with -qf for now. Once I've fixed a number of problems, I'll post updated beta executables. Thanks. I will removed that 'avoiding endless loop' warning as well - it is unimportant and will just bloat the log file again.

[Brian the Roman]

Woohooo!!!! I'm in! I've been holding my breath waiting for this. I hope Howard appreciates how difficult it was to keep from asking 'Are we there yet?' every day.

Running on one of my machines for now:
AMD XP1600, 512MB, Win XP.

Installed and ran without a hitch so far...

ms

[dano]

Windows XP home
AMD XP2600+
500mB RAM

-qf -df -it -rt

I am running the ASCII version from the command line.
I started the client connected to the internet then disconected from the internet and stoped the client.
Then reconnected started the client and it said the "Data file checksum failed".
It will not upload anything after that.

Error log

========================[ Feb 19, 2003 7:45 PM ]========================

========================[ Feb 19, 2003 7:45 PM ]========================
ERROR: [777.000] {ncbi_socket.c, line 864} [SOCK::s_Connect] Failed connect() to beta.distributedfolding.org:80 {errno=10061}
ERROR: [777.000] {ncbi_connutil.c, line 526} [URL_Connect] Socket connect to beta.distributedfolding.org:80 failed: Unknown
ERROR: [777.000] {ncbi_http_connector.c, line 117} [HTTP] Retry attempt(s) exhausted, giving up
ERROR: [777.000] {ncbi_socket.c, line 864} [SOCK::s_Connect] Failed connect() to beta.distributedfolding.org:80 {errno=10061}
ERROR: [777.000] {ncbi_connutil.c, line 526} [URL_Connect] Socket connect to beta.distributedfolding.org:80 failed: Unknown
ERROR: [777.000] {ncbi_http_connector.c, line 117} [HTTP] Retry attempt(s) exhausted, giving up
ERROR: [000.000] {foldtrajlite2.c, line 3616} Error during upload: NO RESPONSE FROM SERVER - WILL TRY AGAIN LATER
ERROR: [777.000] {ncbi_socket.c, line 864} [SOCK::s_Connect] Failed connect() to beta.distributedfolding.org:80 {errno=10061}
ERROR: [777.000] {ncbi_connutil.c, line 526} [URL_Connect] Socket connect to beta.distributedfolding.org:80 failed: Unknown
ERROR: [777.000] {ncbi_http_connector.c, line 117} [HTTP] Retry attempt(s) exhausted, giving up
ERROR: [777.000] {ncbi_socket.c, line 827} [SOCK::s_Connect] Failed SOCK_gethostbyname(bioinfo.mshri.on.ca)
ERROR: [777.000] {ncbi_connutil.c, line 526} [URL_Connect] Socket connect to bioinfo.mshri.on.ca:80 failed: Unknown
ERROR: [777.000] {ncbi_http_connector.c, line 117} [HTTP] Retry attempt(s) exhausted, giving up
ERROR: [000.000] {foldtrajlite2.c, line 1345} Wed Feb 19 19:52:35 2003
Unable to check server status
ERROR: [000.000] {foldtrajlite2.c, line 3502} File XXXXXXXX_0_protein_0000270_min.val is corrupt or has been tampered with; cannot continue - replace file and start again, or manually delete filelist.txt
ERROR: [000.000] {foldtrajlite2.c, line 3616} Error during upload: Data file checksum failed
ERROR: [000.000] {foldtrajlite2.c, line 3502} File XXXXXXXX_0_protein_0000270_min.val is corrupt or has been tampered with; cannot continue - replace file and start again, or manually delete filelist.txt
ERROR: [000.000] {foldtrajlite2.c, line 3616} Error during upload: Data file checksum failed

I replaced my handle with "XXXXXXXX"

progress.txt

Building structure 19
1 until next generation
4 generations buffered

[FoBoT]

can i get a clarification?
to run the beta linux client

you first download the gcc version (normal version) , not the intel compiler version, correct?

thank you

[dano]

Originally posted by FoBoT
can i get a clarification?
to run the beta linux client

you first download the gcc version (normal version) , not the intel compiler version, correct?

thank you

The Intel version worked fine for me.

[FoBoT]

hmm, maybe it doesn't matter

the instructions above to do the symbolic link didn't work so i am installing RH8 instead

don't want to muddy the thread with my low linux skill problems

my windows 98 box is running strong, no issues at all so far

[FBK]

W2K PRO all current service pack(s) and patches
Tyan tiger with AMD XP1800+ procs
500mB RAM

proc 1 = -df -rt -g100 -p -20
proc 2 = -df -rt -g100

I am running 2 instances the ASCII version from the command line, on this machine.

Both instances ran and uploaded for a several hours, but both
eventually generated the same errors.

Example of both error logs

ERROR: [000.000] {foldtrajlite2.c, line 3502} File .\xxxxxxxx_0_xxxxxxxx_protein_2_0000001_min.val is corrupt or has been tampered with; cannot continue - replace file and start again, or manually delete filelist.txt
ERROR: [000.000] {foldtrajlite2.c, line 3616} Error during upload: Data file checksum failed

[frozenchosen]

Downloaded beta version and its running as advertised on Dell 2.4 ghz 1gb RAM Windows XP.

[Pascal]

That morning I had to delete the filelist.txt after an hour - the client (Linux, qf -df -it -rt) repeated endless producing the same structure. Ending and restarting did not help, so I had to delete the filelist.txt


OS SuSE 8.1 (KDE 3.1)

For hardware look there
||
V

[pfb]

Ran the Windows beta for 6 hours overnight on my main PC (XP Home on a P4 @ 2.8, 512MB) with the extra ram switch- no errors reported in error.log - used with and without net access.

Only query is that the 'minimising energy' progress bar sometimes didn't reach the end - and the following generation normally produced the biggest ASCII 'drawings'...

[mighty]

Athlon XP 2100+
Windows 2k
Switches: -df -rt -i t
I left it running in offline mode overnight and stopped it to upload 18 generations this morning. The upload itself (removed -i t and added -u t) went fine, but right after the "UPLOAD COMPLETED - CONTINUING" message the client crashes. This has happened to me before when I used the same procedure. I will try to replicate the error later to see it if crashes everytime I try to upload.

[mighty]

Athlon XP 2100+
Windows 2k
Switches: -df -rt

Client got stuck on the same structure in generation 18 for more than 5 hours. It has been restarting on the same structure every 3 minutes or so, but it just won't go any further. I have tried stopping and starting the client several times and also tried letting it run for 2-3 hours without interuption - same result in the end. I guess technically this is not a bug, but there should be some kind of timelimit so this doesn't happen. Maybe continue to next structure after x*3 minutes?

[Brian the Roman]

I think we can see the benefits of this new algorythm at work. My values have gone from 10.28, to 9.45, to 7.25 to 7.21. Under the old method I never would have seen this type of improvement in less than a day. Howard et al's work seems to be paying off!

ms

[Digital Parasite]

Originally posted by mighty
Client got stuck on the same structure in generation 18 for more than 5 hours. It has been restarting on the same structure every 3 minutes or so, but it just won't go any further. I have tried stopping and starting the client several times and also tried letting it run for 2-3 hours without interuption - same result in the end. I guess technically this is not a bug, but there should be some kind of timelimit so this doesn't happen. Maybe continue to next structure after x*3 minutes?

P3-800
Windows 2XP Pro
Switches: none

Same thing happens to me. My client got stuck on the last protein in generaton 15, it gets up to around 70AA and gets stucks, resetting every 3 minutes but just can't get past it.

What happens to a person if they get stuck like this say for hours, when they upload their results I guess they will just get credit for those structures in that generation even though it may have taken them 20x longer to complete than other people because it got more stuck than others?

Jeff.

[Brian the Fist]

I have added in a feature for the next beta release (maybe early next week) so that if it continues to get 'stuck' several times in a row, it will be more 'lax' on the restraints to allow it to complete. It will keep increasing the laxity (resulting in a poorer quality structure in term of atomic overlaps, etc.) but at leats it will finish. Once that generation finishes, the laxness will reset to the default in the hopes that it will get out of whatever 'tight spot' it was stuck in.

Also, it is normal that the minimization can complete early some of the time (if the structure is already close to a local minimum).

thanks for all the input so far, we are pleased with the improved RMSDs we have achieved so far (in case there was any doubt, this is the same protein we are doing on the 'live' experiment right now).

For the Linux version you can first download either linux release, as you are going to overwrite the executable with the beta anyhow.

Looks like there are a few possible bugs here to iron out but not too bad, looks like I got most of the big ones already. I'll start trying to reproduce the ones you've described here and please try to describe how to reproduce them as precisely as possible if you can, to make it faster for me to fix. Thanks!

[Brian the Fist]

Originally posted by mighty
Athlon XP 2100+
Windows 2k
Switches: -df -rt -i t
I left it running in offline mode overnight and stopped it to upload 18 generations this morning. The upload itself (removed -i t and added -u t) went fine, but right after the "UPLOAD COMPLETED - CONTINUING" message the client crashes. This has happened to me before when I used the same procedure. I will try to replicate the error later to see it if crashes everytime I try to upload.

I am able to reproduce this one and will get it fixed. Thanks for spotting it.

[Digital Parasite]

I just want to clarify that there is no detriment to running this new algorithm "offline" right? The client doesn't get information from the server between generations? It just uses the information from the previous generation that was locally created?

Jef.

[m0ti]

I haven't gotten in on the beta yet, but I will be soon.

I've gotten an early version of dfQ together which seems to be working but I've got a few bugs/features to iron out. Once that's out of the way, I'll release it for beta-testing. Of course I'll need to make changes to accomadate the beta so I might do a beta compliant version too.

Once things settle down on that front I'll be in with the rest of ya crunching away!

Originally posted by Brian the Fist
thanks for all the input so far, we are pleased with the improved RMSDs we have achieved so far (in case there was any doubt, this is the same protein we are doing on the 'live' experiment right now).

Wow! That's truly incredible. Just shows the value of adding a little bit of intelligence to the algorithm. You guys have done 1.7M folds and have already found a better fold than with the standard algorithm in over 9 billion folds! That's about a factor of 4500! :shocked:

[shortfinal]

Originally posted by mighty
Athlon XP 2100+
Windows 2k
Switches: -df -rt -i t
I left it running in offline mode overnight and stopped it to upload 18 generations this morning. The upload itself (removed -i t and added -u t) went fine, but right after the "UPLOAD COMPLETED - CONTINUING" message the client crashes. This has happened to me before when I used the same procedure. I will try to replicate the error later to see it if crashes everytime I try to upload.

Same thing happened to me.

Windows NT V4.0 Sp6
Switches: -rt -g 5 -df -i f
I left it running for only 5-6 generations. Stopped it while it was generating a structure by pressing the 'Q' key. I then used options -df -u t which uploaded fine. Right after the UPLOAD... message the client aborted via Dr Watson. The dialog box said:

Exception: access violation (0xc0000005), Address: 0x00475e5e

[shortfinal]

Originally posted by Brian the Fist
I am able to reproduce this one and will get it fixed. Thanks for spotting it.

Sorry Howard, I didn't see your post before I submitted mine. Never mind.

[Brian the Fist]

Originally posted by Digital Parasite
I just want to clarify that there is no detriment to running this new algorithm "offline" right? The client doesn't get information from the server between generations? It just uses the information from the previous generation that was locally created?

Jef.

Correct. For the time being anyways...

[Welnic]

Running on a linux box with these settings:
-qf -df -rt -g 100 -it

I was moving some stuff around and making some links (thanks dano for the libncurses fix) so I was logged in as root. I fire up the beta and ran it through the first generation and then quit.

I tried to restart as my normal login and got a message that the program crashed, nothing in the log file. I can still run as root.

[Brian the Fist]

Originally posted by FBK
W2K PRO all current service pack(s) and patches
Tyan tiger with AMD XP1800+ procs
500mB RAM

proc 1 = -df -rt -g100 -p -20
proc 2 = -df -rt -g100

I am running 2 instances the ASCII version from the command line, on this machine.

Both instances ran and uploaded for a several hours, but both
eventually generated the same errors.

Example of both error logs

ERROR: [000.000] {foldtrajlite2.c, line 3502} File .\xxxxxxxx_0_xxxxxxxx_protein_2_0000001_min.val is corrupt or has been tampered with; cannot continue - replace file and start again, or manually delete filelist.txt
ERROR: [000.000] {foldtrajlite2.c, line 3616} Error during upload: Data file checksum failed

I have not seen this error occur. Is the named file actually present still, or have you moved on already? Can you specify how to reproduce this error and if you still have the file, please e-mail it to trades@mshri.on.ca and I can take a look. Did both instances have exactly the same error or were the numbers in the filename different? thanks.

[Brian the Fist]

Originally posted by Welnic
Running on a linux box with these settings:
-qf -df -rt -g 100 -it

I was moving some stuff around and making some links (thanks dano for the libncurses fix) so I was logged in as root. I fire up the beta and ran it through the first generation and then quit.

I tried to restart as my normal login and got a message that the program crashed, nothing in the log file. I can still run as root.

That is likely a file permission problem then. Try chown ing the files to your uer account before running it.

[Insidious]

running on one XP machine...

not service, looks fine so far (just started) seeing the "stuck" thingy frequently. No 'problems' yet.

[Welnic]

Originally posted by Brian the Fist
That is likely a file permission problem then. Try chown ing the files to your uer account before running it.

Yes, that is the problem, it actually happens with current software. If I run the software as root any data files that are saved as well as filelist.txt are all owned by root. If I try to run under my normal account the program can't do anything with those files and crashes.

On my linux boxes the folding is started by startup scripts so it does actually leave root owned files around, but I guess that it has always uploaded them when I have had to stop it and start it again as a lesser being.

[Insidious]

I see memory usage switch between the 'normal' 88MB and a much lower ~20MB or so from time to time... normal?

WindowsXP
AXP 2000+ @ 1765
Service install

service.cfg:

service=1
useram=1
progress=15
uploadfreq=999


Switches:

.\foldtrajlite -f protein -n native -qf -df -it -rt

(no shutdowns or strangness so far.....)

Should dfGUI work with this client?

-Sid