linux error

[Grawlfang]

Anyone else getting the following error since the new library has been made available ?

[FIDA] Cryptography Enabled
[FIDA] Initializing
[FIDA] Loading Config File
Loading from: /export/dc/eon/libaloptc.so
[FIDA] Failed - Cannot Load Library: /export/dc/eon/libaloptc.so: cannot open shared object file: No such file or directory [FIDA] Email Address: ???@???.???
[FIDA] Requesting Assignment
[FIDA] Connecting to Server
[FIDA] Updating...
[FIDA] Updating Client Application - Please Wait...
[FIDA] Verifying Signature...
[FIDA] Using public file to verify signature: key.pub
[FIDA] Valid Signature detected (1024 bits)
[FIDA] Created: Thursday November 4, 2004 03:18:17 UTC
[FIDA] Signed: Tuesday May 24, 2005 19:50:03 UTC
Loading from: /export/dc/eon/libemtc.so
Segmentation fault

I'm running Slackware 10.

[Grawlfang]

an strace of the binary gives the following :-

...
...
...
open("/lib/libc.so.6", O_RDONLY) = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\360Y\1"..., 1024) = 1024
fstat64(3, {st_mode=S_IFREG|0755, st_size=1250840, ...}) = 0
old_mmap(NULL, 1237892, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0) = 0x4023d000
mprotect(0x40365000, 25476, PROT_NONE) = 0
old_mmap(0x40365000, 16384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 3, 0x128000) = 0x40365000
old_mmap(0x40369000, 9092, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x40369000
close(3) = 0
open("/lib/ld-linux.so.2", O_RDONLY) = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0 \v\0\000"..., 1024) = 1024
fstat64(3, {st_mode=S_IFREG|0755, st_size=83268, ...}) = 0
old_mmap(NULL, 85900, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0) = 0x400bc000
mprotect(0x400d0000, 3980, PROT_NONE) = 0
old_mmap(0x400d0000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 3, 0x13000) = 0x400d0000
close(3) = 0
--- SIGSEGV (Segmentation fault) @ 0 (0) ---
+++ killed by SIGSEGV +++

Anyone know which distro or version of shared libraries the linux library has been built against (and how it differs from the previous one - which worked fine for me) ?

[rcoulter]

Getting errors on both Linux and Windows. This might be the changeover to the Mg surface or they are making the changes to the server software. They mentioned this a few days ago.

If not, the thingy is broke, guys.

Randy

[Grawlfang]

Originally posted by rcoulter
Getting errors on both Linux and Windows. This might be the changeover to the Mg surface or they are making the changes to the server software. They mentioned this a few days ago.

If not, the thingy is broke, guys.

Randy

Odd, I've fired up a couple of Win98 pc's and they load and run the new dll without any problems, just having trouble with my Linux (and OpenBSD) instances.

[rcoulter]

Windows is working, my error.

Randy

[AMDave]

graeme

does omer have a static version ?

[graeme]

Yeah, emtc is a new library for looking at Mg surface growth. It seems that we messed up with the static linux library -- we'll get it fixed today.

[graeme]

There is a new libemtc.so library on the server for linux. Can someone with a non-working box try to erase the library to force the server to send another and see if things are fixed? I've used the same magic proceedure as before, so I hope we're good.

[AMDave]

that does the trick.
got it and it's running
thanks guys

[Grawlfang]

Hi,

Just tried out the new library and it seems to run at least. Can you confirm that it will not always do the 'image done' and 'good pref' steps as per my output below (the first one does these steps, the subsequent WU does not) ? Or is this a bug in the library code ?


Dimer DoAssignment
Dimer: Start

Atom number [0,192] in the .con will be treated as:
Atomic number: 12.

The atomic number of the atoms being deposited will be: 12.

Image Done: U: 0.316298 FC: 346 TS: 65
Dimer: Done
HSize: 9
Good Pref: 1.98952e+13
Minima, saddle : 0.211131 <- 0.316298 -> -7.92219e-06

[FIDA] Reporting Results
[FIDA] Connecting to Server
Dimer UndoAssignment
[FIDA] Work Unit Completed
[FIDA] Requesting Assignment
[FIDA] Connecting to Server
[FIDA] Processing Assignment

Dimer DoAssignment
Dimer: Start

Atom number [0,192] in the .con will be treated as:
Atomic number: 12.

The atomic number of the atoms being deposited will be: 12.


Dimer: Done

[FIDA] Reporting Results
[FIDA] Connecting to Server
Dimer UndoAssignment
[FIDA] Work Unit Completed

[graeme]

Well, I'm not absolutely sure what the client did, but there are cases in which it can (properly) abort early. The searches done by the clients are somewhat random. Along the way, there are some checks to make sure the physical quantities are reasonable. If the energy gets out of hand, the search is aborted and a new work unit is requested. That's my guess at what happened.

The bad pref and bad mins messages are two other such checks. Perhaps I should be a little more systematic about printing messages to indicate what's going on.

Fortunately there are several younger people cleaning up the code now, so better messages may be forthcoming. I even noticed a few comments in the code yesterday!

BTW, the code is now available through cvs. Information about getting it can be found on our new developers forum at http://theory.cm.utexas.edu/forum .

[Mustard]

Hi Graeme,

Yep, comments are a nice thing! Glad to see the cvs. The latest library/client is interesting. I'm running it fine with no probs so far. Just keep plugging away, as the old saying goes, Rome wasn't built overnight!!

Bruce

[AMDave]

Comments ?! My goodness, graeme.
I thought programmers lived on pizza and coke.
You must be providing beer and chocolate LOL

I have a comment...
I notice that the production rate for the moment is over 9,000 workunits per hour.
That is an impressive rate of growth for the subject structure.

I did note a new messages I had not seen before:
"the bytesize of the data received was not the size expected" (or similar)
and then the clients went into sleeps
this only happened for about half an hour and is not happening now.
just mentioning it as an observation.

[jasong]

I'm getting the libal110c.so error right now, which is especially unfortunate since p-1 runs truly horribly on my quad-core when more than two instances are running, and I'm not even talking about Stage 2.

Think of driving a car with your foot on the gas pedal and the brake at the same time. You think you've had latency issues? Try going from 3hrs45min each with an instance on cores 0 and 2, and then going to 4 instances with 30 hours(yes 30) per core. Horrible, truly horrible.

Edit: Could someone send an email to get them to look at this thread? I haven't had my email set up on a new machine, and I can't seem to type a message in the web form of my ISP without Mozilla complaining about "scripts."

[AMDave]

Edit: Could someone send an email to get them to look at this thread? I haven't had my email set up on a new machine, and I can't seem to type a message in the web form of my ISP without Mozilla complaining about "scripts."

Apparently not.

Myself and another project member have both sent multiple emails to graeme and eric (for other reasons) and there has been no confirmation that the emails have been received or read.

It's a ghost train and the driver fell on the dead-man's handle.

Watch out for the next curve - it might get a bit scary.

[jasong]

As far as bookmarks and email are concerned, I've never had anything incredibly important on my machines. I've lost all my data about 3-4 times, only once due to hardware problems, and the only truly bad thing about it was when I thought my dad was going to physically assault me for continually yelling at my computer.

Yeah, except for a Bible and a good pair of glasses, there really isn't anything I own that I couldn't bear to lose. Now, of course, if a hurricane or an earthquake got rid of EVERYTHING, that would be really bad. But if a few things, including expensive things like this computer, go belly up all of a sudden, it's not a super big deal to go without for a while.

[AMDave]

As far as bookmarks and email are concerned, I've never had anything incredibly important on my machines. I've lost all my data about 3-4 times, only once due to hardware problems, and the only truly bad thing about it was when I thought my dad was going to physically assault me for continually yelling at my computer.

Yeah, except for a Bible and a good pair of glasses, there really isn't anything I own that I couldn't bear to lose. Now, of course, if a hurricane or an earthquake got rid of EVERYTHING, that would be really bad. But if a few things, including expensive things like this computer, go belly up all of a sudden, it's not a super big deal to go without for a while.

wrong thread, perhaps?

[jasong]

wrong thread, perhaps?

Guess I should have put off-topic. I'd had a stressful day at work and was probably free associating a little more than usual.

[jasong]

Okay, just as a heads-up to the people who run the project(as opposed to people who CRUNCH the project):

However much time has passed between when the problem was fixed and I noticed it was fixed: Well, that's 4 instances of Core2Quad instances that you could have gotten from me. If Carlos hadn't sent me a PM, I still wouldn't've known the problem was fixed.