Check pointing?

[zombie67]

Do the RNA apps checkpoint? And if so, how frequently?

[Ananas]

The windows version does not.

If software paket "freeze" (it's not the Phyton "freeze" program I guess) is installed, the Linux version saves a process image for restart on the checkpoints, so that's not really checkpoints but it serves the same purpose.

[Roland Schneider]

No, there is currently no checkpointing for the RNA apps.

The only system where checkpointing can be enabled manually is linux 32 bit, but it not that easy and you have to have superuser access to do it.

[yoyo]

The windows version does not.

If software paket "freeze" (it's not the Phyton "freeze" program I guess) is installed, the Linux version saves a process image for restart on the checkpoints, so that's not really checkpoints but it serves the same purpose.

Freeze is inside the RNA application and is called frequently. If it works depends on the configuration of the Linux kernel. You can see the freeze calls in stderr of the results.
yoyo

[Ananas]

...
Freeze is inside the RNA application and is called frequently. ...

why does it sometimes say "not found" then?

[yoyo]

...
Freeze is inside the RNA application and is called frequently. ...

why does it sometimes say "not found" then?

This should be the case only on linux64 and Windows. For these we do not have a freeze app.
yoyo

[Michael H.W. Weber]

In my case, validation errors occur exclusively on my single 32-Bit Linux box and there is a lot of notes from the "freeze" program in my (non-validated) results. So, I conclude this "freeze thing" has to do with the validateion issues. However, not all WUs on this machine are faulty (it is a minority actually). I do not know how the error occasionally occurs. It would surely be very helpful if we could nail the cause for this.

Michael.

[yoyo]

I don't think, that freeze is the root cause for not validating. The freeze output only show that freeze works and what it does. But you can probably switch off freeze. Just delete freeze in the slot directory.
yoyo

[Ananas]

...
why does it sometimes say "not found" then?

This should be the case only on linux64 and Windows. For these we do not have a freeze app.
yoyo

http://www.rnaworld.de/rnaworld/result. ... tid=120325

Wrapper: writing checkpoint image
wrapper: starting ./freeze chkpoint 4367
sh: ./freeze: not found

Linux 2.6.24-19-generic (nothing about x64, it is Saenger's box)


A possible reason : It doesn't sit in the right directory when it calls freeze, i.e. cwd isn't where freeze is extracted.

[yoyo]

As I know saengers PC is a 64bit Linux, but named themself linux-generix. So it downloaded the 64bit app.
yoyo

[Uwe Sänger Herzke]

The first lines in BOINC after start-up on my machine are this:

Code: Alles auswählen

Sa 16 Jan 2010 13:51:49 CET		Starting BOINC client version 6.10.17 for x86_64-pc-linux-gnu
Sa 16 Jan 2010 13:51:49 CET		Config: use at most 4 CPUs
Sa 16 Jan 2010 13:51:49 CET		log flags: file_xfer, sched_ops, task
Sa 16 Jan 2010 13:51:49 CET		Libraries: libcurl/7.18.0 OpenSSL/0.9.8g zlib/1.2.3.3 c-ares/1.5.1
Sa 16 Jan 2010 13:51:49 CET		Data directory: /home/saenger/BOINC/BOINC
Sa 16 Jan 2010 13:51:49 CET		Processor: 4 GenuineIntel Intel(R) Core(TM)2 Quad  CPU   Q9450  @ 2.66GHz [Family 6 Model 23 Stepping 7]
Sa 16 Jan 2010 13:51:49 CET		Processor: 6.00 MB cache
Sa 16 Jan 2010 13:51:49 CET		Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good pni monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr sse4_1
Sa 16 Jan 2010 13:51:49 CET		OS: Linux: 2.6.24-19-generic
Sa 16 Jan 2010 13:51:49 CET		Memory: 3.87 GB physical, 5.68 GB virtual
Sa 16 Jan 2010 13:51:49 CET		Disk: 132.04 GB total, 67.03 GB free
Sa 16 Jan 2010 13:51:49 CET		Local time is UTC +1 hours
Sa 16 Jan 2010 13:51:49 CET		No usable GPUs found
Sa 16 Jan 2010 13:51:50 CET		Not using a proxy

I don't know where to change the naming of the OS, but it's 64bit.
Edith says: 2.66GHz is wrong as well, it's 3.2GHz

[vaughan]

Check pointing would be nice.

I have a couple of tasks running that are going to run over the deadline.
The task names are incredibly long and I don't know how to cut & paste but the first is:
CMS_MW1_Anaeromxyobacter-sp.-Fw109-5 etc etc
The second is:
CMS_MW1_Clostridium-E88 etc etc
They are at approx. 4% done after 40 hours with estimated 78 hours remaining.
This is on an Intel C2D E8600 cpu @ stock 3.33GHz with 8GB RAM, Windows 7 64-bit Ultimate.

Perhaps its time to abort them.