mdd and sss distributed computing result is worse than one host parallel

[bill wolf]

I run mdd and cs parallel in my cluster. But I found that  the result of parallel is worse than parallel on one computer.The result is in the attachment.

The steps of run parallel in my cluster list below:

1:run zero at user graphic interface

2:modify host and worker in queMM.sh

3:select mdd and Send to remote host

4:delete rproc line and add queMM.sh regions.runs in file project.proc/proc.sh

5:run proc.sh

css's step is the same as mdd step.

Tips:

   mddnmr version :2.5

I don't know where the problem is ,please help me .

[Krzysztof Kazimierczuk]

Hi Bill,

I cannot access the file project.proc/proc.sh
("site can't be reached").
Also, can you say, if you used f180 (half dwell time) in any of the indirect dimensions? We just corrected small bug related to this (in standalone cssolver only) and new version will be soon available.

Kind Regards,
Krzysztof Kazimierczuk

--
--
You received this message because you are subscribed to the Google Groups "mddnmr" group.
To post to this group, send email to mdd...@googlegroups.com rom your registered email address.
To unsubscribe from this group, send email to mddnmr-un...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/mddnmr

---
You received this message because you are subscribed to the Google Groups "mddnmr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mddnmr+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

-- 
Pozdrawiam,
Krzysztof Kazimierczuk<br />


Krzysztof Kazimierczuk, PhD
Laboratory of Intermolecular Interactions
University of Warsaw
Zwirki i Wigury 101, 02-089 Warsaw
tel. +48 22 8220211 ext. 421

[bill wolf]

[Krzysztof Kazimierczuk]

Thanks ,

Can you send proc.sh, fidSP.com, regions.runs and recFT.com ?
Regards,

Krzysztof

--
--
You received this message because you are subscribed to the Google Groups "mddnmr" group.
To post to this group, send email to mdd...@googlegroups.com rom your registered email address.
To unsubscribe from this group, send email to mddnmr-un...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/mddnmr

---
You received this message because you are subscribed to the Google Groups "mddnmr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mddnmr+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[bill wolf]

Hi Krzysztof Kazimierczuk:

    Project.proc,project just a fid,not a special meanings.

    I just modify proc.sh :delete rproc line and add queMM.sh regions.runs.

[bill wolf]

Yes I can.

      I‘ll send all of that you need tomorrow

On Tuesday, December 20, 2016 at 4:24:39 PM UTC+8, Krzysztof Kazimierczuk wrote:

[bill wolf]

[bill wolf]

On Tuesday, December 20, 2016 at 4:50:40 PM UTC+8, Krzysztof Kazimierczuk wrote:

[Krzysztof Kazimierczuk]

Thank you,

Could you send me nls.in file as well and one of the .mdd files from the MDD folder?

Thanks,
Krzysztof

[Krzysztof Kazimierczuk]

Dear mddnmr users,

After days of testing I am not able to confirm the problem reported by Bill. CS works the same in standalone mode and in memory on the datasets and computers I have used... If anybody observes similar effects as Bill, please join this thread and comment.

Kind Regards,
Krzysztof