SPIN polarised spectra in RPA level

[joydev de]

Dear Davide,

Thanks a lot for your quick reply.

I will install it. Let you know.

Sincerely,
Sincerely,
JOYDEV DE, PhD student
NISER,INDIA

[joydev de]

Dear Davide,

According to your suggestion, I have recompiled the yambo source code using this TDDFT_ALDA_G_space.F.
When I run this, It successfully create the SAVE directory but exit with an error message without creating the
LOG file.


Regards,
JOYDEV DE, PhD student
NISER,INDIA

[Daniele Varsano]

Dear Joydev,
there is a bit confusion here.
Here it seems you run p2y and not yambo.
It seems that p2y is launched in parallel but running in serial. Moreover, it seems it has finished even if the log file is messy.
My suggestion is:
1) remove your SAVE directory
2) Run p2y in serial
3) Then run yambo.

Report if you still have a problem.

Best,

Daniele

[joydev de]

Dear Daniele,
In the old version (yambo-4.0.2) , when I change TDDFT_ALDA_G_space.F file inside src/tddft, the
source code is not install properly and always exiting with an error message. In new version (yambo-4.1.2),
it is installed but not working. In new version, first I run p2y in parallel. After that when I am running yambo ,
it always exit with an error message.

According to your suggestion, I run p2y in serial. After that I run yambo in serial for alda level but it is not exit after 24 hours.

The version(yambo-4.0.2) is working for nspin>1 only in the rpa level but it is not working in the alda level.

Sincerely,
JOYDEV DE, PhD student
NISER,INDIA

[Daniele Varsano]

Dear Joydev,
you cannot apply patches generated for a release to another one. You need to apply it to the latest release.
I was suggesting you run p2y in serial and not yambo.

In the new version (yambo-4.1.2), it is installed but not working.

Ok please report what is exactly the problem (input/report/log).

Best,

Daniele

[joydev de]

Dear Daniele,

I use the following ways:-

1. First run p2y in serial.
2. For run set up I run yambo in serial
3. Next I generate the input file using the command: yambo -o c -k alda -V par
But it is not creating the input file for parallel run. It create the input file for serial run.
4.Next I run yambo in serial.
But it takes time.

For the second step when I run yambo for run set up in parallel, always it exit with out creating a LOG directory.
Is it installation problem?

Sincerely,Sincerely,
JOYDEV DE, PhD student
NISER,INDIA

[Daniele Varsano]

Dear Joydev,
1 OK
2 OK
3

Next I generate the input file using the command: yambo -o c -k alda -V par
But it is not creating the input file for parallel run. It create the input file for serial run.

That's strange, anyway you can run it in parallel and default parallelization should apply. You can try by typing -V all
Note your calculation is intense memory consuming:

Code: Select all

<01m-36s> [WF] Performing Wave-Functions I/O from ./SAVE
 <01m-36s> [M 11.432 Gb] Alloc wf_disk ( 0.078)
 <01m-56s> [M 11.354 Gb] Free wf_disk ( 0.078)

A parallilzation over bands here it helps!

For the second step when I run yambo for run[/code] set up in parallel, always it exit with out creating a LOG directory.

Setup is usually a 10 second run, no need to run it in parallel.

Best,

Daniele

[joydev de]

Dear Daniele,

For nspin>1. the yambo-4.1.3 version is running well.
I test my data set.

Is the any different will come for nspin=1 and nspin=2 for same system?

Thanks you for your kind attention.

Sincerely,
JOYDEV DE, PhD student
NISER,INDIA

[Daniele Varsano]

Dear Joydev,

Is the any different will come for nspin=1 and nspin=2 for same system?

This is something that depends on the system, if a spin polarized calculation is needed or not.

Best,

Daniele