restart problem


Just Got Here
Hi all,

I am doing calculation with DFT/COSMO model,

It seems, when I restart the job, it will always restart from the beginning of cosmo part. Does anyone have this problem before? What should I do? There is no variable in the COSMO list can load any information...

And I have this restart problem in TDDFT calculation too, it will always restart from scratch.

Thank you!

Forum Vet
Can you be a little more specific. What is the behavior that you are trying to accomplish:

1. for the DFT/COSMO run?

2. For the TDDFT?

For the latter, I don't think we have restartable eigensolvers yet.

Bert



Quote:Rayamaranth Dec 2nd 2:07 pm
Hi all,

I am doing calculation with DFT/COSMO model,

It seems, when I restart the job, it will always restart from the beginning of cosmo part. Does anyone have this problem before? What should I do? There is no variable in the COSMO list can load any information...

And I have this restart problem in TDDFT calculation too, it will always restart from scratch.

Thank you!

Just Got Here
Thank you for replying!

I think I have problem for both of them, and they have similar behavior,

No matter which calculation I am doing(DFT/COSMO or DFT/TDDFT), it will restart from the last step of dft calculation, it will get converged in 2 steps, and then go to a new COSMO(TDDFT) calculation. So there is no effective restart.

I understand what you said at last, nwchem-6.0 doesn't have mechanism to restart the TDDFT. BUT, I've seen one sentence in the instruction of nwchem-6.1,("proper restart of TDDFT"), does it mean this mechanism will be included in this new version?

Thank you again!



Quote:Bert Dec 3rd 12:03 am
Can you be a little more specific. What is the behavior that you are trying to accomplish:

1. for the DFT/COSMO run?

2. For the TDDFT?

For the latter, I don't think we have restartable eigensolvers yet.

Bert



Quote:Rayamaranth Dec 2nd 2:07 pm
Hi all,

I am doing calculation with DFT/COSMO model,

It seems, when I restart the job, it will always restart from the beginning of cosmo part. Does anyone have this problem before? What should I do? There is no variable in the COSMO list can load any information...

And I have this restart problem in TDDFT calculation too, it will always restart from scratch.

Thank you!

Forum Vet
Yes, there should be bette r restart capabilities in NWChem 6.1 for TDDFT.

Bert


Quote:Rayamaranth Dec 3rd 4:40 am
Thank you for replying!

I think I have problem for both of them, and they have similar behavior,

No matter which calculation I am doing(DFT/COSMO or DFT/TDDFT), it will restart from the last step of dft calculation, it will get converged in 2 steps, and then go to a new COSMO(TDDFT) calculation. So there is no effective restart.

I understand what you said at last, nwchem-6.0 doesn't have mechanism to restart the TDDFT. BUT, I've seen one sentence in the instruction of nwchem-6.1,("proper restart of TDDFT"), does it mean this mechanism will be included in this new version?

Thank you again!



Quote:Bert Dec 3rd 12:03 am
Can you be a little more specific. What is the behavior that you are trying to accomplish:

1. for the DFT/COSMO run?

2. For the TDDFT?

For the latter, I don't think we have restartable eigensolvers yet.

Bert



Quote:Rayamaranth Dec 2nd 2:07 pm
Hi all,

I am doing calculation with DFT/COSMO model,

It seems, when I restart the job, it will always restart from the beginning of cosmo part. Does anyone have this problem before? What should I do? There is no variable in the COSMO list can load any information...

And I have this restart problem in TDDFT calculation too, it will always restart from scratch.

Thank you!


Forum >> NWChem's corner >> General Topics