No output file

[wanghong]

Dear all:
When I ran the 2018 version of Dalton, the system was automatically interrupted without a clear error message.I don't know what to do, so I would like to ask all teachers to help me check the problem.I would appreciate it if you could give me some advice.
Attached are my input files and output files and the content in the image is the content in my run.sh file.
Looking forward to your reply.

[hjaaj]

"-mw 15000" sets work memory to 15000 dp words, not 15000 megawords. I suggest to use the -gb or -mb option instead, which defines work memory in gigabytes and megabytes, respectively.

[magnus]

Also note that it is given as memory per MPI process, so that "mpirun -np 32" combined with "dalton -gb 2" will use 32*2 gb in total.

[wanghong]

Thank you for your reply

[wanghong]

Hello,professor:
I have a problem that I want to ask you.When I set ROOTS 20 in the input file, I will calculate that the file will be abnormal interruption, and there are the following prompts in the output file. Can you help me? Looking forward to your reply.

WARNING: all PP,LR,C6 input ignored because
because cubic or quadratic response has been requested.


@ MPI MASTER, node no.: 0
@ Reason: INQRTR: TOO MANY PROPERTIES SPECIFIED

[hjaaj]

Edit ..../DALTON/include/rspprp.h and change MAXLBL from 1000 to, say, 2000 and rebuild dalton.

[wanghong]

Hello, Sir, I want to ask how many states SOC can calculate at most. I changed the data in rspprp.h file to PARAMETER ( MAXLBL = 20000 ), and only 55 states can be calculated. Do I want to calculate 100 States? I would appreciate it if you could help me.
Looking forward to your reply.

[hjaaj]

I cannot tell if you need 100 states for your project. How many SOC states you can calculate at most depends on how much memory there is available for Dalton and the setting of MAXLBL and maybe other parameters. With a value of MAXLBL=20000 I would have expected you could calculate 80 states, so maybe you encountered another limitation?

I know why the MAXLBL demand grows. For 100 singlet and 100 triplet states, for the quadratic response you need labels for all 10,000 (100*100) energy differences between any triplet state and any singlet state times 3 (the 3 components of the spin orbit operator). So you would need MAXLBL to be more than 30000 to be able to calculate 100 states. By the way, I think (that is, I have not tested it) that you can run your SOC by changing MXLRQR in include/indqr.h file instead of MAXLBL in rspprp.h file. This will reduce the static memory usage by Dalton.

[wanghong]

So how can I changing MXLRQP in include/indqr.h file?

[wanghong]

I'm sorry to tell you that when I changed this data in rspprp.h file to PARAMETER ( MAXLBL = 30000 ), there are still only 55 states in the output file.Maybe what you said to change MXLRQR in include/indqr.h file is right, but I don't know how to change it? I would really appreciate it if you could help me.

[hjaaj]

1) you need to upload the output file for me to have any chance of answering the question about only 55 states
2) changing include/indqr.h file instead of include/rspprp.h file only reduces the static memory in Dalton, it does not make it possible to run more states. To do that, change "MXLRQR=MAXLBL" to "MXLRQR=30000".

[hjaaj]

I have looked at the output file you sent me by private message. The reason you only get 55 states is that 55 is the number of orbital rotations for your small test case, and 55 is therefore also the maximum number of excited states you can calculate in the response (TD-DFT) part.
If you ask for more states than the max dimension in your calculation, then Dalton will reduce the request to the max dimension.

[wanghong]

OK, I see. Thank you very much.