[salmon-user:00346] something wrong happened in TDDFT+maxwell calculaiton

[宁 雅恬]

Dear Salmon developers,

I'm starting the MoS2 calculation(a new user), but I found that the calculation can't proceed and an error occurred. I checked and found that the cause is a memory leak. Is there any way to reduce the memory?There are my input,i use intel-mkl-2021 to make the salmon.

 ##########################

#BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES = RANK 0 PID 1657697 RUNNING AT HPE-#Debian = KILLED BY SIGNAL: 9 (Killed)

############################

Best Wishes!

ytning Ning

[Otobe Tomohito]

Dear Ning-san

 

Thank you for the question.

According to the output file, the number of electrons is not consistent with input file, “Ne=   24095.5187993038”.

I suppose that some of the parameters for time evolution and ground state (GS) calculations may not match.

Could you check the input file for GS calculation?

 

乙部 智仁

 

Tomohito Otobe

 

Senior Principal Researcher

Kansai Photon Science Institute (KPSI)

National Institutes for Quantum Science and Technology (QST)

 

otobe.t...@qst.go.jp

+81-774-80-8899

 

 

 

差出人: 宁 雅恬 <ytni...@outlook.com>
日付: 水曜日, 2025年6月11日 12:43
宛先: salmo...@salmon-tddft.jp <salmo...@salmon-tddft.jp>
件名: [salmon-user:00346] something wrong happened in TDDFT+maxwell calculaiton

[宁 雅恬]

Dear Salmon developers,

 

I have cheaked my gs-inputfile and gs-outputfile, it seems  doing well.Ne=36.0000000 in outputfile.The parameters I have in the gs and td calculations are consistent. But the wfn.bin file is 1.29GB. Is there a problem with this?

---

发件人: Otobe Tomohito <otobe.t...@qst.go.jp>
发送时间: 2025年6月11日 15:13
收件人: salmo...@salmon-tddft.jp <salmo...@salmon-tddft.jp>
主题: [salmon-user:00347] Re: something wrong happened in TDDFT+maxwell calculaiton

 

[Otobe Tomohito]

Dear Ning-san

 

Thank you for the GS information.

In the GS calculation, you defined the MPI process as follows.

   nproc_k = 2

   nproc_ob = 1

   nproc_rgrid(1) =2

   nproc_rgrid(2) = 2

   nproc_rgrid(3) = 4

On the other hand, in RT calculation you did not define MPI parameters.

Please use same MPI processes as GS calculation.

I recommend reuse GS input file, and change “theory” (and add laser parameters ) to avoid such error.

 

Best regards,

 

乙部 智仁

 

Tomohito Otobe

 

Senior Principal Researcher

Kansai Photon Science Institute (KPSI)

National Institutes for Quantum Science and Technology (QST)

 

otobe.t...@qst.go.jp

+81-774-80-8899

 

 

 

差出人: 宁 雅恬 <ytni...@outlook.com>
日付: 水曜日, 2025年6月11日 18:08
宛先: salmo...@salmon-tddft.jp <salmo...@salmon-tddft.jp>
件名: [salmon-user:00348] 回复: something wrong happened in TDDFT+maxwell calculaiton

[Shunsuke Yamada]

Dear Ning-san, Otobe-san,

I would like to add some points I realized.

The error in MoS2_rt_multiscale.out seems to be due to larger time step dt. For stable RT calculations, we should use sufficiently smaller dt for retaining the norm of the wavefunctions.

I recommend trying usual TDDFT calculations (theroy=’TDDFT_pulse’) before the multiscale method. After confirming convergence of grids (r-grid num_rgrid, k-grid num_kgrid, t-grid dt), pls proceed multiscale calculations. The required memory of the multiscale method is exactly equal to the multiple of that of single TDDFT calculation and the number of macro-grid points (here, nx_m = 8).

Sincerely,

Shunsuke

山田俊介
Shunsuke Yamada
量子技術基盤研究部門 関西光量子科学研究所
Kansai Institute for Photon Science, National Institutes for Quantum Science and Technology

2025/06/11 12:15、Otobe Tomohito <otobe.t...@qst.go.jp>のメール:

[Otobe Tomohito]

Dear Yamada-san

 

Thank you for your comments.

Maybe, you are right

Best,

 

乙部 智仁

 

Tomohito Otobe

 

Senior Principal Researcher

Kansai Photon Science Institute (KPSI)

National Institutes for Quantum Science and Technology (QST)

 

otobe.t...@qst.go.jp

+81-774-80-8899

 

 

 

差出人: Shunsuke Yamada <yamada...@gmail.com>
日付: 水曜日, 2025年6月11日 18:34
宛先: salmo...@salmon-tddft.jp <salmo...@salmon-tddft.jp>
件名: [salmon-user:00350] Re: something wrong happened in TDDFT+maxwell calculaiton

[宁 雅恬]

Dear  Otobe and Yamada,

It was running when I tried to reduce the time step size. Thank you very much for your help!

Sincerely,

Yatian Ning

---

发件人: Shunsuke Yamada <yamada...@gmail.com>
发送时间: 2025年6月11日 17:32
收件人: salmo...@salmon-tddft.jp <salmo...@salmon-tddft.jp>
主题: [salmon-user:00350] Re: something wrong happened in TDDFT+maxwell calculaiton