QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: pervanem on August 31, 2018, 15:29

Title: NEGF calculation issue between gold electrode
Post by: pervanem on August 31, 2018, 15:29: Hello,
I try to perform the NEGF calaculation as shown picture3
I was perform in ATK 2016.4 version in a week.
now I can not carry out this calculation (in ATK 2017.2 and QuantumATK 2018.06). I waited more than a week. it is not be...

+----------------------------------------------------------+
+------------------------------------------------------------------------------+
| |
| Transmission Spectrum Analysis |
| |
+------------------------------------------------------------------------------+

|--------------------------------------------------|
Calculating Transmission : ==================================================

+------------------------------------------------------------------------------+
| IV-Curve Calculation |
+------------------------------------------------------------------------------+
| The current will be calculated at 10 points between 0.000 to 2.000 V. |
| Note: The current has already been calculated at 0.000 V. |
| Each bias point will be calculated using 1/1 MPI process. |
| |
| Bias Point Process Group Log File |
| 0.200 0 ivcurve_0.20000V.log |
| 0.400 0 ivcurve_0.40000V.log |
| 0.600 0 ivcurve_0.60000V.log |
| 0.800 0 ivcurve_0.80000V.log |
| 1.000 0 ivcurve_1.00000V.log |
| 1.200 0 ivcurve_1.20000V.log |
| 1.400 0 ivcurve_1.40000V.log |
| 1.600 0 ivcurve_1.60000V.log |
| 1.800 0 ivcurve_1.80000V.log |
| 2.000 0 ivcurve_2.00000V.log |
+------------------------------------------------------------------------------+

after that, how long is the calculation? I passed more than a week but it was not finished...
am I doing wrong? i am making a mistake?

please, can you help me?
Title: Re: NEGF calculation issue between gold electrode
Post by: Ulrik G. Vej-Hansen on August 31, 2018, 15:57: We changed some defaults between 2016 and 2017, so that could be the reason. Can you see if it is progressing as it should, just more slowly than before?
Title: Re: NEGF calculation issue between gold electrode
Post by: pervanem on September 3, 2018, 11:25: it is probably due to default settings... :-\
I think this is taking too long... more than a week.. is it normal?

I added the previous and next log files. (2016 and 2017/2018 version)
would you check them?
thanks...
Title: Re: NEGF calculation issue between gold electrode
Post by: pervanem on September 3, 2018, 11:26: contiuned...files..
Title: Re: NEGF calculation issue between gold electrode
Post by: pervanem on September 3, 2018, 11:29: I deleted part of the file (due to max 400kb limit)
Title: Re: NEGF calculation issue between gold electrode
Post by: Petr Khomyakov on September 3, 2018, 11:40: Please post the python script(s) related to this calculation.
Title: Re: NEGF calculation issue between gold electrode
Post by: pervanem on September 3, 2018, 12:02: I performed as shown in picture..(for 2018.06)

Note: in 2016.4, there is not transmission spectrum calculation (for device cu).
Title: Re: NEGF calculation issue between gold electrode
Post by: Petr Khomyakov on September 3, 2018, 13:21: This is likely because of new defaults in QuantumATK-2018.06. Your scripts were save with 'Minimal' option in 2016.4, i.e., you did not use 'Show defaults' when saving the script. In 2017 and 2018, the default basis sets are different (more accurate, but more expensive at the same time) from that of 2016 version. Also, for example, temperature broadening for the Fermi-Dirac distribution is now set to 1000 K instead of 300 K.
Title: Re: NEGF calculation issue between gold electrode
Post by: pervanem on September 3, 2018, 15:33: thank you your reply...
I noticed the temperature change or other settings...
yes.. in this case, the account will be more cost and more accurate...