search

tfrederiksen / inelastica

Python package for eigenchannels, vibrations and inelastic electron transport based on SIESTA/TranSIESTA DFT
https://tfrederiksen.github.io/inelastica
GNU Lesser General Public License v3.0
33 stars 16 forks source link

Inquiry about EigenChannels limitations #58

Closed el-abed closed 4 years ago

el-abed commented 4 years ago

Good afternoon to whom it may concern, I was trying to calculate the EigenChannels for different phospherene/graphene systems with different number of atoms. Some would end successfully. But as the number of atoms increases ( by a 100 or so) the job would be killed. I wanted to ask if there are limits to EigenChannels? If not, can it be run parallel? thank you very much and looking forward to your thoughts

tfrederiksen commented 4 years ago

It could be that you are running out of memory for large systems, but without more info it is difficult to say. Feel free to share more details here about a problematic run.

Generally speaking, to limit the computational burden of EigenChannels it is essential to keep the device region to a minimum. This is controlled with the keywords -F and -L(or equivalently --DeviceFirst and --DeviceLast).

el-abed commented 4 years ago

Thank you very much Thomas for the reply. I have managed to run in parallel the system with 8 cores and the run finished successfully. I was wondering from a physical stand point, shouldn’t it be more wisely to study the entire system not only the scattering region to find out if conductance does take place or not? So for example if the eigenchannels runs through out the entire system is an indication of high conductance. Where can I read more about such insight? Thank you very much again and hope to read your reply. EL-abed

Sent from Mailhttps://go.microsoft.com/fwlink/?LinkId=550986 for Windows 10

From: Thomas Frederiksenmailto:notifications@github.com Sent: Tuesday, 5 May 2020 7:40 PM To: tfrederiksen/inelasticamailto:inelastica@noreply.github.com Cc: El-abed Haidarmailto:ehai2584@uni.sydney.edu.au; Authormailto:author@noreply.github.com Subject: Re: [tfrederiksen/inelastica] Inquiry about EigenChannels limitations (#58)

It could be that you are running out of memory for large systems, but without more info it is difficult to say. Feel free to share more details here about a problematic run.

Generally speaking, to limit the computational burden of EigenChannels it is essential to keep the device region to a minimum. This is controlled with the keywords -F and -L(or equivalently --DeviceFirst and --DeviceLast).

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/tfrederiksen/inelastica/issues/58#issuecomment-623956155, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AENPWP3NCG2SDDCRHXF4EHDRP7NHPANCNFSM4MXQ36UQ.