I am doing a DFT calculation in Gaussian for a complex which consist of two molecules and 284 atoms, I tried different method and basis set but I received the following error each time. can anyone help me to solve this problem?
the erroe is :
Error termination via Lnk1e in /opt/gaussian/g09/l103.exe at Tue Apr 11 09:13:42 2017.
Job cpu time: 4 days 2 hours 5 minutes 30.2 seconds.
File lengths (MBytes): RWF= 1434 Int= 0 D2E= 0 Chk= 210 Scr= 1
Error: segmentation violation
rax 0000000000000000, rbx ffffffffffffffff, rcx ffffffffffffffff
rdx 000000000000183d, rsp 00007ffd8a819208, rbp 00007ffd8a8197f0
rsi 000000000000000b, rdi 000000000000183d, r8 00002ba450fe3180
r9 0000000000000000, r10 00007ffd8a818f90, r11 0000000000000202
r12 0000000000000000, r13 0000000000000000, r14 00007ffd8a819838
r15 00000000000003e6
--- traceback not available
'Segmentation violation' generally (not specifically in Gaussian) refers to overwriting one or some arrays in the program. So, you should check whether the arrays as defined in your program are sufficiently long for the calculation at hand.
Incidentally, the relevant calculations being time-consuming (taking over 4 days to complete), you should write a small subroutine that calculates the required array lengths and compares these with the declared lengths at the very outset of the calculations, so that if there is something amiss with the array lengths, you know it almost immediately, and not some four days later.
Dear Sara,
Your problem needs more evaluation of your input file, so please upload the input and hopefully I can assist you.
You should run your job at lower level of method ( HF method) for optimization. After optimization, you should run DFT job of optimized geometry which has done at lower level of method. I hope your calculation may be completed in this way.
Dear Dr. Nand Kishor,
I have tried your suggestion before and did calculation with HF and small basis set but there is still the same error
Dear Dr. Sayyed Jalil Mahdizadeh
Thank you so much for your assist. I have changed the input file according to your suggestion if there is still an error I will upload the file for more help. thank you so much.
Yeah, I would like to see the input and the output file. Did it run 4 days then crash?
If you have available several servers then use other software. You can find free software to execute in parallel and distribute tasks and memory. Best regards.
Hello
in first time you can run your job at HF method ( HF method) for optimization. you can run at DFT job of optimized geometry after first level HF which has done at lower level of method. I said your calculation may be completed in this way!
Dear Sara
The l103 error is definitely related to your geometry input, So, please put here what type of atoms you have and your input file.
Dear Dr Mahdizadeh and Dr. Rastegar and other friends, my input file is attached. would you please help me, I tried different methods and basis sets. I started with HF/3-21g but still crash after few days.
please check this input file and if it faces a problem, send the output.
regards
You can easily find the errors related to links from Gaussian with Google...
https://www.ace-net.ca/wiki/Gaussian_Error_Messages#Linear_search_skipped_for_unknown_reason
Follow their suggestion with opt=calcfc
However, looking at the input file I would say that you can perhaps start with even a lower level of theory (perhaps semiempirical or molecular mechanics methods) in order to get a reasonable structure. Specially considering that you can have a very large amount of different conformers...
Dear Sara, my suggestions are as follow:
Use commands as Somayeh recommended but optimize each part of your structure separately. In this way, your input files would be much smaller and easier for optimization. After two parts are optimized, put them next to each other and re-optimize the whole system at the same level of theory. Goodluck
Is your version SMP? (multi processor, one node)? What revision? I've had a segmentation fault in Gaussian 09 D.01 smp whenever i set it to use more processors than available (I had a multithreading i7 that reported to have 8 processors but actually has 4) or when I set it to use more than 2GB of ram, no idea why. Maybe check the number of processors and lower the RAM usage?
Dear Dr. Mahdizadeh
I optimized each system separately first and then built the complex from the beginning. Now I am using the input structure which suggested by Somayyeh. but I used the ONIOM method for optimization. I used DFT for the main part of the molecule which is our interest and semiempirical for the rest. it is still running. do you think this method is acceptable?
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry