why is this file running so slowly

GW, Bethe-Salpeter …

Moderators: maryam.azizi, bruneval

Locked
jiangx3
Posts: 32
Joined: Fri Apr 30, 2010 9:33 pm

why is this file running so slowly

Post by jiangx3 » Fri Jul 16, 2010 3:42 pm

Dear abinitors,

I have two jobs: one is doing gw calculation of optimized Arm chair Graphene Nanoribbons n=6 (AGNR6) and the other is doing gw calculation of non-optimized Graphene Nanoribbons n=5(AGNR5). I attach their input files here.
Attachments
g.in
the input file for optimized AGNR6
(3.95 KiB) Downloaded 302 times
gw.in
the input file for non-optimized AGNR5
(3.82 KiB) Downloaded 265 times

jiangx3
Posts: 32
Joined: Fri Apr 30, 2010 9:33 pm

Re: why is this file running so slowly

Post by jiangx3 » Fri Jul 16, 2010 3:56 pm

I am sorry I have not finished my problem. The optimized AGNR6 one is running successfully within 18 hours. But the non-optimized AGNR5 is running slowly and it seems that it would never finish because it already calculated for more than 50 hours. The output file of the job shows it stays on the same place for almost 20 hours.

I compared the two input files and the difference are (1-non-optimized AGNR5; 2- optimized AGNR6)
< nband2 128 ----- 1
---
> nband2 128 ------2
59,63c59,65
< kptgw 0.0000 0.3750 0.0000
< 0.0000 0.4063 0.0000
< 0.0000 0.4375 0.0000
< 0.0000 0.4688 0.0000
< 0.0000 0.5000 0.0000 ------1
---
> kptgw 0.0 0.0 0.0
> 0.0 0.125 0.0
> 0.0 0.25 0.00
> 0.0 0.375 0.0
> 0.0 0.5 0.0 -------2
>
> bdgw 22 23 22 23 22 23 22 23 22 23 -------1
65d66
< bdgw 26 27 26 27 26 27 26 27 26 27 -------2
69c70
< acell 16.60553 4.24423 6.0 Angstrom -------1
---
> acell 3.0235618000E+01 8.0223409349E+00 1.1338356750E+01 # lengths of the primitive vectors (in Bohr) # 16 4.245240306 6 Angstrom -------2
83,96c84,98
< xangst 1.72520127 3.32947924 0.00000000
< 1.72520127 0.50000000 0.00000000
< 2.95040254 1.20736443 0.00000000
< 5.40080507 1.20736443 0.00000000
< 2.95040254 2.62211481 0.00000000
< 5.40080507 2.62211481 0.00000000
< 4.17560381 3.32947924 0.00000000
< 6.62600634 3.32947924 0.00000000
< 4.17560381 0.50000000 0.00000000
< 6.62600634 0.50000000 0.00000000
< 7.55265161 2.79447592 0.00000000
< 7.55265161 1.03500332 0.00000000
< 0.79855600 1.03500332 0.00000000
< 0.79855600 2.79447592 0.00000000 ----------1
---
> xangst # Reduced coordinate of atoms, xangst, xangst
> 4.4657934906E+00 9.5730038436E-01 0.0000000000E+00
> 4.4657762040E+00 2.3378991241E+00 0.0000000000E+00
> 5.6925709288E+00 2.3571111751E-01 0.0000000000E+00
> 5.6925560909E+00 3.0594817742E+00 0.0000000000E+00
> 6.9373577498E+00 9.3611360256E-01 0.0000000000E+00
> 6.9373548622E+00 2.3590859341E+00 0.0000000000E+00
> 8.1822795909E+00 2.3572609726E-01 0.0000000000E+00
> 8.1822846462E+00 3.0594774338E+00 0.0000000000E+00
> 9.4093299508E+00 9.5725940714E-01 0.0000000000E+00
> 9.4093395107E+00 2.3379620154E+00 0.0000000000E+00
> 3.4741166689E+00 4.1107517197E-01 0.0000000000E+00
> 3.4740828723E+00 2.8840809275E+00 0.0000000000E+00
> 1.0402324056E+01 4.1269718594E-01 0.0000000000E+00
> 1.0402330008E+01 2.8825165771E+00 0.0000000000E+00 ----------2


Both of them are calculating 2 bands and 5 kpoints in self energy. But why is the time consumed so much different?


Thanks and Best regards,
Xueping Jiang
Attachments
gw2.out
the output file for non-optimized AGNR5
(39.98 KiB) Downloaded 256 times

bruneval
Posts: 40
Joined: Mon Aug 17, 2009 11:38 am

Re: why is this file running so slowly

Post by bruneval » Mon Jul 19, 2010 12:25 pm

One of your systems is 8 times larger than the other one... With a tiny bit of imagination, this may lead you to some kind of answer.

Furthermore, the output does not correspond to the input you provide.

Fabien

jiangx3
Posts: 32
Joined: Fri Apr 30, 2010 9:33 pm

Re: why is this file running so slowly

Post by jiangx3 » Mon Jul 19, 2010 3:32 pm

Dear Fabien,

Thanks very much for your reply. Could you tell me where you can see one file is 8 times than the other? Do you mean the memory?
The output file is for the non-optimized one, which is for the 3 and 4 steps of 'gw.in'. Because at first I run the job and it can not finish during the time limit, I restart the job from step 3 with gw2.in. I attach all of the output and input files for non-optimized one.

Do you think the reason I can not run successfully is that the abinit version I used is too old(abinit-5.4.4p)?


Thanks and Best regards,
Xueping Jiang
Attachments
gw2.in
the restart input file
(3.15 KiB) Downloaded 279 times
gw.out
the first out put file for non-optimized GNR
(231.32 KiB) Downloaded 240 times

bruneval
Posts: 40
Joined: Mon Aug 17, 2009 11:38 am

Re: why is this file running so slowly

Post by bruneval » Thu Jul 22, 2010 8:53 am

Check your acell values!

Locked