ERROR: genwf: rkq mismatch

278 views
Skip to first unread message

yu pang

unread,
Sep 12, 2023, 11:23:09 PM9/12/23
to BerkeleyGW Help

Dear BerkeleyGW Developers,

I am facing an issue while running sigma.x with nns subsampling, and I'm getting the following error:

================================================================================
 08:26:59   Dealing with k =  0.000000  0.000000  0.000000                1 / 64
================================================================================

 Reading vxc.dat
 Number of q-points in the irreducible BZ(k) (nrq): 40

 Started calculating Sigma with 861 block(s) at 08:26:59.
 [ 08:27:00 |   0% ] block   1 / 861.
 [ 08:27:07 |   0% ] block   3 / 861, remaining: 2989 s.
 [ 08:28:22 |  10% ] block  87 / 861, remaining: 714 s.
 [ 08:28:32 |  20% ] block 173 / 861, remaining: 358 s.

ERROR: genwf: rkq mismatch

forrtl: error (76): Abort trap signal
Image              PC                Routine            Line        Source
sigma.cplx.x       000000000295BC29  Unknown               Unknown  Unknown
sigma.cplx.x       000000000295A4FE  Unknown               Unknown  Unknown
sigma.cplx.x       00000000029176B2  Unknown               Unknown  Unknown
sigma.cplx.x       00000000028B5558  Unknown               Unknown  Unknown
sigma.cplx.x       00000000028BC652  Unknown               Unknown  Unknown
libpthread.so.0    00002B9CE47985D0  Unknown               Unknown  Unknown
libc.so.6          00002B9CE5997207  Unknown               Unknown  Unknown
libc.so.6          00002B9CE59988F8  Unknown               Unknown  Unknown
sigma.cplx.x       0000000000409CD0  Unknown               Unknown  Unknown
sigma.cplx.x       00000000004E1177  Unknown               Unknown  Unknown
sigma.cplx.x       0000000000558B1D  Unknown               Unknown  Unknown
sigma.cplx.x       0000000000405C36  Unknown               Unknown  Unknown
libc.so.6          00002B9CE59833D5  Unknown               Unknown  Unknown
sigma.cplx.x       0000000000405B29  Unknown               Unknown  Unknown

I'm unsure how to resolve this problem. I have included all relevant input and output files in the appendix for your reference.

I greatly appreciate your assistance and patience!

Best,

Yu Pang

vs2.zip

Mauro Del Ben

unread,
Sep 13, 2023, 8:00:28 PM9/13/23
to yu pang, BerkeleyGW Help
Hi Yu,

I assume the epsilon calculation ran fine, and you see the problem only in sigma. Try to use consistent units in the kgrid.in and wfn.in, that is use angstrom for both lattice parameters and coordinates. Also try to switch off symmetry in wfn.inp, nosym= .true.

Best

-M


--
You received this message because you are subscribed to the Google Groups "BerkeleyGW Help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to help+uns...@berkeleygw.org.
To view this discussion on the web visit https://groups.google.com/a/berkeleygw.org/d/msgid/help/4d1ba4ac-7f1d-4dd4-bd75-f2b0c028ce8an%40berkeleygw.org.

yu pang

unread,
Sep 18, 2023, 9:42:08 PM9/18/23
to BerkeleyGW Help, mde...@lbl.gov, BerkeleyGW Help

Dear -M,

Thank you for your prompt response. Following your advice, I have taken the following steps:

  1. Ensured consistent units (angstroms) for lattice parameters and coordinates in both the kgrid.in and wfn.in files.
  2. Disabled symmetry by setting nosym = .true. in the wfn.in file.

However, I have encountered the following error when attempting to run epsilon.x:

================================================================================
 11:11:19   Dealing with q =  0.000000  0.083333  0.000000               11 / 40
================================================================================

 This is a regular non-zero q-point.
 Rank of the polarizability matrix (nmtx): 3697
 
 BLACS processor grid:   8 x   8; BLOCKSIZE =   32
 
 Number of k-points in the irreducible BZ(q) (nrk): 31
 
 Started calculation of matrix elements with 403 transition(s) at 11:11:22.
 [ 11:11:22 |   0% ] transition   1 / 403.
 [ 11:11:22 |   0% ] transition   3 / 403, remaining: 101 s.
 
ERROR: genwf_mpi: No match for rkq point:   0.083   0.917   0.000 in file WFN
 
 
ERROR: genwf_mpi: No match for rkq point:   0.083   0.917   0.000 in file WFN


 
forrtl: error (76): Abort trap signal
Image              PC                Routine            Line        Source            

epsilon.cplx.x     0000000002DB7759  Unknown               Unknown  Unknown
epsilon.cplx.x     0000000002DB602E  Unknown               Unknown  Unknown
epsilon.cplx.x     0000000002D731E2  Unknown               Unknown  Unknown
epsilon.cplx.x     0000000002D11088  Unknown               Unknown  Unknown
epsilon.cplx.x     0000000002D18182  Unknown               Unknown  Unknown

......

I am uncertain about how to resolve this issue. Notably, I used direct coordinates in the wfn.in file but Cartesian coordinates in the kgrid.in. Could this discrepancy be a factor contributing to the error?

I also attempted to calculate epsilon without NNS subsampling (located in the directory bgw/02-epsilon-03_wfnq), but I still received the same error.

For your reference, I have attached all the relevant input and output files. The output of wfn_rho_vxc_info.x for both WFN and WFNq can be found in their respective folders, specifically in wfn_check.out and wfnq_check.out.

I greatly appreciate your assistance in resolving this matter.

Best,

Yu

vs2.zip

yu pang

unread,
Oct 2, 2023, 12:11:39 AM10/2/23
to BerkeleyGW Help, yu pang, mde...@lbl.gov, BerkeleyGW Help
Dear -M, 

I encountered the same error message when running epsilon.x on both kgrid.in and wfn.in with the same coordinates and lattice parameters, while setting nosym=.true. in wfn.in:

ERROR: genwf_mpi: No match for rkq point: 0.083 0.917 0.000.

I then attempted to use the same coordinates and lattice parameters in both kgrid.in and wfn.in while setting nosym=.false. in wfn.in for epsilon calculations, and it ran successfully. However, when running sigma, I encountered the initial error message again:

ERROR: genwf: rkq mismatch.

I have included the input and output files related to nosym=.true.in the appendix.


Best,

Yu

vs2.zip

yu pang

unread,
Oct 4, 2023, 7:08:45 AM10/4/23
to BerkeleyGW Help, yu pang, mde...@lbl.gov, BerkeleyGW Help

Dear -M,

I am using a 12x12x1 k-grid for the wfn in epsilon calculations and a 18x18x1 k-grid for sigma calculations.

Best,

Yu

yu pang

unread,
Oct 8, 2023, 2:29:04 AM10/8/23
to BerkeleyGW Help, yu pang, mde...@lbl.gov, BerkeleyGW Help

Hi -M,

Thank you for your response.

Following your advice, I have used the same grid for both epsilon and sigma calculations and have set nosym= .False. in the wfn.in file. Both epsilon and sigma calculations proceed smoothly without any errors.

However, I find myself puzzled by the following issue: When I set nosym= .True. and used the same grid for both epsilon and sigma , I encounter the following error during epsilon calculation:

ERROR: genwf_mpi: No match for rkq point: 0.083 0.917 0.000 in file WFN.

Additionally, based on the example provided at https://github.com/BerkeleyGW/BerkeleyGW-examples/tree/master/DFT/MoS2_SOC, it appears that sigma and epsilon calculations can use different grids. I have followed this example for my previous calculations and have not encountered any errors , but I am uncertain why I am receiving an "ERROR: genwf: rkq mismatch" when using a different k-grid this time.

Do you have any insights on these two issues?

Best,

Yu

yu pang

unread,
Oct 8, 2023, 7:54:27 AM10/8/23
to BerkeleyGW Help, yu pang, mde...@lbl.gov, BerkeleyGW Help

Hi -M,

Yes, I generated the grid using the kgrid.x utility of BGW. However, I'm not entirely clear on what you mean by "usually by shifting by half the grid for epsilon." Could you please elaborate a bit more? Are you suggesting that if I want to use an 18X18X1 grid to calculate sigma, I should use a 9X9X1 grid for epsilon? Just like the example provided at https://github.com/BerkeleyGW/BerkeleyGW-examples/tree/master/DFT/MoS2_SOC, where it used a 12X12X1 grid for sigma calculations, while the grid used for epsilon was 6X6X1.

Best, 

Yu

Reply all
Reply to author
Forward
0 new messages