From tjesser at ucdavis.edu Wed Oct 4 10:19:58 2017 From: tjesser at ucdavis.edu (Tyler Esser) Date: Wed, 4 Oct 2017 10:19:58 -0700 Subject: [CIG-SEISMO] MINEOS Installation problems Message-ID: Hi Paul, I think the first thing to try would be to run this command in the terminal you are using: > export CFLAGS='-Wno-error=return-type' > Then you would rerun the *./configure*, *make clean*, and *make* commands. If you still encounter an error, please provide the output of: > gcc --version > as well as the error you received. I assume you are using the gcc version supplied by Xcode? Unfortunately, the Mineos devs are no longer active, so this may be a difficult problem. Tyler Esser -------------- next part -------------- An HTML attachment was scrubbed... URL: From maurya2278satish at gmail.com Wed Oct 4 12:06:04 2017 From: maurya2278satish at gmail.com (satish maurya) Date: Wed, 4 Oct 2017 12:06:04 -0700 Subject: [CIG-SEISMO] Mineos: sensor depth => edepth not working? Message-ID: Dear all, I try to compute synthetic seismogram for receiver's place at depth 0, 50,100...300km, but seismogram is same. I put these depths in .sitechan file and elev is 0.0 in all case. I was wondering, if this parameter is implemented in the code or I am doing something wrong. Any suggestion! Regards Satish Maurya Post-Doc University of California, Berkeley Seismological Laboratory, Dept. of Earth & Planetary Science 201 The Tower Room, McCone Hall Berkeley, CA 94720 -------------- next part -------------- An HTML attachment was scrubbed... URL: From phil.cummins at anu.edu.au Wed Oct 11 01:27:18 2017 From: phil.cummins at anu.edu.au (Phil Cummins) Date: Wed, 11 Oct 2017 19:27:18 +1100 Subject: [CIG-SEISMO] specfem2d: snapshots of domain subset Message-ID: <59DDD5E6.8050307@anu.edu.au> Hi, I'm using spefem2d to compute seismic waves n a sedimentary basin. Does anyone know how I can output snapshots of a subset of the computational domain only, not the whole thing? I want to view the seismic waves in a shallow basin that is barely visible in a snapshot of the whole computational domain. I.e., I want to zoom in on the basin. Many thanks! -- Phil Cummins Prof. Natural Hazards Research School of Earth Sciences Australian National University (+61/0)-2-6125-1217 From komatitsch at lma.cnrs-mrs.fr Wed Oct 11 04:12:15 2017 From: komatitsch at lma.cnrs-mrs.fr (Dimitri Komatitsch) Date: Wed, 11 Oct 2017 13:12:15 +0200 Subject: [CIG-SEISMO] specfem2d: snapshots of domain subset In-Reply-To: <59DDD5E6.8050307@anu.edu.au> References: <59DDD5E6.8050307@anu.edu.au> Message-ID: Hi Phil, It is not implemented, but in src/specfem2D/prepare_color_image.F90 if you replace ! horizontal min and max coordinates of the image xmin_color_image_loc = minval(coord(1,:)) xmax_color_image_loc = maxval(coord(1,:)) ! vertical min and max coordinates of the image, slightly increase it to go beyond maximum topography zmin_color_image_loc = minval(coord(2,:)) zmax_color_image_loc = maxval(coord(2,:)) with the min and max that you want it will probably work (please let us know). Thanks, Dimitri. On 10/11/2017 10:27 AM, Phil Cummins wrote: > Hi, > > I'm using spefem2d to compute seismic waves n a sedimentary basin. > Does anyone know how I can output snapshots of a subset of the > computational domain only, not the whole thing? I want to view the > seismic waves in a shallow basin that is barely visible in a snapshot of > the whole computational domain. I.e., I want to zoom in on the basin. > Many thanks! > -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr From phil.cummins at anu.edu.au Wed Oct 11 07:47:00 2017 From: phil.cummins at anu.edu.au (Phil Cummins) Date: Wed, 11 Oct 2017 14:47:00 +0000 Subject: [CIG-SEISMO] specfem2d: snapshots of domain subset In-Reply-To: References: <59DDD5E6.8050307@anu.edu.au>, Message-ID: Hi Dimitri, Well try it and let yo know. Thanks! - Phil ________________________________ From: Dimitri Komatitsch Sent: Wednesday, October 11, 2017 10:12:15 PM To: cig-seismo at geodynamics.org; Phil Cummins Subject: Re: [CIG-SEISMO] specfem2d: snapshots of domain subset Hi Phil, It is not implemented, but in src/specfem2D/prepare_color_image.F90 if you replace ! horizontal min and max coordinates of the image xmin_color_image_loc = minval(coord(1,:)) xmax_color_image_loc = maxval(coord(1,:)) ! vertical min and max coordinates of the image, slightly increase it to go beyond maximum topography zmin_color_image_loc = minval(coord(2,:)) zmax_color_image_loc = maxval(coord(2,:)) with the min and max that you want it will probably work (please let us know). Thanks, Dimitri. On 10/11/2017 10:27 AM, Phil Cummins wrote: > Hi, > > I'm using spefem2d to compute seismic waves n a sedimentary basin. > Does anyone know how I can output snapshots of a subset of the > computational domain only, not the whole thing? I want to view the > seismic waves in a shallow basin that is barely visible in a snapshot of > the whole computational domain. I.e., I want to zoom in on the basin. > Many thanks! > -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr -------------- next part -------------- An HTML attachment was scrubbed... URL: From komatitsch at lma.cnrs-mrs.fr Wed Oct 11 08:21:13 2017 From: komatitsch at lma.cnrs-mrs.fr (Dimitri Komatitsch) Date: Wed, 11 Oct 2017 17:21:13 +0200 Subject: [CIG-SEISMO] specfem2d: snapshots of domain subset In-Reply-To: References: <59DDD5E6.8050307@anu.edu.au> Message-ID: Hi Phil, Perfect! If you use MPI i.e. if you use more than one processor core, the right change should probably be: ! horizontal min and max coordinates of the image xmin_color_image_loc = max(minval(coord(1,:)), yyyyyy_put_here_the_min_X_that_you_want_yyyyyyyyyy) xmax_color_image_loc = min(maxval(coord(1,:), yyyyyy_put_here_the_max_X_that_you_want_yyyyyyyyyy) ! vertical min and max coordinates of the image, slightly increase it to go beyond maximum topography zmin_color_image_loc = max(minval(coord(2,:), yyyyyy_put_here_the_min_Z_that_you_want_yyyyyyyyyy) zmax_color_image_loc = min(maxval(coord(2,:), yyyyyy_put_here_the_max_Z_that_you_want_yyyyyyyyyy) If you use a single slice (no MPI), this should do it: ! horizontal min and max coordinates of the image xmin_color_image_loc = yyyyy_put_here_the_min_X_that_you_want_yyyyyyyyyy xmax_color_image_loc = yyyyyy_put_here_the_max_X_that_you_want_yyyyyyyyyy ! vertical min and max coordinates of the image, slightly increase it to go beyond maximum topography zmin_color_image_loc = yyyyyy_put_here_the_min_Z_that_you_want_yyyyyyyyyy zmax_color_image_loc = yyyyyy_put_here_the_max_Z_that_you_want_yyyyyyyyyy Best regards, Dimitri. On 10/11/2017 04:47 PM, Phil Cummins wrote: > Hi Dimitri, > > Well try it and let yo know. Thanks! > > - Phil > > ------------------------------------------------------------------------ > *From:* Dimitri Komatitsch > *Sent:* Wednesday, October 11, 2017 10:12:15 PM > *To:* cig-seismo at geodynamics.org; Phil Cummins > *Subject:* Re: [CIG-SEISMO] specfem2d: snapshots of domain subset > > Hi Phil, > > It is not implemented, but in src/specfem2D/prepare_color_image.F90 if > you replace > >    ! horizontal min and max coordinates of the image >    xmin_color_image_loc = minval(coord(1,:)) >    xmax_color_image_loc = maxval(coord(1,:)) > >    ! vertical min and max coordinates of the image, slightly increase it > to go beyond maximum topography >    zmin_color_image_loc = minval(coord(2,:)) >    zmax_color_image_loc = maxval(coord(2,:)) > > with the min and max that you want it will probably work (please let us > know). > > Thanks, > Dimitri. > > On 10/11/2017 10:27 AM, Phil Cummins wrote: >> Hi, >> >> I'm using spefem2d to compute seismic waves n a sedimentary basin. >> Does anyone know how I can output snapshots of a subset of the >> computational domain only, not the whole thing? I want to view the >> seismic waves in a shallow basin that is barely visible in a snapshot of >> the whole computational domain. I.e., I want to zoom in on the basin. >> Many thanks! >> > > -- > Dimitri Komatitsch, CNRS Research Director (DR CNRS) > Laboratory of Mechanics and Acoustics, Marseille, France > http://komatitsch.free.fr -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr From komatitsch at lma.cnrs-mrs.fr Tue Oct 17 08:15:08 2017 From: komatitsch at lma.cnrs-mrs.fr (Dimitri Komatitsch) Date: Tue, 17 Oct 2017 17:15:08 +0200 Subject: [CIG-SEISMO] Fwd: Anisotropy Model in SPECFEM3D_GLOBE In-Reply-To: <2BD75E80-249F-4E30-92DB-891A2B270883@terpmail.umd.edu> References: <2BD75E80-249F-4E30-92DB-891A2B270883@terpmail.umd.edu> Message-ID: <85450815-3330-4c2c-3c3d-3ae5ff63aa8f@lma.cnrs-mrs.fr> Dear all, Would anybody know the answer to Quancheng Huang's question below? Thank you very much, Best regards, Dimitri. -------- Forwarded Message -------- Subject: Anisotropy Model in SPECFEM3D_GLOBE Date: Mon, 16 Oct 2017 15:41:54 -0400 From: Quancheng Huang To: komatitsch at lma.cnrs-mrs.fr Dear Prof. Komatitsch, I’m a PhD student from University of Maryland. Recently, I’m using SPECFEM3D_GLOBE to do synthetic modeling about anisotropic structures in upper mantle. When I switch the model to *3D_anisotropic* in Par_file, I will always run into a bug like this for creating meshes: /[compute-a20-7:69690] *** Process received signal ***/ /[compute-a20-7:69690] Signal: Floating point exception (8)/ /[compute-a20-7:69690] Signal code: Floating point divide-by-zero (3)/ /[compute-a20-7:69690] Failing at address: 0x4507c3/ / / /[compute-b18-12:42501] [ 0] /lib64/libpthread.so.0(+0xf7e0) [0x7f0c777d47e0]/ /[compute-b18-12:42501] [ 1] /lustre/qchuang/specfem3d_globe/bin/xmeshfem3D(get_vpvs_minmax_+0x1763) [0x4507c3]/ /[compute-b18-12:42501] [ 2] /lustre/qchuang/specfem3d_globe/bin/xmeshfem3D(check_mesh_resolution_+0x6c3) [0x451183]/ /[compute-b18-12:42501] [ 3] /lustre/qchuang/specfem3d_globe/bin/xmeshfem3D(create_regions_mesh_+0xbb8) [0x48a678]/ /[compute-b18-12:42501] [ 4] /lustre/qchuang/specfem3d_globe/bin/xmeshfem3D(create_meshes_+0x6b4) [0x46f304]/ /[compute-b18-12:42501] [ 5] /lustre/qchuang/specfem3d_globe/bin/xmeshfem3D(main+0x3e) [0x417fee]/ /[compute-b18-12:42501] [ 6] /lib64/libc.so.6(__libc_start_main+0xfd) [0x7f0c7744fd1d]/ /[compute-b18-12:42501] [ 7] /lustre/qchuang/specfem3d_globe/bin/xmeshfem3D() [0x41802d]/ /[compute-b18-12:42501] *** End of error message ***/ / / It is supposed to read in the Montagner (2002) model, which is a build-in model in the codes, when the model flag is switched to *3D_anisotropic. *I think the bug came from the *get_vpvs_minmax***subroutine but I can’t figure it out. I can run all the other tomography models except this 3D anisotropy model. Do you have any idea what the bug is? Best Regards, Quancheng Huang Quancheng Huang PhD Student Department of Geology University of Maryland College Park, MD 20742 Email: qchuang at umd.edu Tel: 202-848-8110 From maurya2278satish at gmail.com Tue Oct 17 10:32:21 2017 From: maurya2278satish at gmail.com (satish maurya) Date: Tue, 17 Oct 2017 10:32:21 -0700 Subject: [CIG-SEISMO] Mineos: sensor depth => edepth not working? In-Reply-To: References: Message-ID: Reminder! Any one... Dear all, I try to compute synthetic seismogram for receiver's place at depth 0, 50,100...300km, but seismogram is same. I put these depths in .sitechan file and elev is 0.0 in all case. I was wondering, if this parameter is implemented in the code or I am doing something wrong. Any suggestion! Regards Satish Maurya Post-Doc University of California, Berkeley Seismological Laboratory, Dept. of Earth & Planetary Science 201 The Tower Room, McCone Hall Berkeley, CA 94720 On Wed, Oct 4, 2017 at 12:06 PM, satish maurya wrote: > Dear all, > > I try to compute synthetic seismogram for receiver's place at depth 0, > 50,100...300km, but seismogram is same. > I put these depths in .sitechan file and elev is 0.0 in all case. > I was wondering, if this parameter is implemented in the code or I am > doing something wrong. > > Any suggestion! > > Regards > > > > Satish Maurya > Post-Doc > University of California, Berkeley > Seismological Laboratory, Dept. of Earth & Planetary Science > 201 The Tower Room, McCone Hall > Berkeley, CA 94720 > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andrew.valentine at anu.edu.au Tue Oct 17 15:16:43 2017 From: andrew.valentine at anu.edu.au (Andrew Valentine) Date: Wed, 18 Oct 2017 09:16:43 +1100 Subject: [CIG-SEISMO] Mineos: sensor depth => edepth not working? In-Reply-To: References: Message-ID: Not a definitive answer, as I don’t have a lot of experience with this aspect of Mineos, but… A quick look at the code (green.f & fdb/fdbs.f in mineos v1.0.2) suggests that although the station elevation and depth parameters are read in from the file (arrays elev_site and edepth_sitechan in common block /c_site/, defined in fdb/fdb_site.h), they are not actually used anywhere in green.f. I therefore suspect that this functionality is not implemented, and all seismograms are computed at the surface of the spherical body. All the best, Andrew > On 18 Oct 2017, at 04:32, satish maurya wrote: > > Reminder! > Any one... > > Dear all, > > I try to compute synthetic seismogram for receiver's place at depth 0, 50,100...300km, but seismogram is same. > I put these depths in .sitechan file and elev is 0.0 in all case. > I was wondering, if this parameter is implemented in the code or I am doing something wrong. > > Any suggestion! > > Regards > > Satish Maurya > Post-Doc > University of California, Berkeley > Seismological Laboratory, Dept. of Earth & Planetary Science > 201 The Tower Room, McCone Hall > Berkeley, CA 94720 > > On Wed, Oct 4, 2017 at 12:06 PM, satish maurya wrote: > Dear all, > > I try to compute synthetic seismogram for receiver's place at depth 0, 50,100...300km, but seismogram is same. > I put these depths in .sitechan file and elev is 0.0 in all case. > I was wondering, if this parameter is implemented in the code or I am doing something wrong. > > Any suggestion! > > Regards > > > > Satish Maurya > Post-Doc > University of California, Berkeley > Seismological Laboratory, Dept. of Earth & Planetary Science > 201 The Tower Room, McCone Hall > Berkeley, CA 94720 > > _______________________________________________ > CIG-SEISMO mailing list > CIG-SEISMO at geodynamics.org > http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo -- Dr Andrew Valentine Fellow, Seismology & Mathematical Geophysics Research School of Earth Sciences The Australian National University 142 Mills Road, Acton ACT 2601 +61 (0)2 612 53424 Office: Jaeger 2, 147a http://rses.anu.edu.au/~andrewv From seis_phd at yahoo.com Sun Oct 22 08:35:38 2017 From: seis_phd at yahoo.com (seisphd) Date: Sun, 22 Oct 2017 15:35:38 +0000 (UTC) Subject: [CIG-SEISMO] specfem2d velocity gradient implementation References: <866375629.1980443.1508686538571.ref@mail.yahoo.com> Message-ID: <866375629.1980443.1508686538571@mail.yahoo.com> Dear developers and users Downloading and starting using specfem2D I appreciated the fact that its free software and was amazed by its potentials. I have a question though, which I can not solve through the manual. Using an "internal velocity model" format;( like "simple_topography_and_also_a_simple_fluid_layer/interfaces_simple_topo_curved.dat"). Which parameters needs to be modified  into the Par_file to introduce a vertical velocity gradient? So for example the nodes of the first layer instead being all 3000.d0will start 2900 at the top and increase gradually with  depth to end with 3100 at the bottom. Thank you in advance for your help and time -------------- next part -------------- An HTML attachment was scrubbed... URL: From seis_phd at yahoo.com Sun Oct 22 08:10:19 2017 From: seis_phd at yahoo.com (seisphd) Date: Sun, 22 Oct 2017 15:10:19 +0000 (UTC) Subject: [CIG-SEISMO] velocity gradient specfem2D References: <73468088.1962913.1508685019149.ref@mail.yahoo.com> Message-ID: <73468088.1962913.1508685019149@mail.yahoo.com> Dear developers and users Downloading and starting using specfem2D I appreciated the fact that its free software and was amazed by its potentials. I have a question though, which I can not solve through the manual. Using an "internal velocity model" format;( like "simple_topography_and_also_a_simple_fluid_layer/interfaces_simple_topo_curved.dat"). Which parameters needs to be modified  into the Par_file to introduce a vertical velocity gradient? So for example the nodes of the first layer instead being all 3000.d0will start 2900 at the top and increase gradually with  depth to end with 3100 at the bottom. Thank you in advance for your help and time -------------- next part -------------- An HTML attachment was scrubbed... URL: From komatitsch at lma.cnrs-mrs.fr Sun Oct 22 08:47:30 2017 From: komatitsch at lma.cnrs-mrs.fr (Dimitri Komatitsch) Date: Sun, 22 Oct 2017 17:47:30 +0200 Subject: [CIG-SEISMO] specfem2d velocity gradient implementation In-Reply-To: <866375629.1980443.1508686538571@mail.yahoo.com> References: <866375629.1980443.1508686538571.ref@mail.yahoo.com> <866375629.1980443.1508686538571@mail.yahoo.com> Message-ID: Hi, You need to use the "define_external_model()" subroutine. Best regards, Dimitri. On 10/22/2017 05:35 PM, seisphd wrote: > Dear developers and users > > Downloading and starting using specfem2D I appreciated the fact that its > free software and was amazed by its potentials. > > I have a question though, which I can not solve through the manual. > > Using an "internal velocity model" format; > ( like > "simple_topography_and_also_a_simple_fluid_layer/interfaces_simple_topo_curved.dat"). > Which parameters needs to be modified  into the *Par_file* to introduce > a *vertical velocity gradient*? > > So for example the nodes of the first layer instead being all > 3000.d0 > will start 2900 at the top and increase gradually with  depth to end > with 3100 at the bottom. > > Thank you in advance for your help and time > > > _______________________________________________ > CIG-SEISMO mailing list > CIG-SEISMO at geodynamics.org > http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo > -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr From jchodacki at gig.eu Wed Oct 25 02:05:30 2017 From: jchodacki at gig.eu (Jacek Chodacki) Date: Wed, 25 Oct 2017 11:05:30 +0200 Subject: [CIG-SEISMO] SPECFEM3D - help Message-ID: <59F06FFA020000F40002D883@poczta_int.gw.local> Dear Madame/Sir, I am geophysicist and for some time I have been trying to use SPECFEM3D for predicting surface vibrations from seismic events (local seismicity connected with coal mine operation). Because the results are not satisfying I stepped back and calculated an example from SPECFEM3D (from directory ‘meshfem3d_examples -> simple model’). In output files I received for example pick value of displacement=0.0008m, pick value of velocity=0.0001m/s and pick value of acceleration=0.00005m/s^2. Is there any mistake in calculation or there is something I don’t know about (units of results in output files are [m] for displacement, [m/s] for velocity and [m/s^2] for acceleration)? Thank you in advance for help. Best regards Jacek Chodacki Central Mining Institute, Katowice, Poland -------------- next part -------------- An HTML attachment was scrubbed... URL: From carene at lanl.gov Wed Oct 25 15:59:26 2017 From: carene at lanl.gov (Larmat, Carene) Date: Wed, 25 Oct 2017 22:59:26 +0000 Subject: [CIG-SEISMO] Seismograms = zero with cuda version Message-ID: Hello, I want to share this observation: I am running specfem3d cartesian version on a CPU-GPU cluster. I download a recent update and use the master branch which is at commit b1822f93e6faafae4c204 (Jun 14 2017). I use the cuda library: CUDA Version 8.0.61. At that commit, the seismograms produced are all at 0.0. I went back to the commit 8b09ba951d2f and all is well. Let me know if you need more information. Cheers! Carene Larmat, EES-17, MS D452 carene at lanl.gov, 505 667 2074 -------------- next part -------------- An HTML attachment was scrubbed... URL: From jerome.azzola at unistra.fr Thu Oct 26 04:59:54 2017 From: jerome.azzola at unistra.fr (=?utf-8?q?AZZOLA_Jerome_=28JDS=29?=) Date: Thu, 26 Oct 2017 13:59:54 +0200 Subject: [CIG-SEISMO] Problems in seismograms - Specfem3D Message-ID: Hello, First, I would like to thank CIG Seismology for providing Specfem3D Cartesian and other very useful Software with free access. I was working with Specfem2D to simulate propagation of waves in a plate with holes that I mesh using the external mesher GMSH. I would like now to model the same problem using Specfem3D. For my convenience, I have to use GMSH to mesh the propagation medium, which is still a perforated plate. I join my data files and an output. I would like to propose you the link above to have a closer look to the geometry and the mesh files I use also: https://drive.google.com/open?id=0BzMCuLE_smWxR001V1lZbFM4Y2s The problem is that the seismograms that I obtain show wiggles that I relate to computational instability. If I use a higher sampling rate, the hysteresis is even increased. I don't understand the origin of the problem. It seems that the mesh has no influence because the increase of the number of elements is not a solution. I have the same problem when I simulate the propagation in a plate without holes. I checked the stability coefficient for wave velocity in the output files and the skewness of my elements. Do you have any idea about the origin of this instability ? Cordially, -- Jérôme AZZOLA Doctorant en géosciences Tél : +33 (0)7 68 54 92 57 azzola at unistra.fr jerome.azzola at gmail.com EOST - Ecole et Observatoire des Sciences de la Terre 5, rue René Descartes F - 67000 Strasbourg -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: STATIONS Type: application/octet-stream Size: 169 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: FORCESOLUTION Type: application/octet-stream Size: 277 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Par_file Type: application/octet-stream Size: 15281 bytes Desc: not available URL: -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: output_solver.txt URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: sismos_xx.pdf Type: application/pdf Size: 257895 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: sismos_xy.pdf Type: application/pdf Size: 278615 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: sismos_xz.pdf Type: application/pdf Size: 248215 bytes Desc: not available URL: From komatitsch at lma.cnrs-mrs.fr Thu Oct 26 16:01:09 2017 From: komatitsch at lma.cnrs-mrs.fr (Dimitri Komatitsch) Date: Fri, 27 Oct 2017 01:01:09 +0200 Subject: [CIG-SEISMO] SPECFEM3D - help In-Reply-To: <59F06FFA020000F40002D883@poczta_int.gw.local> References: <59F06FFA020000F40002D883@poczta_int.gw.local> Message-ID: <02501e07-6217-576d-6f4d-3295a3c4a3ff@lma.cnrs-mrs.fr> Dear Jacek, In many of the examples, in particular those that use FORCESOLUTION rather than CMTSOLUTION to describe the source, what is used is relative amplitudes (comparing amplitude variations between seismograms recorded at different locations). If you need absolute amplitudes, CMTSOLUTION is safe but I am not sure if force sources (i.e. FORCESOLUTION) are OK, see https://github.com/geodynamics/specfem3d/issues/842 . Best regards, Dimitri. On 10/25/2017 11:05 AM, Jacek Chodacki wrote: > Dear Madame/Sir, > > I am geophysicist and for some time I have been trying to use SPECFEM3D > for predicting surface vibrations from seismic events (local seismicity > connected with coal mine operation). Because the results are not > satisfying I stepped back and calculated an example from SPECFEM3D (from > directory ‘meshfem3d_examples -> simple model’). In output files I > received for example pick value of displacement=0.0008m, pick value of > velocity=0.0001m/s and pick value of acceleration=0.00005m/s^2. Is there > any mistake in calculation or there is something I don’t know about > (units of results in output files are [m] for displacement, [m/s] for > velocity and [m/s^2] for acceleration)? > > Thank you in advance for help. > > Best regards > > Jacek Chodacki > > Central Mining Institute, Katowice, Poland > > > > _______________________________________________ > CIG-SEISMO mailing list > CIG-SEISMO at geodynamics.org > http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo > -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr From surendra at iith.ac.in Thu Oct 26 11:25:35 2017 From: surendra at iith.ac.in (Surendra Nadh Somala) Date: Thu, 26 Oct 2017 23:55:35 +0530 Subject: [CIG-SEISMO] Problems in seismograms - Specfem3D In-Reply-To: References: Message-ID: Dear Jérôme, I recommend that you use CUBIT/Python interface to create mesh files that are usable by scotch (the graphical partitioner), unlike gmsh. Best, Surendra ___________ *Surendra Nadh Somala* Assistant Professor Department of Civil Engineering (Block-E, 208) Indian Institute of Technology (IIT) Hyderabad Kandi, Sangareddy Telangana, India - 502285 Phone : +91 (0)40 2301-8457 http://civil.iith.ac.in/surendra/ On Thu, Oct 26, 2017 at 5:29 PM, AZZOLA Jerome (JDS) < jerome.azzola at unistra.fr> wrote: > Hello, > > First, I would like to thank CIG Seismology for providing Specfem3D > Cartesian and other very useful Software with free access. > I was working with Specfem2D to simulate propagation of waves in a plate > with holes that I mesh using the external mesher GMSH. I would like now to > model the same problem using Specfem3D. For my convenience, I have to use > GMSH to mesh the propagation medium, which is still a perforated plate. I > join my data files and an output. I would like to propose you the link > above to have a closer look to the geometry and the mesh files I use also: > https://drive.google.com/open?id=0BzMCuLE_smWxR001V1lZbFM4Y2s > The problem is that the seismograms that I obtain show wiggles that I > relate to computational instability. If I use a higher sampling rate, the > hysteresis is even increased. I don't understand the origin of the problem. > It seems that the mesh has no influence because the increase of the number > of elements is not a solution. I have the same problem when I simulate the > propagation in a plate without holes. I checked the stability coefficient > for wave velocity in the output files and the skewness of my elements. Do > you have any idea about the origin of this instability ? > Cordially, > > -- > > Jérôme AZZOLA > Doctorant en géosciences > Tél : +33 (0)7 68 54 92 57 > azzola at unistra.fr > jerome.azzola at gmail.com > [image: www.unistra.fr] > EOST - Ecole et Observatoire des Sciences de la Terre > 5, rue René Descartes > F - 67000 Strasbourg > > _______________________________________________ > CIG-SEISMO mailing list > CIG-SEISMO at geodynamics.org > http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo > -------------- next part -------------- An HTML attachment was scrubbed... URL: From dylanmikesell at boisestate.edu Fri Oct 27 12:17:43 2017 From: dylanmikesell at boisestate.edu (Dylan Mikesell) Date: Fri, 27 Oct 2017 13:17:43 -0600 Subject: [CIG-SEISMO] SPECFEM3D Message-ID: Hi SPECFEM developers, Is anyone currently working on the PML implementation for the adjoint. A student and I are working on an adjoint tomography and planning to use SPECFEM3D. We tried to save the forward simulations today and got the following. STOP PML_CONDITIONS is still under test for adjoint simulation I checked the devel branch and this is still there. Is anyone actively working on this topic? It would be nice to get the PML going with the adjoint. Best wishes, Dylan -------------- next part -------------- An HTML attachment was scrubbed... URL: From jas11 at Princeton.EDU Mon Oct 30 10:44:56 2017 From: jas11 at Princeton.EDU (James A. Smith) Date: Mon, 30 Oct 2017 17:44:56 +0000 Subject: [CIG-SEISMO] SPECFEM3D In-Reply-To: References: Message-ID: Hi Dylan, I have tested PML for the adjoint simulation. It seems to work, although I am still playing with getting higher quality inversion results (as compared to Stacey). Please see this issue: https://github.com/geodynamics/specfem3d/issues/312 You will need to enable UNDO_ATTENUATION for PML to work in this case (in setup/constanst.h). You will also have to comment out the lines in the source code that stop the code if PML is turned on for an adjoint simulation and recompile. I hope it works for you. Best, James ________________________________ From: CIG-SEISMO [cig-seismo-bounces at geodynamics.org] on behalf of Dylan Mikesell [dylanmikesell at boisestate.edu] Sent: Friday, October 27, 2017 2:17 PM To: cig-seismo at geodynamics.org Subject: [CIG-SEISMO] SPECFEM3D Hi SPECFEM developers, Is anyone currently working on the PML implementation for the adjoint. A student and I are working on an adjoint tomography and planning to use SPECFEM3D. We tried to save the forward simulations today and got the following. STOP PML_CONDITIONS is still under test for adjoint simulation I checked the devel branch and this is still there. Is anyone actively working on this topic? It would be nice to get the PML going with the adjoint. Best wishes, Dylan -------------- next part -------------- An HTML attachment was scrubbed... URL: From komatitsch at lma.cnrs-mrs.fr Mon Oct 30 12:59:25 2017 From: komatitsch at lma.cnrs-mrs.fr (Dimitri Komatitsch) Date: Mon, 30 Oct 2017 20:59:25 +0100 Subject: [CIG-SEISMO] SPECFEM3D In-Reply-To: References: Message-ID: <66ed3ae0-9174-f0e6-b71a-70d866925fb5@lma.cnrs-mrs.fr> Hi Dylan, Yes, just use UNDO_ATTENUATION, which can undo any dissipative medium exactly, including PML. It is described in http://komatitsch.free.fr/preprints/GJI_undo_attenuation_2016.pdf . We do that routinely here for FWI with PML, works beautifully. Please cite that paper if you use it. Thanks, Best, Dimitri. On 10/30/2017 06:44 PM, James A. Smith wrote: > Hi Dylan, > > I have tested PML for the adjoint simulation. It seems to work, although > I am still playing with getting higher quality inversion results (as > compared to Stacey). Please see this issue: > https://github.com/geodynamics/specfem3d/issues/312 > > You will need to enable UNDO_ATTENUATION for PML to work in this case > (in setup/constanst.h). You will also have to comment out the lines in > the source code that stop the code if PML is turned on for an adjoint > simulation and recompile. I hope it works for you. > > Best, > James > > ------------------------------------------------------------------------ > *From:* CIG-SEISMO [cig-seismo-bounces at geodynamics.org] on behalf of > Dylan Mikesell [dylanmikesell at boisestate.edu] > *Sent:* Friday, October 27, 2017 2:17 PM > *To:* cig-seismo at geodynamics.org > *Subject:* [CIG-SEISMO] SPECFEM3D > > Hi SPECFEM developers, > > Is anyone currently working on the PML implementation for the adjoint. A > student and I are working on an adjoint tomography and planning to use > SPECFEM3D. We tried to save the forward simulations today and got the > following. > > STOP PML_CONDITIONS is still under test for adjoint simulation > > I checked the devel branch and this is still there. Is anyone actively > working on this topic? It would be nice to get the PML going with the > adjoint. > > Best wishes, > > Dylan > > > _______________________________________________ > CIG-SEISMO mailing list > CIG-SEISMO at geodynamics.org > http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo > -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr From komatitsch at lma.cnrs-mrs.fr Mon Oct 30 13:07:34 2017 From: komatitsch at lma.cnrs-mrs.fr (Dimitri Komatitsch) Date: Mon, 30 Oct 2017 21:07:34 +0100 Subject: [CIG-SEISMO] SPECFEM3D / PS In-Reply-To: <66ed3ae0-9174-f0e6-b71a-70d866925fb5@lma.cnrs-mrs.fr> References: <66ed3ae0-9174-f0e6-b71a-70d866925fb5@lma.cnrs-mrs.fr> Message-ID: <431b155c-5782-a9b5-1f31-fe8e4b8f921a@lma.cnrs-mrs.fr> Hi all, PS: we could thus consider renaming the UNDO_ATTENUATION flag to UNDO_ATTENUATION_AND_OR_PML. I know people do not like when we rename flags because Par_files then need to be updated, but this one has become a bit confusing. Best, Dimitri. On 10/30/2017 08:59 PM, Dimitri Komatitsch wrote: > > Hi Dylan, > > Yes, just use UNDO_ATTENUATION, which can undo any dissipative medium > exactly, including PML. It is described in > http://komatitsch.free.fr/preprints/GJI_undo_attenuation_2016.pdf . > We do that routinely here for FWI with PML, works beautifully. > Please cite that paper if you use it. > > Thanks, > Best, > Dimitri. > > On 10/30/2017 06:44 PM, James A. Smith wrote: >> Hi Dylan, >> >> I have tested PML for the adjoint simulation. It seems to work, >> although I am still playing with getting higher quality inversion >> results (as compared to Stacey). Please see this issue: >> https://github.com/geodynamics/specfem3d/issues/312 >> >> You will need to enable UNDO_ATTENUATION for PML to work in this case >> (in setup/constanst.h). You will also have to comment out the lines in >> the source code that stop the code if PML is turned on for an adjoint >> simulation and recompile. I hope it works for you. >> >> Best, >> James >> >> ------------------------------------------------------------------------ >> *From:* CIG-SEISMO [cig-seismo-bounces at geodynamics.org] on behalf of >> Dylan Mikesell [dylanmikesell at boisestate.edu] >> *Sent:* Friday, October 27, 2017 2:17 PM >> *To:* cig-seismo at geodynamics.org >> *Subject:* [CIG-SEISMO] SPECFEM3D >> >> Hi SPECFEM developers, >> >> Is anyone currently working on the PML implementation for the adjoint. >> A student and I are working on an adjoint tomography and planning to >> use SPECFEM3D. We tried to save the forward simulations today and got >> the following. >> >> STOP PML_CONDITIONS is still under test for adjoint simulation >> >> I checked the devel branch and this is still there. Is anyone actively >> working on this topic? It would be nice to get the PML going with the >> adjoint. >> >> Best wishes, >> >> Dylan >> >> >> _______________________________________________ >> CIG-SEISMO mailing list >> CIG-SEISMO at geodynamics.org >> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo >> > -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr From dmikesell at cgiss.boisestate.edu Mon Oct 30 13:01:18 2017 From: dmikesell at cgiss.boisestate.edu (Dylan Mikesell) Date: Mon, 30 Oct 2017 14:01:18 -0600 Subject: [CIG-SEISMO] SPECFEM3D In-Reply-To: <66ed3ae0-9174-f0e6-b71a-70d866925fb5@lma.cnrs-mrs.fr> References: <66ed3ae0-9174-f0e6-b71a-70d866925fb5@lma.cnrs-mrs.fr> Message-ID: Hi Dimitri and James, Thank you for the help! We will give it a try. Best wishes, Dylan On Mon, Oct 30, 2017 at 1:59 PM, Dimitri Komatitsch < komatitsch at lma.cnrs-mrs.fr> wrote: > > Hi Dylan, > > Yes, just use UNDO_ATTENUATION, which can undo any dissipative medium > exactly, including PML. It is described in http://komatitsch.free.fr/prep > rints/GJI_undo_attenuation_2016.pdf . > We do that routinely here for FWI with PML, works beautifully. > Please cite that paper if you use it. > > Thanks, > Best, > Dimitri. > > On 10/30/2017 06:44 PM, James A. Smith wrote: > >> Hi Dylan, >> >> I have tested PML for the adjoint simulation. It seems to work, although >> I am still playing with getting higher quality inversion results (as >> compared to Stacey). Please see this issue: >> https://github.com/geodynamics/specfem3d/issues/312 >> >> You will need to enable UNDO_ATTENUATION for PML to work in this case (in >> setup/constanst.h). You will also have to comment out the lines in the >> source code that stop the code if PML is turned on for an adjoint >> simulation and recompile. I hope it works for you. >> >> Best, >> James >> >> ------------------------------------------------------------------------ >> *From:* CIG-SEISMO [cig-seismo-bounces at geodynamics.org] on behalf of >> Dylan Mikesell [dylanmikesell at boisestate.edu] >> *Sent:* Friday, October 27, 2017 2:17 PM >> *To:* cig-seismo at geodynamics.org >> *Subject:* [CIG-SEISMO] SPECFEM3D >> >> Hi SPECFEM developers, >> >> Is anyone currently working on the PML implementation for the adjoint. A >> student and I are working on an adjoint tomography and planning to use >> SPECFEM3D. We tried to save the forward simulations today and got the >> following. >> >> STOP PML_CONDITIONS is still under test for adjoint simulation >> >> I checked the devel branch and this is still there. Is anyone actively >> working on this topic? It would be nice to get the PML going with the >> adjoint. >> >> Best wishes, >> >> Dylan >> >> >> _______________________________________________ >> CIG-SEISMO mailing list >> CIG-SEISMO at geodynamics.org >> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo >> >> > -- > Dimitri Komatitsch, CNRS Research Director (DR CNRS) > Laboratory of Mechanics and Acoustics, Marseille, France > http://komatitsch.free.fr > > _______________________________________________ > CIG-SEISMO mailing list > CIG-SEISMO at geodynamics.org > http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jas11 at Princeton.EDU Mon Oct 30 13:14:42 2017 From: jas11 at Princeton.EDU (James A. Smith) Date: Mon, 30 Oct 2017 20:14:42 +0000 Subject: [CIG-SEISMO] SPECFEM3D / PS In-Reply-To: <431b155c-5782-a9b5-1f31-fe8e4b8f921a@lma.cnrs-mrs.fr> References: <66ed3ae0-9174-f0e6-b71a-70d866925fb5@lma.cnrs-mrs.fr>, <431b155c-5782-a9b5-1f31-fe8e4b8f921a@lma.cnrs-mrs.fr> Message-ID: Hi Dimitri, Shouldn't we also remove the code that stops the simulation if you attempt to run with PML for adjoint run? Also, the UNDO_ATTENUATION flag is in setup/constants.h. Does it make more sense for it to be in Par_file if users will change it occasionally? Best, James ________________________________________ From: Dimitri Komatitsch [komatitsch at lma.cnrs-mrs.fr] Sent: Monday, October 30, 2017 3:07 PM To: cig-seismo at geodynamics.org; James A. Smith; Dylan Mikesell Cc: Vadim Monteiller Subject: Re: [CIG-SEISMO] SPECFEM3D / PS Hi all, PS: we could thus consider renaming the UNDO_ATTENUATION flag to UNDO_ATTENUATION_AND_OR_PML. I know people do not like when we rename flags because Par_files then need to be updated, but this one has become a bit confusing. Best, Dimitri. On 10/30/2017 08:59 PM, Dimitri Komatitsch wrote: > > Hi Dylan, > > Yes, just use UNDO_ATTENUATION, which can undo any dissipative medium > exactly, including PML. It is described in > http://komatitsch.free.fr/preprints/GJI_undo_attenuation_2016.pdf . > We do that routinely here for FWI with PML, works beautifully. > Please cite that paper if you use it. > > Thanks, > Best, > Dimitri. > > On 10/30/2017 06:44 PM, James A. Smith wrote: >> Hi Dylan, >> >> I have tested PML for the adjoint simulation. It seems to work, >> although I am still playing with getting higher quality inversion >> results (as compared to Stacey). Please see this issue: >> https://github.com/geodynamics/specfem3d/issues/312 >> >> You will need to enable UNDO_ATTENUATION for PML to work in this case >> (in setup/constanst.h). You will also have to comment out the lines in >> the source code that stop the code if PML is turned on for an adjoint >> simulation and recompile. I hope it works for you. >> >> Best, >> James >> >> ------------------------------------------------------------------------ >> *From:* CIG-SEISMO [cig-seismo-bounces at geodynamics.org] on behalf of >> Dylan Mikesell [dylanmikesell at boisestate.edu] >> *Sent:* Friday, October 27, 2017 2:17 PM >> *To:* cig-seismo at geodynamics.org >> *Subject:* [CIG-SEISMO] SPECFEM3D >> >> Hi SPECFEM developers, >> >> Is anyone currently working on the PML implementation for the adjoint. >> A student and I are working on an adjoint tomography and planning to >> use SPECFEM3D. We tried to save the forward simulations today and got >> the following. >> >> STOP PML_CONDITIONS is still under test for adjoint simulation >> >> I checked the devel branch and this is still there. Is anyone actively >> working on this topic? It would be nice to get the PML going with the >> adjoint. >> >> Best wishes, >> >> Dylan >> >> >> _______________________________________________ >> CIG-SEISMO mailing list >> CIG-SEISMO at geodynamics.org >> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo >> > -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr From komatitsch at lma.cnrs-mrs.fr Mon Oct 30 13:25:27 2017 From: komatitsch at lma.cnrs-mrs.fr (Dimitri Komatitsch) Date: Mon, 30 Oct 2017 21:25:27 +0100 Subject: [CIG-SEISMO] SPECFEM3D / PS In-Reply-To: References: <66ed3ae0-9174-f0e6-b71a-70d866925fb5@lma.cnrs-mrs.fr> <431b155c-5782-a9b5-1f31-fe8e4b8f921a@lma.cnrs-mrs.fr> Message-ID: <59c722db-91b6-72e6-716e-3c277fa24da6@lma.cnrs-mrs.fr> Hi James, Hi all, Very good idea. Tomorrow I will rename it and will move it to the Par_file, and will also remove the stop statement (more precisely, I will convert it to a stop statement in case UNDO_ATT is not set but PML is set). Vadim is cleaning his local implementation of UNDO_ATT, we will soon commit it to the development branch (see https://github.com/geodynamics/specfem3d/issues/14 ). For now only SPECFEM3D_GLOBE and SPECFEM2D have it, regarding SPECFEM3D (Cartesian) we are almost done. Best, Dimitri. On 10/30/2017 09:14 PM, James A. Smith wrote: > Hi Dimitri, > > Shouldn't we also remove the code that stops the simulation if you attempt to run with PML for adjoint run? > > Also, the UNDO_ATTENUATION flag is in setup/constants.h. Does it make more sense for it to be in Par_file if users will change it occasionally? > > Best, > James > > ________________________________________ > From: Dimitri Komatitsch [komatitsch at lma.cnrs-mrs.fr] > Sent: Monday, October 30, 2017 3:07 PM > To: cig-seismo at geodynamics.org; James A. Smith; Dylan Mikesell > Cc: Vadim Monteiller > Subject: Re: [CIG-SEISMO] SPECFEM3D / PS > > Hi all, > > PS: we could thus consider renaming the UNDO_ATTENUATION flag to > UNDO_ATTENUATION_AND_OR_PML. I know people do not like when we rename > flags because Par_files then need to be updated, but this one has become > a bit confusing. > > Best, > Dimitri. > > On 10/30/2017 08:59 PM, Dimitri Komatitsch wrote: >> >> Hi Dylan, >> >> Yes, just use UNDO_ATTENUATION, which can undo any dissipative medium >> exactly, including PML. It is described in >> http://komatitsch.free.fr/preprints/GJI_undo_attenuation_2016.pdf . >> We do that routinely here for FWI with PML, works beautifully. >> Please cite that paper if you use it. >> >> Thanks, >> Best, >> Dimitri. >> >> On 10/30/2017 06:44 PM, James A. Smith wrote: >>> Hi Dylan, >>> >>> I have tested PML for the adjoint simulation. It seems to work, >>> although I am still playing with getting higher quality inversion >>> results (as compared to Stacey). Please see this issue: >>> https://github.com/geodynamics/specfem3d/issues/312 >>> >>> You will need to enable UNDO_ATTENUATION for PML to work in this case >>> (in setup/constanst.h). You will also have to comment out the lines in >>> the source code that stop the code if PML is turned on for an adjoint >>> simulation and recompile. I hope it works for you. >>> >>> Best, >>> James >>> >>> ------------------------------------------------------------------------ >>> *From:* CIG-SEISMO [cig-seismo-bounces at geodynamics.org] on behalf of >>> Dylan Mikesell [dylanmikesell at boisestate.edu] >>> *Sent:* Friday, October 27, 2017 2:17 PM >>> *To:* cig-seismo at geodynamics.org >>> *Subject:* [CIG-SEISMO] SPECFEM3D >>> >>> Hi SPECFEM developers, >>> >>> Is anyone currently working on the PML implementation for the adjoint. >>> A student and I are working on an adjoint tomography and planning to >>> use SPECFEM3D. We tried to save the forward simulations today and got >>> the following. >>> >>> STOP PML_CONDITIONS is still under test for adjoint simulation >>> >>> I checked the devel branch and this is still there. Is anyone actively >>> working on this topic? It would be nice to get the PML going with the >>> adjoint. >>> >>> Best wishes, >>> >>> Dylan >>> >>> >>> _______________________________________________ >>> CIG-SEISMO mailing list >>> CIG-SEISMO at geodynamics.org >>> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo >>> >> > > -- > Dimitri Komatitsch, CNRS Research Director (DR CNRS) > Laboratory of Mechanics and Acoustics, Marseille, France > http://komatitsch.free.fr > -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr From komatitsch at lma.cnrs-mrs.fr Tue Oct 31 14:17:25 2017 From: komatitsch at lma.cnrs-mrs.fr (Dimitri Komatitsch) Date: Tue, 31 Oct 2017 22:17:25 +0100 Subject: [CIG-SEISMO] SPECFEM3D / PS In-Reply-To: <59c722db-91b6-72e6-716e-3c277fa24da6@lma.cnrs-mrs.fr> References: <66ed3ae0-9174-f0e6-b71a-70d866925fb5@lma.cnrs-mrs.fr> <431b155c-5782-a9b5-1f31-fe8e4b8f921a@lma.cnrs-mrs.fr> <59c722db-91b6-72e6-716e-3c277fa24da6@lma.cnrs-mrs.fr> Message-ID: <3e15a521-1a67-18a1-764b-64c4f7ed6707@lma.cnrs-mrs.fr> Hi all, Done! (in 2D and in 3D_Cartesian; in 3D_GLOBE we do not have PMLs and thus I did not rename the flag). Best regards, Dimitri. On 10/30/2017 09:25 PM, Dimitri Komatitsch wrote: > > Hi James, Hi all, > > Very good idea. Tomorrow I will rename it and will move it to the > Par_file, and will also remove the stop statement (more precisely, I > will convert it to a stop statement in case UNDO_ATT is not set but PML > is set). > > Vadim is cleaning his local implementation of UNDO_ATT, we will soon > commit it to the development branch (see > https://github.com/geodynamics/specfem3d/issues/14 ). For now only > SPECFEM3D_GLOBE and SPECFEM2D have it, regarding SPECFEM3D (Cartesian) > we are almost done. > > Best, > Dimitri. > > On 10/30/2017 09:14 PM, James A. Smith wrote: >> Hi Dimitri, >> >> Shouldn't we also remove the code that stops the simulation if you >> attempt to run with PML for adjoint run? >> >> Also, the UNDO_ATTENUATION flag is in setup/constants.h. Does it make >> more sense for it to be in Par_file if users will change it occasionally? >> >> Best, >> James >> >> ________________________________________ >> From: Dimitri Komatitsch [komatitsch at lma.cnrs-mrs.fr] >> Sent: Monday, October 30, 2017 3:07 PM >> To: cig-seismo at geodynamics.org; James A. Smith; Dylan Mikesell >> Cc: Vadim Monteiller >> Subject: Re: [CIG-SEISMO] SPECFEM3D / PS >> >> Hi all, >> >> PS: we could thus consider renaming the UNDO_ATTENUATION flag to >> UNDO_ATTENUATION_AND_OR_PML. I know people do not like when we rename >> flags because Par_files then need to be updated, but this one has become >> a bit confusing. >> >> Best, >> Dimitri. >> >> On 10/30/2017 08:59 PM, Dimitri Komatitsch wrote: >>> >>> Hi Dylan, >>> >>> Yes, just use UNDO_ATTENUATION, which can undo any dissipative medium >>> exactly, including PML. It is described in >>> http://komatitsch.free.fr/preprints/GJI_undo_attenuation_2016.pdf . >>> We do that routinely here for FWI with PML, works beautifully. >>> Please cite that paper if you use it. >>> >>> Thanks, >>> Best, >>> Dimitri. >>> >>> On 10/30/2017 06:44 PM, James A. Smith wrote: >>>> Hi Dylan, >>>> >>>> I have tested PML for the adjoint simulation. It seems to work, >>>> although I am still playing with getting higher quality inversion >>>> results (as compared to Stacey). Please see this issue: >>>> https://github.com/geodynamics/specfem3d/issues/312 >>>> >>>> You will need to enable UNDO_ATTENUATION for PML to work in this case >>>> (in setup/constanst.h). You will also have to comment out the lines in >>>> the source code that stop the code if PML is turned on for an adjoint >>>> simulation and recompile. I hope it works for you. >>>> >>>> Best, >>>> James >>>> >>>> ------------------------------------------------------------------------ >>>> >>>> *From:* CIG-SEISMO [cig-seismo-bounces at geodynamics.org] on behalf of >>>> Dylan Mikesell [dylanmikesell at boisestate.edu] >>>> *Sent:* Friday, October 27, 2017 2:17 PM >>>> *To:* cig-seismo at geodynamics.org >>>> *Subject:* [CIG-SEISMO] SPECFEM3D >>>> >>>> Hi SPECFEM developers, >>>> >>>> Is anyone currently working on the PML implementation for the adjoint. >>>> A student and I are working on an adjoint tomography and planning to >>>> use SPECFEM3D. We tried to save the forward simulations today and got >>>> the following. >>>> >>>> STOP PML_CONDITIONS is still under test for adjoint simulation >>>> >>>> I checked the devel branch and this is still there. Is anyone actively >>>> working on this topic? It would be nice to get the PML going with the >>>> adjoint. >>>> >>>> Best wishes, >>>> >>>> Dylan >>>> >>>> >>>> _______________________________________________ >>>> CIG-SEISMO mailing list >>>> CIG-SEISMO at geodynamics.org >>>> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo >>>> >>> >> >> -- >> Dimitri Komatitsch, CNRS Research Director (DR CNRS) >> Laboratory of Mechanics and Acoustics, Marseille, France >> http://komatitsch.free.fr >> > -- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr