[cig-commits] r16302 - in seismo/3D: SPECFEM3D_GLOBE/trunk SPECFEM3D_GLOBE/trunk/USER_MANUAL SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures SPECFEM3D_SESAME/trunk SPECFEM3D_SESAME/trunk/USER_MANUAL SPECFEM3D_SESAME/trunk/USER_MANUAL/figures
dkomati1 at geodynamics.org
dkomati1 at geodynamics.org
Sun Feb 21 16:44:16 PST 2010
Author: dkomati1
Date: 2010-02-21 16:44:15 -0800 (Sun, 21 Feb 2010)
New Revision: 16302
Modified:
seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/bibliography.bib
seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/PKPdf_all_15s500s.pdf
seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/bolivia_trans.pdf
seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/bolivia_vertical.pdf
seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/vanuatu_trans.pdf
seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/vanuatu_vertical.pdf
seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/manual_SPECFEM3D_GLOBE.pdf
seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/manual_SPECFEM3D_GLOBE.tex
seismo/3D/SPECFEM3D_GLOBE/trunk/todo_list_please_dont_remove.txt
seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/bibliography.bib
seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/figures/gauss_vs_triangle_mod.pdf
seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/figures/source_timing.pdf
seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/manual_SPECFEM3D_SESAME.pdf
seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/manual_SPECFEM3D_SESAME.tex
seismo/3D/SPECFEM3D_SESAME/trunk/todo_list_please_dont_remove.txt
Log:
updated the manuals
Modified: seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/bibliography.bib
===================================================================
--- seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/bibliography.bib 2010-02-21 22:02:11 UTC (rev 16301)
+++ seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/bibliography.bib 2010-02-22 00:44:15 UTC (rev 16302)
@@ -10649,7 +10649,7 @@
}
@INPROCEEDINGS{SavWiTr05,
- author = {Savage, B. and D.A. Wiens and J. Tromp},
+ author = {Savage, B. and D. A. Wiens and J. Tromp},
title = {Effects of 3{D} Velocity and Attenuation in the {T}onga-{F}iji Subduction
Zone},
booktitle = {Expanded abstracts of the American Geophysical Union},
Modified: seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/PKPdf_all_15s500s.pdf
===================================================================
(Binary files differ)
Modified: seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/bolivia_trans.pdf
===================================================================
(Binary files differ)
Modified: seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/bolivia_vertical.pdf
===================================================================
(Binary files differ)
Modified: seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/vanuatu_trans.pdf
===================================================================
(Binary files differ)
Modified: seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/figures/vanuatu_vertical.pdf
===================================================================
(Binary files differ)
Modified: seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/manual_SPECFEM3D_GLOBE.pdf
===================================================================
(Binary files differ)
Modified: seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/manual_SPECFEM3D_GLOBE.tex
===================================================================
--- seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/manual_SPECFEM3D_GLOBE.tex 2010-02-21 22:02:11 UTC (rev 16301)
+++ seismo/3D/SPECFEM3D_GLOBE/trunk/USER_MANUAL/manual_SPECFEM3D_GLOBE.tex 2010-02-22 00:44:15 UTC (rev 16302)
@@ -137,7 +137,7 @@
for details. It was a finalist again in 2008 for a run at 0.16 petaflops (sustained) on 149,784 processors of the `Jaguar' Cray XT5 system at Oak Ridge National Laboratories (USA) \citep{CaKoLaTiMiLeSnTr08}.
Future improvements will include support for GPU graphics card acceleration \citep{KoMiEr09,KoErGoMi10,MiKo10}
-as well as Convolutional-Perfectly Matched absorbing Layers (C-PML) \citep{KoMa07}.
+as well as Convolutional-Perfectly Matched absorbing Layers (C-PML) \citep{KoMa07,MaKoEz08,MaKo09,MaKoGeBr10}.
\section{Citation}
@@ -156,7 +156,7 @@
Grant No. ANR-05-CIGC-002, and European FP6 Marie Curie International
Reintegration Grant No. MIRG-CT-2005-017461.
Older versions of the code were initially developed by Dimitri Komatitsch at Institut de Physique du Globe (France)
-and then at Harvard University (USA) by Dimitri Komatitsch and Jeroen Tromp.
+and then by Dimitri Komatitsch and Jeroen Tromp at Harvard University (USA).
Any opinions, findings, and conclusions or recommendations expressed in this material are
those of the authors and do not necessarily reflect the views of the
U.S. National Science Foundation, CNRS, INRIA, ANR or the European
@@ -1229,6 +1229,11 @@
\begin{lyxcode}
\noindent {\small Station~Network~Latitude~(degrees)~Longitude~(degrees)~Elevation~(m)~burial~(m)~}{\small \par}
\end{lyxcode}
+If you want to put a station on the ocean floor, just set
+elevation and burial depth in the STATIONS file to 0.
+Equivalently you can also set elevation to a negative value equal
+to the ocean depth, and burial depth to 0.
+
Solver output is provided in the \texttt{OUTPUT\_FILES} directory
in the \texttt{output\_solver.txt} file. Output can be directed to
the screen instead by uncommenting a line in \texttt{constants.h}:
@@ -2767,10 +2772,10 @@
Jeroen Tromp. The manual's cover graphic was created by Santiago Lombeyda
from Caltech's Center for Advanced Computing Research (CACR) \url{http://www.cacr.caltech.edu/}.
Older versions of the code were initially developed by Dimitri Komatitsch at Institut de Physique du Globe (France)
-and then at Harvard University (USA) by Dimitri Komatitsch and Jeroen Tromp.
+and then by Dimitri Komatitsch and Jeroen Tromp at Harvard University (USA).
Please e-mail your feedback, questions, comments, and suggestions
-to Jeroen Tromp \url{jtromp-AT-gps.caltech.edu} or to the CIG Computational Seismology Mailing List \url{cig-seismo at geodynamics.org}.
+to Jeroen Tromp \url{jtromp-AT-princeton.edu} or to the CIG Computational Seismology Mailing List \url{cig-seismo at geodynamics.org}.
\chapter*{\label{cha:Copyright}Copyright}
Modified: seismo/3D/SPECFEM3D_GLOBE/trunk/todo_list_please_dont_remove.txt
===================================================================
--- seismo/3D/SPECFEM3D_GLOBE/trunk/todo_list_please_dont_remove.txt 2010-02-21 22:02:11 UTC (rev 16301)
+++ seismo/3D/SPECFEM3D_GLOBE/trunk/todo_list_please_dont_remove.txt 2010-02-22 00:44:15 UTC (rev 16302)
@@ -2,8 +2,28 @@
To-do list for SPECFEM3D_GLOBE, by Dimitri Komatitsch:
------------------------------------------------------
-Things that could be done in a future version 4.1:
+Things that could be done in a future version:
+- there is something in SPECFEM3D_GLOBE that we noticed a few years ago
+regarding attenuation but never fixed: on page 813 of our 1999 paper
+I used a trick suggested by Robertsson et al. (1994)
+to use a non-staggered Runge-Kutta (RK4) scheme for the attenuation
+equations while using a staggered finite-difference (Newmark)
+time scheme for all the other equations.
+(combining staggered and non-staggered formulations being difficult)
+The trick works fine in most cases but there is a hidden problem for very long
+simulations, and in particular for multi-orbit surface waves:
+because of that trick, the RK4 is not really a real Runge-Kutta scheme
+(because grad(displacement) is used as a source but not known half way between
+time steps, therefore I replaced it with an average between
+t and t + Delta_t, which implies that the approximation is not fourth-order
+accurate any more because of that smoothing/interpolation). Therefore
+dispersion due to attenuation is not very accurately computed in the case of
+very long runs; that matters mostly for surface waves.
+We should fix that one day. The only thing to do would be to design a better time
+integration scheme for the attenuation equation, without that trick; the rest
+(Newmark etc) is fine and does not need to change.
+
- use a potential of (rho * u) instead of u in the fluid, in case of
fluid-fluid discontinuities
Modified: seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/bibliography.bib
===================================================================
--- seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/bibliography.bib 2010-02-21 22:02:11 UTC (rev 16301)
+++ seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/bibliography.bib 2010-02-22 00:44:15 UTC (rev 16302)
@@ -10649,7 +10649,7 @@
}
@INPROCEEDINGS{SavWiTr05,
- author = {Savage, B. and D.A. Wiens and J. Tromp},
+ author = {Savage, B. and D. A. Wiens and J. Tromp},
title = {Effects of 3{D} Velocity and Attenuation in the {T}onga-{F}iji Subduction
Zone},
booktitle = {Expanded abstracts of the American Geophysical Union},
Modified: seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/figures/gauss_vs_triangle_mod.pdf
===================================================================
(Binary files differ)
Modified: seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/figures/source_timing.pdf
===================================================================
(Binary files differ)
Modified: seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/manual_SPECFEM3D_SESAME.pdf
===================================================================
(Binary files differ)
Modified: seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/manual_SPECFEM3D_SESAME.tex
===================================================================
--- seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/manual_SPECFEM3D_SESAME.tex 2010-02-21 22:02:11 UTC (rev 16301)
+++ seismo/3D/SPECFEM3D_SESAME/trunk/USER_MANUAL/manual_SPECFEM3D_SESAME.tex 2010-02-22 00:44:15 UTC (rev 16302)
@@ -136,12 +136,12 @@
for details. It was a finalist again in 2008 for a run at 0.16 petaflops (sustained) on 149,784 processors of the `Jaguar' Cray XT5 system at Oak Ridge National Laboratories (USA) \citep{CaKoLaTiMiLeSnTr08}.
Future improvements will include support for GPU graphics card acceleration \citep{KoMiEr09,KoErGoMi10,MiKo10}
-as well as Convolutional-Perfectly Matched absorbing Layers (C-PML) \citep{KoMa07}.
+as well as Convolutional-Perfectly Matched absorbing Layers (C-PML) \citep{KoMa07,MaKoEz08,MaKo09,MaKoGeBr10}.
\section{Citation}
If you use SPECFEM3D\_SESAME for your own research, please cite at least one
-of the following articles: \cite{TrKoLi08,ChKoViCaVaFe07,KoLiTrSuStSh04,KoTr99} or \cite {KoVi98}.
+of the following articles: \cite{TrKoLi08,ChKoViCaVaFe07,KoLiTrSuStSh04,KoTr99} or \cite{KoVi98}.
The corresponding Bib\TeX{} entries may be found in file \texttt{USER\_MANUAL/bibliography.bib}
or in comments at the beginning of file \texttt{specfem3D.f90}.
@@ -154,7 +154,7 @@
Grant No. ANR-05-CIGC-002, and European FP6 Marie Curie International
Reintegration Grant No. MIRG-CT-2005-017461.
Older versions of the code were initially developed by Dimitri Komatitsch at Institut de Physique du Globe (France)
-and then at Harvard University (USA) by Dimitri Komatitsch and Jeroen Tromp.
+and then by Dimitri Komatitsch and Jeroen Tromp at Harvard University (USA).
Any opinions, findings,
and conclusions or recommendations expressed in this material are
those of the authors and do not necessarily reflect the views of the
@@ -1600,7 +1600,7 @@
Carl Tape, and Jeroen Tromp. The manual's cover graphic was created
by Santiago Lombeyda from Caltech's Center for Advanced Computing Research (CACR) \url{http://www.cacr.caltech.edu/}.
Older versions of the code were initially developed by Dimitri Komatitsch at Institut de Physique du Globe (France)
-and then at Harvard University (USA) by Dimitri Komatitsch and Jeroen Tromp.
+and then by Dimitri Komatitsch and Jeroen Tromp at Harvard University (USA).
Please e-mail your feedback, questions, comments, and suggestions
to Jeroen Tromp \url{jtromp-AT-princeton.edu} or to the CIG Computational Seismology Mailing List \url{cig-seismo at geodynamics.org}.
Modified: seismo/3D/SPECFEM3D_SESAME/trunk/todo_list_please_dont_remove.txt
===================================================================
--- seismo/3D/SPECFEM3D_SESAME/trunk/todo_list_please_dont_remove.txt 2010-02-21 22:02:11 UTC (rev 16301)
+++ seismo/3D/SPECFEM3D_SESAME/trunk/todo_list_please_dont_remove.txt 2010-02-22 00:44:15 UTC (rev 16302)
@@ -1,41 +1,45 @@
-To-do list for SPECFEM3D, by Dimitri Komatitsch
------------------------------------------------
+To-do list for SPECFEM3D_SESAME, by Dimitri Komatitsch
+------------------------------------------------------
-- we should keep and modify the old mesher from SPECFEM3D_BASIN in order
-to be able to quickly generate an analytical mesh for a simple basin without
-using CUBIT, because many users may still want to be able to use the old and
-simple mesher and not the (more complex) CUBIT package, and/or they might not
-have access to CUBIT. Therefore the old mesher should be modified to save its
-files in CUBIT/Abaqus format so that they can be used as input to "decompose_mesh".
+- there is something in SPECFEM3D that we noticed a few years ago
+regarding attenuation but never fixed: on page 813 of our 1999 paper
+I used a trick suggested by Robertsson et al. (1994)
+to use a non-staggered Runge-Kutta (RK4) scheme for the attenuation
+equations while using a staggered finite-difference (Newmark)
+time scheme for all the other equations.
+(combining staggered and non-staggered formulations being difficult)
+The trick works fine in most cases but there is a hidden problem for very long
+simulations, and in particular for multi-orbit surface waves:
+because of that trick, the RK4 is not really a real Runge-Kutta scheme
+(because grad(displacement) is used as a source but not known half way between
+time steps, therefore I replaced it with an average between
+t and t + Delta_t, which implies that the approximation is not fourth-order
+accurate any more because of that smoothing/interpolation). Therefore
+dispersion due to attenuation is not very accurately computed in the case of
+very long runs; that matters mostly for surface waves.
+We should fix that one day. The only thing to do would be to design a better
+time integration scheme for the attenuation equation, without that trick; the rest
+(Newmark etc) is fine and does not need to change.
-- Regarding domain decomposition we should forget about both
-METIS and ParMETIS, which are both inefficient
-for very large meshes. Pieyre has done detailed tests (Pieyre,
-could you please send the Excel file to all of us with a few lines
-of description of the figures?).
-SCOTCH (serial) and PT-SCOTCH (parallel) is much better.
-(faster, uses far less memory etc.)
-Because of that, we have removed support for METIS in SESAME.
-It can still be called indirectly from SCOTCH (there is an option
-in SCOTCH to call METIS instead). But this is a bad idea because
-the partition obtained is worse.
+- we could add support for the TetGen mesh creation package, and cut each
+tetrahedron into four hexahedra using the barycenter. This could help design
+meshes for sedimentary basins, in particular near basin edges
+- we could add support for different polynomial degrees in different elements (p-adaptivity)
+
- Regarding memory size (getting an estimate of memory consumption in SESAME),
somebody should just cut and paste my SPECFEM3D_GLOBE routine
"SPECFEM3D_GLOBE/version41_beta/src/memory_eval.f90", which
I call from "SPECFEM3D_GLOBE/version41_beta/src/create_header_file.f90".
It would work for SESAME as well (with minor modifications).
-- re-add the kernel calculations
-
- add SOURCESOLUTION in addition to CMTSOLUTION, to choose between a CMT source and a force source
- put a flag to choose between a Ricker source and a Heaviside source when a force source is used
(we can cut and paste the Heaviside implementation from the case of a CMT source)
-- automatic detection (coloring / flags) of the free surface and of the PML absorbing elements in CUBIT;
-read this from input files in the solver
+- automatic detection (coloring / flags) of the free surface and of the PML absorbing elements in CUBIT; read this from input files in the solver
- Pieyre should add C-PML following Roland's 2D implementation
@@ -116,3 +120,17 @@
- Added attenuation
- Added moment sources
+
+ - Put kernel calculations back
+
+- Regarding domain decomposition, METIS and ParMETIS can become inefficient
+for very large meshes. Pieyre has done detailed tests (Pieyre,
+could you please send the Excel file to all of us with a few lines
+of description of the figures?).
+SCOTCH (serial) and PT-SCOTCH (parallel) is then better.
+(faster, uses less memory etc.)
+Because of that, we have removed support for METIS in SESAME.
+It can still be called indirectly from SCOTCH (there is an option
+in SCOTCH to call METIS instead). But this is a bad idea because
+the partition obtained is worse.
+
More information about the CIG-COMMITS
mailing list