4d214c162f
sf_pml: user defined grading function using the fparser lib
2010-07-30 17:02:21 +02:00
b2ac68d835
sf_pml: moved Zm into GetKappaGraded
2010-07-30 15:29:19 +02:00
e85f8ef8b7
allow boundary conditions strings e.g. "PEC" "PMC" or "PML_8"
2010-07-30 15:28:15 +02:00
3d1c7f22b9
update & fixes to sf_pml to support z-direction pml in clyindrical coords
2010-07-29 18:32:57 +02:00
8316b1c2bd
Operator: GetNodeArea & Update/fix in Calc_EffMatPos method
...
- new method GetNodeArea will return the area of a node for a given direction
- methods Calc_ECPos & Calc_EffMatPos now const
- Calc_EffMatPos in Operator class updated to use new functions GetMeshDelta & GetNodeArea
- Calc_EffMatPos introduced (fixed) in Operator_Cylinder (need some testing)
- treatment of E_alpha=0 at r==0 moved from Calc_ECPos to Electric-BC
2010-07-29 18:30:50 +02:00
56154aeb05
processfields: mesh correction for non-interpolated case
2010-07-20 16:30:41 +02:00
1f65648776
added discretization info to processfields.cpp
...
The coordinate snapping needs to be revised!
2010-07-20 11:42:47 +02:00
e5930c7774
field probes: add info to text files
2010-07-20 11:37:19 +02:00
e3904c0f18
changed node-interpolation in H-field dumps
...
average the H-fields not the currents
2010-07-19 09:04:13 +02:00
1a93650fa0
new H-field probe
2010-07-19 08:41:53 +02:00
efe64fca4d
pml warning only if using the pml
2010-07-17 15:22:51 +02:00
d3434906a3
new: split field pml implementation started
...
todo:
- pml edges and corners
- flexible profile definition
- lots of testing !!!
Signed-off-by: Thorsten Liebig <Thorsten.Liebig@gmx.de>
2010-07-16 17:25:32 +02:00
0ccbbab593
new probe type: E-field
2010-07-16 15:55:35 +02:00
55ac7137a3
minor documentation updates
2010-07-16 10:41:12 +02:00
5e904b112c
write H-field excitation into file "ht"
...
If we have a separate signal for H-field excitation, we need to dump it.
The question is, do we really need a separate Signal?
2010-07-16 10:33:54 +02:00
f7edb2f1fd
changed openEMS dumps from drawing units to unit of meters
...
the better approach is to create a new class (e.g. Dump)which contains all dump formats (currently vtk and hdf5). This class will then be responsible to choose a dump mode: in drawing units or in meters
2010-07-15 12:58:48 +02:00
050a595cd3
docu update & typo
2010-07-15 12:12:52 +02:00
7744bf6030
fix: missing Engine_Extension deconstructor
2010-07-15 12:11:48 +02:00
fb164ba1fe
update to console output regarding used timestep method & typo fix
2010-07-13 15:37:56 +02:00
59ffbb5100
Improve the speed of the SSE engine by using flush-to-zero
...
This change modifies the behaviour of the complete program, if the SSE engine is used.
A better approach may be to only enable flush-to-zero in IterateTS()...
2010-07-13 10:47:40 +02:00
911f7c5528
show extensions status & define const methods
2010-07-11 23:45:41 +02:00
03b03b9834
sse compression output: yes & no instead of 0/1
2010-07-11 19:53:02 +02:00
9b8c280504
dispersive engine extension: support inline methods
2010-07-11 17:21:59 +02:00
b524b57be0
new: dispersive materials / first and simple lorentz/drude dispersive materials
...
todo:
- more material parameters for the lorents model (currently only the plasma frequencies)
- code documentation
- examples & validation
- speed optimizations
Signed-off-by: Thorsten Liebig <Thorsten.Liebig@gmx.de>
2010-07-06 10:01:26 +02:00
a083c6d739
make excitation error message a bit more elaborate
2010-06-29 16:10:53 +02:00
91ecb547d6
fix in multithreading linesPerThread
...
Signed-off-by: Thorsten Liebig <Thorsten.Liebig@gmx.de>
2010-06-29 12:44:47 +02:00
0f4a78cdbf
new ProcessingIntegral as abstract base class to voltage and current processing
2010-06-28 19:45:19 +02:00
e880fa3ee8
Processing: some code cleanups & consistencies
...
Signed-off-by: Thorsten Liebig <Thorsten.Liebig@gmx.de>
2010-06-28 18:27:41 +02:00
6f06497dab
Added frequency domain probe support
...
Signed-off-by: Thorsten Liebig <Thorsten.Liebig@gmx.de>
2010-06-28 18:05:03 +02:00
dd71ae33ae
bugfix and optimization of compressed sse operator
2010-06-25 09:38:49 +02:00
ebacc62812
fix: new timestep not (always) stable for cylinder-coords, using the old one...
2010-06-22 12:49:51 +02:00
3fd58b7e7c
rewritten current calculation...
2010-06-18 12:52:02 +02:00
b776061f7f
new CalcTimestep for larger a timestep
2010-06-18 12:37:37 +02:00
87b8e22bf7
Add Probe-Box weighting (new in CSXCAD)
2010-06-16 12:50:19 +02:00
88b319cbba
some typos & consol message fixes
2010-06-07 23:08:38 +02:00
11df0f5c37
array_ops: continue templates
...
Signed-off-by: Thorsten Liebig <thorsten.liebig@gmx.de>
2010-06-06 20:22:05 +02:00
97de0e84ee
make memory allocators templates (to allow for other types than FDTD_FLOAT)
...
Signed-off-by: Thorsten Liebig <thorsten.liebig@gmx.de>
2010-06-06 20:00:24 +02:00
ab0a4aaacb
number of threads received by cylindrical operator & engine
...
Signed-off-by: Thorsten Liebig <thorsten.liebig@gmx.de>
2010-06-06 01:50:58 +02:00
08fd499194
operator: multithreaded CalcPEC
2010-06-06 01:47:32 +02:00
e816d50cc1
operator: split up CalcPEC as preparation for easier multithreading
...
Signed-off-by: Thorsten Liebig <thorsten.liebig@gmx.de>
2010-06-06 00:53:05 +02:00
b4c98632bf
MT-Operator: now using multithreading for Calc_EC
...
Calc_PEC & Excitation should be next
Signed-off-by: Thorsten Liebig <thorsten.liebig@gmx.de>
2010-06-05 11:50:13 +02:00
f0141f6c7e
Operator: split Calc_EC into Init_EC + Calc_EC
2010-06-05 11:47:21 +02:00
0954177dc6
operator debug dumps: more console infos & check first if file can be created
...
Signed-off-by: Thorsten Liebig <thorsten.liebig@gmx.de>
2010-06-04 14:08:42 +02:00
7172e9e53d
obey mesh type in material/operator/PEC dump
2010-06-02 17:21:58 +02:00
610f4e62da
processfields: allow different mesh types for vtk dump, e.g. cylindrical mesh dump
2010-06-02 16:41:26 +02:00
82befba245
new debug option: dump PEC
...
command line parameter: --debug-PEC
writes a file PEC_dump.vtk which can be visualized with paraview
visualize one component (x,y,z) at a time using arrow glyphs
2010-06-02 16:37:21 +02:00
96144ed3a1
bugfix: snapping of current probes
...
(sorry...)
2010-06-02 08:57:45 +02:00
27bf0aac7d
Warning for unused primitives
2010-05-29 17:40:18 +02:00
7ad8f4ac3c
fixed broken sse-compressed operator!!!
2010-05-29 17:14:36 +02:00
0aada84f80
DumpOperator2File now available for all types of operators
2010-05-29 12:47:07 +02:00
d96a592e05
Defining the compressed sse operator access functions
2010-05-29 12:45:38 +02:00
4bda515078
processfields: allow low level access to vtk write methods
2010-05-29 12:17:15 +02:00
f671b52986
added command line option --showProbeDiscretization
2010-05-28 15:13:14 +02:00
5163c69e7e
fix: data size is MiB instead of MB
2010-05-21 10:14:09 +02:00
f71960c4e0
Adept compression cache to grid size
2010-05-21 08:22:41 +02:00
6bc146a3fb
cylinder operator now based on multithreaded-operator
2010-05-20 23:51:40 +02:00
dab5be229b
multithreaded engine now a derived class of compressed sse engine, created by new operator_multithread class
...
- also fixed double init() call of engines
- operator_multithread open for future multithreaded operator creation
2010-05-20 22:02:06 +02:00
df17dc23c6
evil hack to get multithreading working again
...
future: restructure the multithreading engine to only work on top the sse_compressed engine
2010-05-20 17:39:39 +02:00
c71b0154a1
allow multiple excitations at the same edges (ignore priorities)
2010-05-20 15:17:30 +02:00
8212f039ed
more console output regarding operator & engine usage
2010-05-19 21:25:15 +02:00
f868eb328c
operator_cylinder with flexible base class -> now sse-compressed
2010-05-19 21:10:48 +02:00
009eeef96a
fix: number of lines already defined by engine base-class
2010-05-19 21:08:38 +02:00
23b917be5f
update to Operator::ShowStat and derived operators
2010-05-19 11:42:56 +02:00
de5bdc8ef6
introducing compressed sse operator & engine
...
use --engine=sse-compressed option to enable
2010-05-19 11:41:35 +02:00
72330844cc
critical fix in engine field access methods
2010-05-18 14:10:11 +02:00
25777f7f27
cylinder extensions: two critical bug fixes
2010-05-17 15:44:03 +02:00
f98314d55c
mur_abc extension speedup by using inline engine functions
2010-05-17 13:10:14 +02:00
98c10c7628
define engine type and overload all GetVolt/GetCurr
2010-05-17 13:08:27 +02:00
b2c436282e
fix: init numTS in engine constructor
2010-05-17 13:05:41 +02:00
a0c215a40c
allow node-interpolated hdf5 dump
2010-05-12 20:40:35 +02:00
a1ef3426f2
FIX in sse engine: init fields & fixed reset calls
2010-05-11 22:54:08 +02:00
10bbb9db79
write interpolation type info to vtk file
2010-05-11 21:03:16 +02:00
5e5a278ac7
new: node interpolated dump in processfields
2010-05-11 20:38:58 +02:00
279b348724
correction in H-field cell-interpolation
2010-05-11 20:38:09 +02:00
4245f26e5e
fix of excite signal dump et
2010-05-11 19:11:38 +02:00
abbe10bd9f
fix: init sse correctly
2010-05-11 19:11:38 +02:00
7f70f0fb4d
engine independent multi-threading
2010-05-11 19:11:38 +02:00
54d0671d33
new: operator should create its engine
2010-05-11 19:11:38 +02:00
74ba9eceff
cylinder-operator docu
2010-05-11 19:11:37 +02:00
8cc2a2dd44
Moved special cylinder operations into op extentions
2010-05-11 19:11:37 +02:00
0a2f5fee5a
update to Operator::ShowStat + count PEC
2010-05-10 09:14:29 +02:00
7cc2a7fcf2
bug fix in current integration
2010-05-07 22:59:02 +02:00
5b3ccd3f53
bug fix in interpolated h-field dump
2010-05-07 22:58:56 +02:00
065c05eeae
limit the gauss excitation to max timesteps
2010-05-05 17:28:00 +02:00
09cc2c7360
added extension support to cylinder-operator/engine & removed some unused code
2010-05-03 23:08:47 +02:00
ca8d2f8833
check whether extensions are cylinder-coords approved or not...
2010-05-03 22:37:29 +02:00
fe483946d2
cylinder-operator: use operator.h EC calc
2010-05-03 22:36:04 +02:00
4cff6eebc5
minor fix
2010-05-03 21:20:49 +02:00
f762214eae
split excitation from operator
...
The excitation variables and functions are separated into a class Excitation.
This allows completely different operator implementations (e.g. TLM) to use the excitations functions without deriving from class Operator.
2010-05-03 18:33:14 +02:00
99a082e8a6
sse: small code-cleanup
2010-05-01 13:06:13 +02:00
63ef1b3ebf
sse engine: changed the vector order to reduce copy operations
2010-05-01 12:57:43 +02:00
d26986fe0d
sse: removed more unnecessary methods & new operator access & names that make more sense
2010-04-30 19:41:44 +02:00
2a3e8dd20f
new: operator access functions
2010-04-30 19:40:08 +02:00
3e09b22b4a
removed unnecessary methodes in sse engine & fixed excitation
2010-04-30 19:09:05 +02:00
209c066732
Merge branch 'master' into sse
...
Conflicts:
TESTSUITE/helperscripts/invoke_openEMS.m
2010-04-30 15:29:21 +02:00
ab7a25d3ee
separate excitation signal for voltages and currents
2010-04-30 11:28:53 +02:00
d16545c776
dump time information to hdf5 file
2010-04-29 19:26:45 +02:00
b1b1e71126
Operator: added H-field/current excitation
2010-04-29 00:28:07 +02:00
8787d793aa
update to field excitation
2010-04-28 22:42:55 +02:00
885fc8ee26
Merge branch 'master' into sse
...
Conflicts:
FDTD/engine.h
openEMS.pro
openems.cpp
2010-04-28 22:35:12 +02:00
d5c4747118
operator BC handling changed
2010-04-28 22:01:03 +02:00
138624d2d8
mur_abc_extension: now allows material distribution
2010-04-28 21:45:05 +02:00
03d8935cb8
fix in field processing + subsampling + cell-interpolated
2010-04-28 21:03:15 +02:00
3c3edd32e0
Mur_ABC: start after wall-excitation
2010-04-28 18:21:02 +02:00
4fcee4588f
win32 compile fixes
2010-04-28 11:15:59 +02:00
b296c441f9
extention updates & new extention: Mur's absorbing boundary condition
2010-04-27 23:06:42 +02:00
9c5c5e9057
new: operator & engine extensions
2010-04-25 21:59:05 +02:00
13bd884e8e
removed invalid inline specifier
2010-04-25 20:28:17 +02:00
e9cdadb01a
processing functions are now independent of the underlying engine data format
2010-04-25 20:28:00 +02:00
71e8febc3b
cylinder operator: some fixes
2010-04-24 14:06:00 +02:00
7b81cab005
bug fix in cylindrical operator
2010-04-23 16:31:00 +02:00
dca73e45fc
some fixes in cylindrical operator
2010-04-23 08:19:56 +02:00
5535a386ff
get correct mesh direction names for dump
2010-04-23 08:17:42 +02:00
1fbffaa502
removed invalid inline specifier
2010-04-21 15:43:39 +02:00
d0ebff17e8
removed sse code from normal operator -> class Operator_sse
2010-04-21 15:38:15 +02:00
b489d3342d
completed sse engine and enabled it
2010-04-21 14:29:02 +02:00
9e66b0e544
processing functions are now independent of the underlying engine data format
2010-04-21 14:28:16 +02:00
95188dd15e
sse tests
2010-04-21 11:18:22 +02:00
d81576cedd
write time domain excitation function to file et1
2010-04-21 11:16:30 +02:00
71bd794cc1
multithreaded engine: use common excitation function
2010-04-20 13:10:25 +02:00
d611e5c8b1
debug: write boxes to vtk-file
2010-04-19 16:09:41 +02:00
c814015f31
make it const
2010-04-19 16:08:32 +02:00
b3e2d9428a
fix in Operator_Cylinder
2010-04-14 11:01:17 +02:00
5314235474
CylinderCoords example using r=0 / cleanup
2010-04-13 20:06:06 +02:00
ebff88207d
suppress all compiler warnings/ code-cleanup
2010-04-13 18:51:44 +02:00
86832d0d3a
CylinderCoords: include r=0 and many fixes & necessary changes
...
- r=0 case included... needs lots of testing...
- field processing can't access mesh directly --> use operator methods
2010-04-13 18:40:43 +02:00
22210247f4
Cylinder_Engine: removed unnecessary Init()/Reset() method's
2010-04-13 18:34:26 +02:00
5bf478abbb
removed false inline function attribute
2010-04-13 18:28:46 +02:00
b4bc990b0f
define dump precision
2010-04-12 09:38:24 +02:00
4db42917bb
new in CylinderCoords: closed alpha field domain
...
todo:
- include r=0
- make sure a varying mesh in alpha-direction is OK
2010-04-11 23:52:38 +02:00
3b29514d16
Access function for number of Lines
2010-04-11 23:42:54 +02:00
101e55202d
fix: current dump timestamp
2010-04-09 21:15:44 +02:00
3ca8609020
fix: don't process current integration outside of simulation domain
2010-04-09 15:58:15 +02:00
2c21303833
NEW: first implementation of Cylindrical Coord FDTD
...
todo:
- allow closed alpha mesh
- allow r=0 case
- adept field processing
- ...
2010-04-09 15:51:37 +02:00
9ae8ac12ab
new excitation: Custom excite and some excitation related code-cleanup
2010-04-07 16:31:23 +02:00
fc600f53ba
fix: Don't create .h5 files if dump is disabled
2010-04-07 16:29:55 +02:00
52f5764976
Supporting sub-sampled dumps
2010-04-07 12:57:45 +02:00
924f0a6d40
first hdf5-file dump implementation
2010-04-05 20:22:03 +02:00
8cc42b911c
added processing steps and fix in end criteria check, now checking max excite TS
2010-04-03 17:36:50 +02:00
f73d6f80f8
fixed nyquist handling
2010-04-02 17:20:18 +02:00
0274f4e253
removed friend-classes from operator
2010-04-02 17:07:56 +02:00
37ff221c18
multithreaded engine works
...
but it's slow...
2010-04-01 16:11:55 +02:00
98063bba6e
new excitation step
2010-04-01 09:38:56 +02:00
f7a17f3678
bugfix: SetProcessInterval
2010-04-01 09:38:08 +02:00
70874e56ab
changed some functions to const
2010-03-31 16:35:43 +02:00
2409a77399
bugfix: multithreaded excitation
2010-03-31 15:37:39 +02:00
813683304b
new commandline switch: --numThreads=
2010-03-30 13:13:00 +02:00
9fdc9096d3
bugfix: length of excitation
2010-03-30 13:10:42 +02:00
4b12054f66
merged
2010-03-30 13:10:23 +02:00
55c0719b4e
Merge branch 'master' into multithreading
2010-03-30 08:33:47 +02:00
b1cb1e30e3
minor fixes
2010-03-30 08:19:41 +02:00
abc705767f
update to console output
2010-03-29 22:11:24 +02:00
09364107cf
Merge remote branch 'origin/master' into multithreading
2010-03-29 10:18:33 +02:00
5a47a6443a
some more int vs. u-int fixes
2010-03-29 10:12:38 +02:00
fd2623d740
removed engine friend classes, use only engine methods to access protected data
2010-03-29 10:01:38 +02:00
b8180287e9
time debug code can now be disabled
2010-03-28 13:10:16 +02:00
ea2823377d
added Curve for excitation and dump excitation in operator-dump
2010-03-27 23:05:45 +01:00
fb193ac25b
debugging
2010-03-27 22:57:25 +01:00
54344b1b86
bug: field dumper directly access protected numTS member variable...
2010-03-27 15:26:20 +01:00
e108e17cec
bugfix: calculate correct number of timesteps
2010-03-27 11:32:06 +01:00
d01900a48f
make it virtual
2010-03-27 11:29:58 +01:00
79b0e6a2e0
first implementation of a multithreaded engine (bugs and mem leaks present!)
2010-03-26 23:53:05 +01:00
98f72a855e
fixed lots of int/u-int comparison and missing return values
2010-03-26 19:55:53 +01:00
207e22f1db
fix in operator curve handling
2010-03-26 19:33:44 +01:00
a7ccec254d
static createEngine() added to work around virtual constructor problems
2010-03-26 17:24:43 +01:00
c916059a30
new engine: multithreaded
2010-03-26 12:57:52 +01:00
686cb3f4cb
derived engines need access to special operator properties
2010-03-26 11:28:54 +01:00
b4bf720466
fix in gauss excite...
2010-03-26 08:39:26 +01:00
8e3f44b0ca
operator: removed some debugging code
2010-03-26 08:23:37 +01:00
d79b15f49e
Curve/Wire support, Helix update
2010-03-22 08:19:17 +01:00
1f20f7ae42
new command-line option: dump operator to vtk-file
2010-03-17 23:16:41 +01:00
0019eeb9b7
some fixes in processing routines
2010-03-16 21:41:17 +01:00
3ff4eec004
rewrite E-Field-Excitation
2010-03-16 21:39:37 +01:00
e4ecd7ca2a
fixed: graded mesh bug in operator
2010-03-16 09:40:01 +01:00
c7ed4a2e76
fix in end-criteria and field-energy-calc
2010-03-15 22:19:51 +01:00
8dc89ecac2
end Criteria with energy estimate; more engine stats while running
2010-03-15 16:59:37 +01:00
b9a3165aac
Update to vtk-field dumps + material-dump for debugging
2010-03-12 20:39:04 +01:00
9b638b1568
fix in operator PEC-calc
2010-03-12 08:38:07 +01:00
73e2b153b6
GPL licensing
2010-03-11 16:47:40 +01:00
cd1df70591
unused AdrOp removed from Engine
2010-03-11 16:30:22 +01:00
98e4790c0e
Operator check if electrode is valid or not
2010-03-11 15:03:24 +01:00
4efb7661bd
Operator::SetGeometryCSX returns success
2010-03-11 10:48:47 +01:00
0e1eba5028
Processing updates: CalcLineIntegral
2010-03-11 10:48:00 +01:00
f73bf210ed
new ProcessingArray; Processing returns next interval for process()
...
Processing now uses a timestep interval in which Process() will do it's work
and returns the next interval it doesn't need another Process() call
all Processing routines can be put into these array
and it will take care of calling Process() at the right time
2010-03-10 12:15:14 +01:00
c8dacd8c31
Operator updates
2010-03-09 21:49:30 +01:00
d31799e912
default dump not-interpolated
2010-03-09 21:35:57 +01:00
9e86a96436
Sinusoidal excitation
2010-03-09 21:34:23 +01:00
4b1044942a
don't process fields in case no file pattern is defined
2010-03-07 21:59:48 +01:00
5208d3f9a2
Major Change: Moved CartOperator into Operator
...
No more seperate Operator for cartesian and future cylindrical operator.
2010-03-05 14:20:25 +01:00
161bdf9c28
Seperate Operator init
2010-03-05 13:53:05 +01:00
5ee107fcbc
Processing default enabled
2010-03-04 16:54:16 +01:00
0a39d32a07
A couple of new features
...
- ProcessCurrent
- Weighted Excitation and Material
- Testing
2010-03-04 11:53:58 +01:00
f1fe4b5277
Minor changes to console output
2010-03-03 08:00:25 +01:00
84ba010482
PEC support added
2010-03-02 22:55:50 +01:00
5a04affca2
PEC-BC fix again!
2010-03-02 22:25:40 +01:00
1c92ab2920
Engine + BC bug fix
2010-03-02 19:01:03 +01:00
6d2e974cc1
fix in snapping, E-Dump and new H-Dump
2010-03-02 15:37:00 +01:00
ca44334da5
Field dumps (E-field) TD, hard E-source
2010-03-02 14:54:50 +01:00
61f69dd240
Some testing...
2010-03-01 20:50:42 +01:00
eea86d4184
Engine fix!, processing class, lots of fixes
2010-03-01 19:35:28 +01:00
Thorsten Liebig
Sebastian Held
Sebastian Held
Sebastian Held