98b143cd22
Switched all loop indices in Matrix/Vector to size_t, with improved results in timeGaussianFactorGraph
2010-01-20 14:42:05 +00:00
b4f9163296
Added more labels and results (with higher numbers of repetitions) for timeGaussianFactorGraph
2010-01-20 14:31:24 +00:00
8db99db57e
Small performance improvement for matrix_augmented, with accompanying test in timeGaussianFactor
2010-01-20 14:09:44 +00:00
6bc0462669
Checked Matrix::collect() with different timing metric, and memcpy has approx 25% speedup over original version. Matrix_range is slower or the same speed as the original version.
2010-01-20 13:45:22 +00:00
977a97eacf
rename variable
2010-01-20 09:47:09 +00:00
aae49e43f0
code factorization, speedup
2010-01-20 06:49:19 +00:00
ae908316e0
add comments
2010-01-20 06:34:25 +00:00
19dc8bf4b1
Fixed size() crash when BayesTree empty, added clear() function to BayesTree, added and removed #includes to fix link errors
2010-01-20 04:23:35 +00:00
e767996b45
same bug fix as in ISAM2
2010-01-20 03:54:20 +00:00
55e414cbdd
Tried noalias() for collect(), but no real improvement
2010-01-20 03:25:55 +00:00
3bf15333af
move ordering into the solver, and the nonlinear optimizer is now exact <G, T, L>
2010-01-20 02:28:23 +00:00
03ec3e3e62
Implemented a version of collect() using matrix_range, but it had worse performance than memcpy. New version of vector_scale[column] with new pointer arithmetic implementation for approx 20% improvement.
2010-01-20 02:28:25 +00:00
820ae5682d
Optimized row version of vector_scale with approx 20% improvement
2010-01-20 02:04:25 +00:00
ecde508318
incremental nonlinear working
2010-01-20 01:52:40 +00:00
b860f06614
Switched Matrix::collect() over to a memcpy, but the results weren't actually an improvement
2010-01-20 01:42:36 +00:00
d2291a38d1
bug fixes after COLAMD activiation: BayesTree::insert was wrong, BayesTree::removeTop/Path fixed and simplified
2010-01-20 01:24:32 +00:00
66607897a5
Added vector_scale timing benchmarks
2010-01-20 01:05:40 +00:00
ff83a7b18e
Fixed compile error in testNoiseModel
2010-01-20 01:05:18 +00:00
b47438a86c
Smart named constructors Covariance/Variances/Variance
2010-01-20 00:26:49 +00:00
f577b27f52
Small optimizations on Matrix::collect(), additional timing capabilities, and now collect() can have dimensions specified to avoid need for lookup.
2010-01-19 21:49:22 +00:00
67878830a6
Added timeMatrix.cpp to test specific matrix operations
2010-01-19 20:59:22 +00:00
3cdbaf81c7
Factorgraph-specific QR factorization now done by the NoiseModel: slow Gram-Schmidt for Constrained, fast Housholder for any other (Gaussian) model.
2010-01-19 19:06:02 +00:00
6f2b5a904a
save on fabs
2010-01-19 19:04:37 +00:00
9f4da004c2
got rid of verbose printing
2010-01-19 19:04:18 +00:00
fb0ca07bf1
Added and tested whitening parameter for matrix() and matrix_augmented()
2010-01-19 16:52:01 +00:00
fbee0bd749
Fixing compile issue in SubgraphPreconditioner.h
2010-01-19 13:29:40 +00:00
3806125096
add linear system as a template parameter in nonlinear optimizer
...
fixed a nasty bug and change the internal data type of subgraph preconditioner from reference to boost shared pointer. reference is not a good idea for class members, because no type checking will happen
2010-01-19 10:46:12 +00:00
2eac3b7235
minor typo - comment only
2010-01-19 05:46:04 +00:00
306ac0a4b2
Fixed clobbered checkin
2010-01-19 05:45:25 +00:00
88e465910a
SQP tests and implementation now use the new Key system
2010-01-19 05:33:44 +00:00
2a57a04ba6
get_model
2010-01-19 05:29:31 +00:00
6dcc275432
small comment/print
2010-01-19 04:39:54 +00:00
b01c111a1d
GaussianFactor now uses noiseModel internally - not yet externally
2010-01-19 04:39:28 +00:00
81aaceb890
Whiten now class-specific, access to sigmas
2010-01-19 04:38:58 +00:00
ebbdfc79aa
comment out overflow exception
2010-01-19 04:18:59 +00:00
ef2fb243c2
fixed unit test after colamd change
2010-01-19 03:43:10 +00:00
2507b6aaca
Cleaned up commented line after fix to use the colamd ordering.
2010-01-19 03:16:09 +00:00
0311d2a2a2
Very very small change :-)
2010-01-19 02:57:08 +00:00
0769af6f20
comment
2010-01-19 01:41:44 +00:00
a45893a434
selective incremental linearization works; bug fix in landmark initialization (frank01 again much faster)
2010-01-19 01:18:28 +00:00
a2bce15c4b
selective update of linearization point
2010-01-19 00:15:46 +00:00
0e6607d160
relinearization working, but currently applied to all variables
2010-01-18 22:37:44 +00:00
5e15564525
Speedups: Matrix mult. in Gaussian NoiseModel. Removed size check in operator* in Matrix.h - ublas does this check when in debug mode.
2010-01-18 21:45:44 +00:00
e935f1745e
stop copying matrices...
2010-01-18 21:17:25 +00:00
f8ef284b30
fixed bearing/range, large speedup for batch; incremental creation of Config works
2010-01-18 20:17:31 +00:00
1f0de30a62
Removed unnecessary headers, now only included when GTSAM_MAGIC_KEY is defined
2010-01-18 19:36:02 +00:00
e523b148cf
Moved tree stuff to graph, removed ordering-inl.h
2010-01-18 19:11:22 +00:00
8e3a8f08c8
incremental processing of planarSLAM; bug fix in ISAM2-inl
2010-01-18 18:13:53 +00:00
dfeacb218e
Added tests for Keys, Added TypedLabeledSymbol class to allow for runtime symbols on keys
2010-01-18 16:18:02 +00:00
7653aa8747
I changed myconfigure to
...
./configure --prefix=$HOME --with-toolbox=$HOME/toolbox/ --with-boost=/opt/local/include/ CXXFLAGS=" -pipe -g -O2 -march=core2 -DNDEBUG" --disable-static
based on reading From http://www.gentoo.org/doc/en/gcc-optimization.xml , including:
-O2: A step up from -O1. This is the recommended level of optimization unless you have special needs. -O2 will activate a few more flags in addition to the ones activated by -O1. With -O2, the compiler will attempt to increase code performance without compromising on size, and without taking too much compilation time.
-O3: This is the highest level of optimization possible, and also the riskiest. It will take a longer time to compile your code with this option, and in fact it should not be used system-wide with gcc 4.x. The behavior of gcc has changed significantly since version 3.x. In 3.x, -O3 has been shown to lead to marginally faster execution times over -O2, but this is no longer the case with gcc 4.x. Compiling all your packages with -O3 will result in larger binaries that require more memory, and will significantly increase the odds of compilation failure or unexpected program behavior (including errors). The downsides outweigh the benefits; remember the principle of diminishing returns. Using -O3 is not recommended for gcc 4.x.
2010-01-18 16:13:07 +00:00
Alex Cunningham
Kai Ni
Michael Kaess
Richard Roberts
Frank Dellaert