TimQu
39549f6ebd
Moved some functionality of StandardMinMaxSolver into a subclass
9 years ago
TimQu
c5ebfb74fb
repeatedMultiply methods of MinMaxSolvers now get the b-Vector as const
9 years ago
dehnert
a85f4fdc89
replaced some StoRMs and Storms by storm, reworked version output a bit
9 years ago
dehnert
5b09b91ae1
fixed more warnings
9 years ago
Sebastian Junges
d246517757
removed src prefix in all includes
9 years ago
Sebastian Junges
e1d201c85e
c++ code compiles again after rename
9 years ago
Sebastian Junges
3a7ee7867b
rename files (does not compile)
9 years ago
sjunges
548ba8bbeb
somehow managed my way through the policy guessing, several minor extensions to solvers
Former-commit-id: c4bb6453e7
10 years ago
sjunges
051ad702a7
solvers updated, constants updated
Former-commit-id: 011251c695
10 years ago
dehnert
83c4b1647c
solvers now can allocated auxiliary memory
Former-commit-id: 76dc1a1679
10 years ago
dehnert
95b95d9c64
fixed some minor issues and renamed equation solver methods slightly to make the names a bit more compact
Former-commit-id: de103e19ad
10 years ago
dehnert
b4e0cabef6
started working on general min-max solver that uses an underlying linear equation solver. provided necessary factories. adapted code and removed old min-max solvers
Former-commit-id: c1895472c7
10 years ago
dehnert
2a7dc0fad0
renamed MarkovChainSettings
Former-commit-id: 39024731f8
10 years ago
dehnert
00d331ebb4
moved linear equation solver factories to the respective solver files (and away from utility). restructured settings in factories and the way they are forwarded to the linear equation solvers. fixed all resulting errors
Former-commit-id: 27e1ae2466
10 years ago
Mavo
effadc5cca
Split into general settings and markov chain settings
Former-commit-id: 619a2e3622
10 years ago
Mavo
67d77608bd
Refactoring of settings
Former-commit-id: ea4350fc1c
10 years ago
sjunges
e0980de0ba
first version of storm without log4cplus as a dependency
Former-commit-id: 5aa64fabd7
10 years ago
dehnert
08bed36579
fixed an issue in performance tests and renamed all remaining LOG4CPLUS macro invocations to that of storm
Former-commit-id: 8536943978
10 years ago
TimQu
dae0faa2a0
refactored the quick implementation of the recent bugfix
Former-commit-id: 5c0d1fa3b9
10 years ago
TimQu
ca917a651c
MinMaxLinearEqSolvers can now use some initial policy as a first guess.
First steps to use this for region approximation
Former-commit-id: 9a8151607f
10 years ago
sjunges
8568ee3986
only one optimization direction enum -- towards integration of termination criterions on the model checker
Former-commit-id: 648855264e
11 years ago
sjunges
faf31156e0
fix for last changes + is probabilistic
Former-commit-id: 38df3f515f
11 years ago
sjunges
3c2040f4b7
Removed many superfluous includes, added some source files -- towards faster compilation
Former-commit-id: a575a97d40
11 years ago
dehnert
e338cbe069
fixed a lot of warnings in the tests
Former-commit-id: b6752202ac
11 years ago
dehnert
04f789619c
some work towards eliminating compiler warnings
Former-commit-id: d1eca470a4
11 years ago
dehnert
3b4dca1a03
Improved Jacobi method a bit.
Former-commit-id: f4affeebf6
11 years ago
dehnert
eb5d4100a6
Renamed Nondeterminstic equation solver as this name is more than misleading.
Former-commit-id: 7f08ed130c
11 years ago
dehnert
1990567b84
Started to improve performance of sparse CTMC model checker.
Former-commit-id: 1d014412ec
11 years ago
dehnert
d545fac471
Restructured solvers a bit: they now get the matrix upon construction and the model checkers use factories to retrieve solvers.
Former-commit-id: 9c727f41f9
11 years ago
David_Korzeniewski
4dc69dd6f5
Fixed performance tests, and again things concerning templates I never heard of before.
Former-commit-id: 1d110c6aad
11 years ago
David_Korzeniewski
7515ca5293
Fixed compile errors caused by parts of the c++ standard I've never heard of before...
Former-commit-id: 8dbd813f42
11 years ago
David_Korzeniewski
8ebc0e4640
Final touches on cuda nondeterministic linear equation solver & modelchecker
Former-commit-id: c549ae0401
11 years ago
David_Korzeniewski
ea2e616196
All tests for CUDA based TopologicalValueIterationMdpPrctlModelChecker passing on Windows.
Former-commit-id: 68cafa6f84
11 years ago
David_Korzeniewski
84f8a41302
More tests adapted, decreased verbosity of TopologicalValueIterationNondeterministicLinearEquationSolver
Former-commit-id: 6e0b492533
11 years ago
David_Korzeniewski
ee4c961cc9
fixes for compile errors. target "storm" builds without errors
tests not compiling because of property modifications.
Former-commit-id: 0366cf99cd
11 years ago
David_Korzeniewski
2279710443
Directly use Matrix with Decomposition
Former-commit-id: 745fa7c5c9
11 years ago
David_Korzeniewski
3e4495cad0
small fixes
Former-commit-id: 0d9cc58d75
11 years ago
David_Korzeniewski
78d3a392a5
Created settings module for TopologicalValueIterationNondeterministicLinearEquationSolver and integrated that with the solver.
Former-commit-id: fa1ad5ce2a
11 years ago
PBerger
c8e05f7137
Added explicit template instance.
Former-commit-id: 1ffcf0e47f
12 years ago
PBerger
493f93a94b
Added __restrict__ keyword to CUDA kernel. This should enhance compiler optimization.
Refactored TopologicalValueIterationNondeterministicLinearEquationSolver to support "down-casting" to float.
Added better timing output.
Former-commit-id: 688c40decb
12 years ago
PBerger
71320239f1
Some debug output.
Former-commit-id: dd0a60f099
12 years ago
PBerger
ea427fcde1
Fixed include directories for CUDA Plugin in CMakeLists.txt
Refactored all code related to the SPMV kernels to work with float.
Wrote a test that determines whether the compiler uses 64bit boundary alignments on std::pairs of uint64 and float.
Introduced functions that allow for conversions between different ValueTypes (e.g. from float to double and backwards).
Former-commit-id: 830d24064f
12 years ago
PBerger
422a317407
Made the OptimalSCC algorithm MUCH faster.
Fixed error reporting in AtomicPropositionLabelingParser.cpp and SparseStateRewardParser.cpp.
Former-commit-id: 77ba352a29
12 years ago
PBerger
7e77fbb6bb
Some testing stuff.
Former-commit-id: d7a9085af5
12 years ago
PBerger
73ddba5b29
Merged master, applied fixes.
Added feedback from the cuda plugin and return of iteration count.
Former-commit-id: 711ca3d9ec
12 years ago
PBerger
d2f4c85711
Made changes to comply with new SparseMatrix Interface (YUCK).
Fixed tests, all that stuff.
Former-commit-id: c78de5f8ce
12 years ago
PBerger
a4a17de4fc
Added timing for PRCTL formula checking.
Replaced std::sort with std::inplace_merge, saving another factor 2.
Former-commit-id: 961c31bb68
12 years ago
PBerger
26500ff4a8
Refactored the CUDA Kernel to once again use the "hacked" combination of column indices and values with a bit of reinterpret_cast magic.
Refactored the CUDA-SCC grouping algorithm as is took 80x longer to calculate the groups than it took to calculate the entire solution.
Former-commit-id: 5a5ffabe38
12 years ago
PBerger
0922921b24
Updated cudaForStorm/CMakeLists.txt to make use of the new GIT based version schema.
Added version functions to the Cuda Plugin.
Edited storm.cpp to show version infos for the CUDA Plugin.
Fixed a critical error in basicValueIteration.cu which causes random SEGFAULTs... :P
Streamlined the TopologicalValueIterationNondeterministicLinearEquationSolver.cpp. The SCC group optimizer now returns flat_sets instead of a vector as the sets are ordered, which is required for the Solver to work.
This is now a stable version of StoRM containing a fully wor
Former-commit-id: 47d5c2825c
12 years ago
PBerger
05814f5d73
Fixed a bug in the equalModuloPrecision function of the CUDA Kernel
Added more debug output to the CUDA handler functions
Added a function for grouping of SCCs for better performance
Added functionality and accessors to the SparseMatrix
Former-commit-id: 770aec1b09
12 years ago