PRISM
=====
Version: 4.5.dev
Date: Sat Mar 14 14:06:05 UTC 2020
Hostname: e72bdd194fc5
Memory limits: cudd=4g, java(heap)=11g
Command line: prism -javamaxmem 11g -cuddmaxmem 4g -ii -heuristic speed -e 1e-6 -ddextraactionvars 100 -maxiters 1000000 kanban.prism kanban.props --property throughput -const t=5
Parsing model file "kanban.prism"...
Type: CTMC
Modules: k1 k2 k3 k4
Variables: w1 x1 y1 z1 w2 x2 y2 z2 w3 x3 y3 z3 w4 x4 y4 z4
Parsing properties file "kanban.props"...
1 property:
(1) "throughput": R{"throughput"}=? [ S ]
---------------------------------------------------------------------
Model checking: "throughput": R{"throughput"}=? [ S ]
Model constants: t=5
Warning: Switching to sparse engine and (backwards) Gauss Seidel (default for heuristic=speed).
Building model...
Model constants: t=5
Computing reachable states...
Reachability (BFS): 71 iterations in 0.35 seconds (average 0.004873, setup 0.00)
Time for model construction: 0.235 seconds.
Type: CTMC
States: 2546432 (1 initial)
Transitions: 24460016
Rate matrix: 6308 nodes (14 terminal), 24460016 minterms, vars: 48r/48c
SCCs: 1, BSCCs: 1, non-BSCC states: 0
BSCC sizes: 1:2546432
Computing steady state probabilities for BSCC 1
Computing probabilities...
Engine: Sparse
Building sparse matrix... [n=2546432, nnz=24460016, compact] [95.7 MB]
Creating vector for diagonals... [dist=187, compact] [4.9 MB]
Allocating iteration vector... [19.4 MB]
TOTAL: [120.0 MB]
Starting iterations...
Iteration 61: max relative diff=0.005125, 5.07 sec so far
Iteration 122: max relative diff=0.000052, 10.14 sec so far
Backwards Gauss-Seidel: 175 iterations in 28.84 seconds (average 0.083114, setup 14.30)
BSCC 1 Reward: 0.30712474345277757
All states are in BSCCs (so no reachability probabilities computed)
Value in the initial state: 0.30712474345277757
Time for model checking: 29.556 seconds.
Result: 0.30712474345277757 (value in the initial state)
Overall running time: 30.346 seconds.
---------------------------------------------------------------------
Note: There was 1 warning during computation.