3
0
Fork 0
mirror of https://github.com/Z3Prover/z3 synced 2025-08-31 15:24:55 +00:00
Commit graph

1235 commits

Author SHA1 Message Date
Nikolaj Bjorner
d218083145 The demodulator doesn't produce proofs so remove code path that depends it does. 2022-12-04 04:48:48 -08:00
Nikolaj Bjorner
7fe6787748 ufbv-rewriter is really a demodulator rewriter and does not reference ufbv
so moving first the rewriter into place of other rewriters
2022-12-04 04:44:02 -08:00
Nikolaj Bjorner
79e6d4e32d tune and debug elim-unconstrained (v2 - for simplifiers infrastructure) 2022-12-04 03:53:31 -08:00
Nikolaj Bjorner
cf7bba6288 use ast_manager as an attribute 2022-12-04 03:53:30 -08:00
Nikolaj Bjorner
b084821a0c wip - dependent expr simpliifer
- simplify iterator over current indices
- add more simplifiers used by asserted_formulas
- improve diagnostics printing
2022-11-30 13:41:40 +07:00
Nikolaj Bjorner
caf204ab95 hoist macro-replacer as shared utility, update eliminate-predicates and model reconstruction 2022-11-24 19:45:51 +07:00
Nikolaj Bjorner
eceeb295fc fix #6457 2022-11-24 14:41:50 +07:00
Nikolaj Bjorner
9a2693bb72 tune euf-completion 2022-11-23 16:39:20 +07:00
Nikolaj Bjorner
22353c2d6c new core perf - add merge_tf and enable_cgc distinction
perf fix for propagation behavior for equalities in the new core.
The old behavior was not to allow congruence closure on equalities.
The new behavior is to just not allow merging tf with equalities unless they appear somewhere in a foreign context (not under a Boolean operator)

The change re-surfaces merge_tf and enable_cgc distinction from the old core.
They can both be turned on or off.

merge_enabled renamed to cgc_enabled

The change is highly likely to introduce regressions in the new core.

Change propagation of literals from congruence:
- track antecedent enode. There are four ways to propagate
literals from the egraph.
- the literal is an equality and the two arguments are congruent
- the antecedent is merged with node n and the antecedent has a Boolean variable assignment.
- the antecedent is true or false, they are merged.
- the merge_tf flag is toggled to true but the node n has not been merged with true/false
2022-11-23 11:37:24 +07:00
Nikolaj Bjorner
ba68652c72 add destructive equality resolution to existentials 2022-11-19 18:43:46 +07:00
Nikolaj Bjorner
98fc8c99db add shortcut to equality mk utility 2022-11-16 03:47:01 -08:00
Nikolaj Bjorner
255414f4a9 fix regression crash
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
2022-11-15 11:20:12 -08:00
Nikolaj Bjorner
9845c33236 add shortcuts in rewriter, eliminate redundancies in dependent_expr tactic 2022-11-15 09:13:13 -08:00
Nikolaj Bjorner
3a37cfca30 switch to solve_eqs2 tactic 2022-11-08 12:23:36 -08:00
Nikolaj Bjorner
f769e2f1f6 have bool rewriter use flat_and_or, and integrate hoist rewriter 2022-11-08 12:21:50 -08:00
Nikolaj Bjorner
3a4b8e2334 add rewrite rules to bv-rewriter 2022-11-08 12:20:51 -08:00
Nikolaj Bjorner
a34701471f clean up hoist rewriter 2022-11-08 12:20:25 -08:00
Nikolaj Bjorner
6c12aaad74 wip - testing solve-eqs2, added as tactic 2022-11-05 22:42:59 -07:00
Nikolaj Bjorner
154b09309b fixing build, wip on model reconstruction integration into dependent-expr-state 2022-11-04 14:04:44 -07:00
Nikolaj Bjorner
de9368bab0 Update expr_replacer.h 2022-11-04 11:25:39 -07:00
Nikolaj Bjorner
28668c6efc set up model reconstruction trail 2022-11-04 11:25:39 -07:00
Nikolaj Bjorner
1646a41b2f minor fixes
- ensure mk_extract performs simplification to distribute over extract and removing extract if the range is the entire bit-vector
- ensure bool_rewriter simplifeis disjunctions when applicable.
2022-11-02 08:44:55 -07:00
Nikolaj Bjorner
a409a4a677 enforce flat within QF_BV tactic, cap in-processing var-elim loops 2022-10-27 20:10:55 -07:00
Nikolaj Bjorner
1fae3aa152 rename set-flat to set-flat-and-or to allow to differentiate parameters 2022-10-27 11:22:57 -07:00
Nikolaj Bjorner
fe1b4bf5ce disable ternary, fixes to propagation, make bv_rewrites for multiplier n-ary 2022-10-26 23:44:38 -07:00
Nikolaj Bjorner
5352a0106d fix #6426 2022-10-26 12:20:55 -07:00
Nikolaj Bjorner
c62c5e9d23 add opportunistic, missing, bv rewrites
- x >> x logical = 0
- ~x = -1 -x
- x * (y << z) = (x * y) << z
2022-10-25 10:29:48 -07:00
Nikolaj Bjorner
876ca2f1a5 fix #6371 2022-09-30 14:51:28 -04:00
Nikolaj Bjorner
bd4db4c41f add option to rewrite and for arithmetic simplification 2022-09-18 17:22:59 -07:00
Nikolaj Bjorner
088898834c filter length limits to be non-skolems and under concat/""/unit 2022-09-15 07:41:13 -07:00
Nikolaj Bjorner
c47ca341b7 fix #6343
The bug was that axiom generation was not enabled on last_index, so no axioms got created to constrain last-index.
With default settings the solver is now very slow on this example. It is related to that the smallest size of a satisfying assignment is above 24. Pending a good heuristic to find initial seeds and increments for iterative deepening, I am adding another parameter smt.seq.min_unfolding that when set to 30 helps for this example.
2022-09-14 10:17:25 -07:00
Nikolaj Bjorner
53611f47df modify clauses used by not-contains
The literal "emp" can be true in the current assignment, in which case the clause
cnt or emp or ~postf is true and does not contribute to propagation.
This saves, potentially, for generating lemmas for postf.

Add a lemma a = "" or |s| >= idx when a = tail(s, idx)
The lemma ensures that length bounding on s is enforced
(the branch that expands not-contains for long sequences s is closed).
2022-09-11 05:48:17 -07:00
Nikolaj Bjorner
6df711254b fix type error when mapping over the empty sequence 2022-09-10 16:03:52 -07:00
Nikolaj Bjorner
8311525472 map and fold cannot be treated as variables 2022-09-10 16:03:24 -07:00
Nikolaj Bjorner
46383a0811 AG - unary datatypes, tester always is true. 2022-09-01 09:45:56 -07:00
Nikolaj Bjorner
48b13291d1 add bv-size reduce #6137
- add option smt.bv.reduce_size.
  - it allows to apply incremental pre-processing of bit-vectors by identifying ranges that are known to be constant.
    This rewrite is beneficial, for instance, when bit-vectors are constrained to have many high-level bits set to 0.
2022-08-16 16:35:14 -07:00
Bruce Mitchener
5014b1a34d Use = default for virtual constructors. 2022-08-05 18:11:46 +03:00
Nikolaj Bjorner
80c516bb50 squash stores 2022-08-05 13:57:35 +03:00
Bruce Mitchener
5d0dea05aa
Remove empty leaf destructors. (#6211) 2022-07-30 10:07:03 +01:00
Bruce Mitchener
1eb84fe4b9
Mark override methods appropriately. (#6207) 2022-07-29 23:29:15 +02:00
Bruce Mitchener
3e38bbb009
Make sure all headers do #pragma once. (#6188) 2022-07-23 10:41:14 -07:00
Nikolaj Bjorner
004139b320 rewrites for characters
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
2022-07-02 11:37:21 -07:00
Nikolaj Bjorner
ea2a843325 flat only
remove option for uzers (users who are in reality fuzzers) to toggle flat option. The legacy arithmetic solver bakes in assumptions about flat format so it isn't helpful to expose this to fuzzers, I mean uzers.
2022-06-30 19:59:46 -07:00
Nuno Lopes
41deed59a3 fix bug in array rewriter introduced in 202ce1e 2022-06-21 22:40:40 +01:00
Nikolaj Bjorner
36a1f758bc mask regression 2022-06-21 14:34:47 -07:00
Nikolaj Bjorner
ab9aee189b perf #6100 2022-06-21 13:49:52 -07:00
Nikolaj Bjorner
202ce1edf0 #6100 - two perf fixes
remaining perf bug is dealing with very large bit-widths. mod 2^n should be computed natively based on n instead of 2^n because we pre-populate an array with all values up to n. Suppose n is 10000, the array has size 10000.
2022-06-21 12:45:29 -07:00
Nikolaj Bjorner
477e9625ef Don't reset the cache between applications of replace
tactic/lia2card shows a huge slowdown because the same replace function is called on thousands of assertions. Each time the cache gets reset with thousands of entries - they are all the same.
So don't reset the cache just because... Instead reset the cache if m_refs grows large.
2022-06-16 15:40:01 -07:00
Nikolaj Bjorner
9cd339841a for Arie
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
2022-06-10 18:07:54 -07:00
Nikolaj Bjorner
994dab8eb6 add pre-filter for F* use case
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
2022-06-10 17:56:48 -07:00