# Tabling: another core dump

**URL:** <https://swi-prolog.discourse.group/t/tabling-another-core-dump/842>\
**Category:** General\
**Created:** [June 20, 2019, 3:35am UTC](https://swi-prolog.discourse.group/t/tabling-another-core-dump/842 "2019-06-20T03:35:39Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\
**Post date:** [June 20, 2019, 3:35am UTC](https://swi-prolog.discourse.group/t/tabling-another-core-dump/842/1 "2019-06-20T03:35:39Z")

</div>

Running this example from [Stack overflow](https://stackoverflow.com/questions/30026151/uneven-tabling-performance-in-bprolog-8-1):

```prolog

:- table posInt_CollatzSteps/2. 
posInt_CollatzSteps(I,N) :-
   ( I == 1
   -> N = 0 % base case
   ; 1 is I /\ 1
   -> I0 is I*3+1, posInt_CollatzSteps(I0,N0), N is N0+1 % odd
   ; I0 is I>>1, posInt_CollatzSteps(I0,N0), N is N0+1 % even
   ).

i0_i_maxSteps0_maxSteps(I0,I,M0,M) :-
   ( I0 > I
   -> M0 = M
   ; posInt_CollatzSteps(I0,N0),
      I1 is I0+1,
      M1 is max(M0,N0),
      i0_i_maxSteps0_maxSteps(I1,I,M1,M)
   ).

```

And this query:

```prolog
2 ?- time(i0_i_maxSteps0_maxSteps(1,1 000 000,0,MaxSteps)).
[Thread 1 (main) at Wed Jun 19 22:46:21 2019] ../src/pl-trie.c:1234: find_var: Assertion
 failed: index == state->max_var_seen+1
C-stack trace labeled "assert_fail":
  [0] save_backtrace() at :? [0x7f09721abadf]
  [1] __assert_fail() at ??:? [0x7f09721e01c6]
  [2] unify_key() at :? [0x7f09721d9026]
  [3] trie_gen() at :? [0x7f09721db239]
  [4] pl_tbl_answer_update_dl2_va.lto_priv.0() at :? [0x7f09721db474]
  [5] PL_next_solution() at ??:? [0x7f09722685bc]
  [6] query_loop() at :? [0x7f097222d192]
  [7] prologToplevel() at :? [0x7f097222d26b]
  [8] PL_toplevel() at ??:? [0x7f0972279cae]
  [9] swipl(+0x1075) [0x55d3ede7b075]
  [10] __libc_start_main() at ??:? [0x7f0971f97ee3]
  [11] swipl(+0x10be) [0x55d3ede7b0be]
[1] 23511 abort (core dumped) swipl

```

It works fine without tabling (albeit it takes a very long time).

---

<div class="post-metadata">

**Author:** ![jan](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/jan/32/4_2.png) [@jan](https://swi-prolog.discourse.group/u/jan)\
**Post date:** [June 20, 2019, 11:22am UTC](https://swi-prolog.discourse.group/t/tabling-another-core-dump/842/3 "2019-06-20T11:22:51Z")

</div>

With latest git (8.1.8-25-g583b8636c):

```prolog
103 ?- time(i0_i_maxSteps0_maxSteps(1, 1 000 000, 0, R)).
% 72,659,302 inferences, 12.005 CPU in 12.061 seconds (100% CPU, 6052541 Lips)
R = 524.

104 ?- time(i0_i_maxSteps0_maxSteps(1, 1 000 000, 0, R)).
% 5,000,001 inferences, 1.165 CPU in 1.169 seconds (100% CPU, 4290451 Lips)
R = 524.

```

These patches also fix [Tabling: Core dump](https://swi-prolog.discourse.group/t/tabling-core-dump/839/2)

---

<div class="post-metadata">

**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\
**Post date:** [June 20, 2019, 11:32am UTC](https://swi-prolog.discourse.group/t/tabling-another-core-dump/842/4 "2019-06-20T11:32:33Z")

</div>

Great! Tremendous response time. I think you surpassed Linus on this one 🙂