Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.lang.prolog > #14230
| From | Mild Shock <janburse@fastmail.fm> |
|---|---|
| Newsgroups | comp.lang.prolog |
| Subject | Re: variant_term/2 is faster than variant/1 (Was: Request for comments, Novacore the sequel to ISO modules) |
| Date | 2024-10-13 00:54 +0200 |
| Message-ID | <veeump$b5p9$1@solani.org> (permalink) |
| References | <db8a6771-3e54-485b-b391-310658dd6f52n@googlegroups.com> <aa869ef6-7f90-405b-a84a-2c11ac7979a3n@googlegroups.com> <vcukob$n3q2$1@solani.org> <veekkd$b9uf$1@solani.org> <veeq19$bcbb$1@solani.org> |
Ok, I have to redo my nok/2 test for Scryer Prolog,
since Scryer Prolog doesn't work as per ISO spec:
/* Scryer Prolog Playground */
?- builtins:variant(f(A,B), f(C,D)).
A = A, B = A, C = A, D = A.
It should not leave some bindings. At least this is
what other Prolog systems do:
/* Trealla Prolog */
?- variant(f(A,B), f(C,D)).
true.
/* SWI-Prolog */
?- variant(f(A,B), f(C,D)).
true.
The performance tests are not affected, but I guess
it is better to test nok/2 like this:
nok2(L, R) :- length(L, 6), length(R, 6),
enum_list(_, L), enum_list(_, R),
boole(variant_term(L, R), A), boole(builtins:variant(L, R), B),
A \== B.
Besides the binding glitch, it seems to be ok:
?- nok2(L, R).
false.
Mild Shock schrieb:
> So what can we do with this insight. Here is
> a little performance of a new distinct/1,
> that is eager und non-variant enumerating:
>
> /* eagerness */
> ?- distinct(repeat), !.
> true.
>
> d(f(A,A)).
> d(f(a,a)).
> d(f(B,B)).
>
> /* non-variant enumerating */
> ?- distinct(d(X)).
> X = f(_70321, _70321);
> X = f(a, a);
> fail.
>
> It is implemented with variant_term/2, since the
> remembered list of archived witnesses so far,
> is anyway copies of terms. So variant_term/2
>
> is appropriate, we never have to check
> two terms that have variables in common.
>
> Cool!
>
> Mild Shock schrieb:
>> How much faster is it?
>>
>> Here some test harness:
>>
>> test :- length(L, 6), length(R, 6),
>> enum_list(_, L), enum_list(_, R),
>> variant(L, R), fail; true.
>>
>> test2 :- length(L, 6), length(R, 6),
>> enum_list(_, L), enum_list(_, R),
>> variant_term(L, R), fail; true.
>>
>> Here some results:
>>
>> - SWI-Prolog 9.3.11:
>>
>> ?- time(test).
>> % 2,126,498 inferences, 0.734 CPU in 0.752 seconds
>> (98% CPU, 2895657 Lips)
>> true.
>>
>> ?- time(test2).
>> % 2,159,795 inferences, 0.234 CPU in 0.236 seconds
>> (99% CPU, 9215125 Lips)
>> true.
>>
>> - Trealla Prolog 2.57.16:
>>
>> ?- time(test).
>> % Time elapsed 3.128s, 14827949 Inferences, 4.741 MLips
>> true.
>>
>> ?- time(test2).
>> % Time elapsed 1.079s, 7775516 Inferences, 7.206 MLips
>> true.
>>
>> - Scryer Prolog :
>>
>> ?- time(test3).
>> % CPU time: 5.653s, 5_192_831 inferences
>> true.
>>
>> ?- time(test2).
>> % CPU time: 1.544s, 6_116_163 inferences
>> true.
>>
>> Note: test3 is like test, only it uses builtins:variant/2.
>>
>> Mild Shock schrieb:
>>> The ISO core standard probably set the
>>> stage for a couple of performance sins.
>>> In 7.1.6.1 Variants of a term we find
>>>
>>> these test cases:
>>>
>>> - f(A, B, A) is a variant of f(X, Y, X).
>>> - g(A, B) is a variant of g(_, _).
>>> - P+Q is a variant of P+Q.
>>>
>>> What is doubious here, is the last test
>>> case with P+Q. Do we need to test terms
>>> that have common variables?
>>>
>>> Lets assume we have situations where we
>>> don't need variant working with common
>>> variables in the two argument terms, what
>>>
>>> about then using this bootstrapping:
>>>
>>> variant_term(X, Y) :-
>>> subsumes_term(X, Y),
>>> subsumes_term(Y, X).
>>>
>>> Here some testing, does it work ok? Take this code:
>>>
>>> enum_arg(_, 1).
>>> enum_arg(_, _).
>>> enum_arg(X, X).
>>>
>>> enum_list(_, []).
>>> enum_list(X, [H|T]) :- enum_arg(X, H), enum_list(X, T).
>>>
>>> boole(G, 1) :- G, !.
>>> boole(_, 0).
>>>
>>> nok(L, R) :- length(L, 6), length(R, 6),
>>> enum_list(_, L), enum_list(_, R),
>>> boole(variant(L, R), A), boole(variant_term(L, R), B),
>>> A \== B.
>>>
>>> Seems to work fine:
>>>
>>> ?- nok(L, R).
>>> false.
>>>
>>>
>>
>
Back to comp.lang.prolog | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-08-17 05:18 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-08-17 06:37 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-08-17 07:00 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-08-17 13:28 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-08-31 02:59 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-08-31 03:04 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-08-31 03:10 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-08-31 10:18 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-08-31 10:22 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-09 16:06 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-09 16:12 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-09 16:13 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-09 16:14 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-09 23:57 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-09 23:59 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-12 05:11 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-12 10:46 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-12 10:48 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-12 13:22 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-13 05:53 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-13 07:32 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-13 11:16 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-13 11:18 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-13 11:36 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-13 13:28 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-14 01:41 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-14 01:43 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-14 07:41 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <janburse@fastmail.fm> - 2022-10-15 14:30 +0200
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-19 07:42 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-19 07:50 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-19 07:51 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-19 07:52 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-19 18:55 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-19 18:58 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-19 19:46 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-25 05:55 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-25 05:59 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-25 08:36 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-25 10:53 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-26 07:01 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-10-26 07:02 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-11-14 14:21 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-11-14 14:31 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-11-17 06:08 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2022-11-17 06:21 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-11 12:45 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-11 12:52 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-11 14:24 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-13 01:56 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-13 01:58 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-13 01:59 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-13 02:01 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-13 04:49 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-16 15:03 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-16 15:07 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-17 00:49 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-01-17 03:22 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-02-27 03:08 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-02-27 03:08 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-02-27 10:31 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-02-27 10:33 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-02-27 16:05 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-01 09:01 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-01 09:02 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-02 04:06 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-02 04:08 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-08 01:45 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-08 01:47 -0800
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-18 05:55 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-18 05:56 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-18 06:01 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-03-18 06:06 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-05-13 06:36 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-05-13 06:41 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mostowski Collapse <bursejan@gmail.com> - 2023-05-21 03:54 -0700
Differences among the "bomb" and "xbetween" (Was: Request for comments, Novacore the sequel to ISO modules) Mild Shock <janburse@fastmail.fm> - 2024-09-24 17:10 +0200
The issue with free speech Mild Shock <janburse@fastmail.fm> - 2024-10-09 15:29 +0200
failure of formal verification [software.imdea.org] (Was: The issue with free speech) Mild Shock <janburse@fastmail.fm> - 2024-11-06 09:33 +0100
Re: failure of formal verification [software.imdea.org] (Was: The issue with free speech) Mild Shock <janburse@fastmail.fm> - 2024-11-06 09:35 +0100
Re: failure of formal verification [software.imdea.org] (Was: The issue with free speech) Mild Shock <janburse@fastmail.fm> - 2024-11-06 09:37 +0100
variant_term/2 is faster than variant/1 (Was: Request for comments, Novacore the sequel to ISO modules) Mild Shock <janburse@fastmail.fm> - 2024-10-12 22:02 +0200
Re: variant_term/2 is faster than variant/1 (Was: Request for comments, Novacore the sequel to ISO modules) Mild Shock <janburse@fastmail.fm> - 2024-10-12 23:16 +0200
Re: variant_term/2 is faster than variant/1 (Was: Request for comments, Novacore the sequel to ISO modules) Mild Shock <janburse@fastmail.fm> - 2024-10-12 23:34 +0200
Re: variant_term/2 is faster than variant/1 (Was: Request for comments, Novacore the sequel to ISO modules) Mild Shock <janburse@fastmail.fm> - 2024-10-13 00:54 +0200
Re: Request for comments, Novacore the sequel to ISO modules Mild Shock <bursejan@gmail.com> - 2023-07-29 04:51 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mild Shock <bursejan@gmail.com> - 2023-09-07 17:15 -0700
Re: Request for comments, Novacore the sequel to ISO modules Mild Shock <janburse@fastmail.fm> - 2023-11-19 19:56 +0100
Re: Request for comments, Novacore the sequel to ISO modules Mild Shock <janburse@fastmail.fm> - 2023-11-19 19:59 +0100
Re: Request for comments, Novacore the sequel to ISO modules Mild Shock <janburse@fastmail.fm> - 2023-11-19 20:01 +0100
New milestone float formatting [LoL] (Was: Request for comments, Novacore the sequel to ISO modules) Mild Shock <janburse@fastmail.fm> - 2024-07-28 14:32 +0200
Re: New milestone float formatting [LoL] (Was: Request for comments, Novacore the sequel to ISO modules) Mild Shock <janburse@fastmail.fm> - 2024-07-28 14:38 +0200
Re: New milestone float formatting [LoL] (Was: Request for comments, Novacore the sequel to ISO modules) Mild Shock <janburse@fastmail.fm> - 2024-07-28 16:42 +0200
Re: Differences among the "bomb" and "xbetween" (Was: New milestone float formatting [LoL]) Mild Shock <janburse@fastmail.fm> - 2024-09-24 17:20 +0200
New milestone time formatting (Was: Differences among the "bomb" and "xbetween") Mild Shock <janburse@fastmail.fm> - 2024-09-26 13:22 +0200
More Prolog Improvement Proposals (PIPs) [PIP401 - PIP404] Mild Shock <janburse@fastmail.fm> - 2024-10-09 09:43 +0200
Even More Prolog Improvement Proposals (PIPs) [PIP301 - PIP303] (Was: More Prolog Improvement Proposals (PIPs) [PIP401 - PIP404]) Mild Shock <janburse@fastmail.fm> - 2024-10-09 09:45 +0200
Re: Even More Prolog Improvement Proposals (PIPs) [PIP301 - PIP303] (Was: More Prolog Improvement Proposals (PIPs) [PIP401 - PIP404]) Mild Shock <janburse@fastmail.fm> - 2024-10-09 10:00 +0200
Re: Even More Prolog Improvement Proposals (PIPs) [PIP301 - PIP303] (Was: More Prolog Improvement Proposals (PIPs) [PIP401 - PIP404]) Mild Shock <janburse@fastmail.fm> - 2024-10-09 11:22 +0200
A FFI for evaluable functions Mild Shock <janburse@fastmail.fm> - 2024-10-10 00:14 +0200
Name resolution is a blackbox (Re: A FFI for evaluable functions) Mild Shock <janburse@fastmail.fm> - 2024-10-10 00:16 +0200
plonk in discourse (Was: The issue with free speech) Mild Shock <janburse@fastmail.fm> - 2024-10-10 00:48 +0200
When do two Prolog terms marry? (Was: The issue with free speech) Mild Shock <janburse@fastmail.fm> - 2024-10-14 19:15 +0200
Re: When do two Prolog terms marry? (Was: The issue with free speech) Mild Shock <janburse@fastmail.fm> - 2024-10-14 19:22 +0200
native implementation of variant/2 can be fast (Was: When do two Prolog terms marry?) Mild Shock <janburse@fastmail.fm> - 2024-10-15 01:35 +0200
KISS principle pays off (Was: A FFI for evaluable functions) Mild Shock <janburse@fastmail.fm> - 2024-11-03 00:52 +0100
Waste of EU money / bread and butter of statistics (Was: failure of formal verification [software.imdea.org]) Mild Shock <janburse@fastmail.fm> - 2024-11-06 09:51 +0100
Re: Waste of EU money / bread and butter of statistics (Was: failure of formal verification [software.imdea.org]) Mild Shock <janburse@fastmail.fm> - 2024-11-06 10:09 +0100
Re: Waste of EU money / bread and butter of statistics (Was: failure of formal verification [software.imdea.org]) Julio Di Egidio <julio@diegidio.name> - 2024-11-06 10:43 +0100
SICStus Prolog is overrated (Was: Waste of EU money / bread and butter of statistics) Mild Shock <janburse@fastmail.fm> - 2024-11-06 12:57 +0100
Crashing under its own bloath, Scryer Prolog (Was: SICStus Prolog is overrated) Mild Shock <janburse@fastmail.fm> - 2024-11-06 13:03 +0100
“Luce,” which is Italian for “light.” (Was: Crashing under its own bloath, Scryer Prolog) Mild Shock <janburse@fastmail.fm> - 2024-11-06 13:12 +0100
More Poisened Meat balls? [NIVIDIA & OpenAI] (Way: “Luce,” which is Italian for “light.”) Mild Shock <janburse@fastmail.fm> - 2024-11-06 13:33 +0100
Re: More Poisened Meat balls? [NIVIDIA & OpenAI] (Way: “Luce,” which is Italian for “light.”) Julio Di Egidio <julio@diegidio.name> - 2024-11-06 14:17 +0100
Not all Red-Black Trees are Okasaki (Was: Request for comments, Novacore the sequel to ISO modules) Mild Shock <janburse@fastmail.fm> - 2024-11-07 00:59 +0100
Re: Request for comments, Novacore the sequel to ISO modules Mild Shock <bursejan@gmail.com> - 2023-11-19 10:54 -0800
csiph-web