max/lean4-htt
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 2
11661 commits 1 branch 0 tags 5.3 GiB
Commit graph

8347 commits

Author SHA1 Message Date
Sebastian Ullrich
b9dc1c251e fix(library/inductive_compiler/ginductive_decl): assertion 2017-03-02 08:02:55 -08:00
Sebastian Ullrich
852df8872b fix(library/constructions/drec): assertion 2017-03-02 08:01:38 -08:00
Daniel Selsam
1f6306d068 perf(library/inductive_compiler): simplification with sizeof lemmas 2017-03-01 21:13:20 -08:00
Johannes Hölzl
f44cbb896c fix(src/library/equations_compiler/elim_match): handle mixing of inaccessible terms and variables 2017-03-01 21:12:42 -08:00
Leonardo de Moura
1542cd750f feat(library/tactic/induction_tactic): use drec in the induction tactic 
The new test failed before this change.
 2017-03-01 18:34:24 -08:00
Leonardo de Moura
e9f18502a7 chore(library/user_recursors): add drec_on and dcases_on 2017-03-01 15:56:06 -08:00
Leonardo de Moura
1ded3b70b8 feat(library/constructions/drec): add dcases_on 2017-03-01 15:46:19 -08:00
Leonardo de Moura
132a629eb7 fix(library/tactic/induction_tactic): use whnf when inferring C.rec name 2017-03-01 14:29:26 -08:00
Leonardo de Moura
7b0a18167b feat(library/constructions/drec): add drec_on and refactor 2017-03-01 14:12:10 -08:00
Leonardo de Moura
5fa7d765ac chore(emacs/lean-server): remove leftover from experiment 2017-02-28 22:57:20 -08:00
Leonardo de Moura
17556758cb feat(library/constructions,library/inductive_compiler): automatically generate dependent eliminator for inductive predicates 
The dependent eliminator for an inductive predicate C is called C.drec

TODO: construct dcases_on and drec_on using C.drec

We need this recursor for implementing dependent elimination for
inductive predicates.

We don't need to define acc.drec and eq.drec in the standard library anymore.
 2017-02-28 20:58:04 -08:00
Leonardo de Moura
cad379333d feat(util/memory_pool): put limit on the size of memory_pool free_lists 
See #1405

Memory consumption is still high, but I didn't manage to cross the 2Gb
limit anymore with this commit even after hundreds of modifications.

@gebner I'm not seeing a big difference betwee Lean without memory_pool,
with bounded memory_pool and unbounded memory_pool. We may even consider
removing it in the future after a more careful benchmarking.

In the benchmark (https://gist.github.com/leodemoura/b27fb4203a13a67274b388a602149303),
I'm getting the following numbers:

- No memory_pool: runtimes between 3.532s - 3.556s

- With memory_pool bounded by 8192: runtimes between 3.32s - 3.44s

- With memory_pool (with no limit): runtimes between 3.32s - 3.44s

On the other hand, the small object allocator makes a big difference.
I used your list_rev.lean example.

- with:    2.62s
- without: 3.75s
 2017-02-28 15:16:43 -08:00
Leonardo de Moura
1fa1312c57 chore(CMakeLists,frontends/lean): add SAVE_SNAPSHOT=OFF SAVE_INFO=OFF compilation options 2017-02-28 12:09:19 -08:00
Gabriel Ebner
f5ae60e667 feat(init/init): write allocation stats to stderr 2017-02-28 11:56:52 -08:00
diakopter
cbaf060a80 chore(library): override warning 2017-02-28 11:24:21 -08:00
Leonardo de Moura
db5b709d9c feat(CMakeLists,util): add CUSTOM_ALLOCATORS=OFF compilation option 2017-02-28 11:13:54 -08:00
Leonardo de Moura
544d5c67f4 feat(CMakeLists,kernel): add TRACK_LIVE_EXPRS=ON compilation option 2017-02-28 10:56:14 -08:00
Leonardo de Moura
ba641ce29f feat(CMakeLists,util): add TRACK_CUSTOM_ALLOCATORS=ON compilation option 2017-02-28 10:38:09 -08:00
Leonardo de Moura
1da93c4d28 feat(CMakeLists, shell/lean): add -D ALPHA=ON cmake option for enabling alpha features (i.e., wip) 
@jroesch This commit also marks --compile as an alpha feature.
 2017-02-26 23:06:45 -08:00
Leonardo de Moura
2386c18ee5 fix(library/tactic/eval): make sure old position information nested in the expression being evaluated is not used in type error messages 
see #1401
 2017-02-26 22:38:24 -08:00
Leonardo de Moura
184d505d51 fix(library/compiler/preprocess): do not unfold noncomputable definitions in the compiler 
see #1401
 2017-02-26 22:05:16 -08:00
Leonardo de Moura
d7674dd060 chore(CMakeLists): bumped master branch to 3.1.1 2017-02-25 18:12:04 -08:00
Gabriel Ebner
fcc0f30b84 fix(frontends/lean/definition_cmds): set scope_pos_info when checking examples 2017-02-25 12:56:46 -08:00
Leonardo de Moura
52221cdbd1 fix(frontends/lean/elaborator): {} elaboration issue 2017-02-24 21:20:39 -08:00
Gabriel Ebner
7e3be51a52 chore(*): remove last remnants of mpfr 2017-02-24 21:42:58 +01:00
Gabriel Ebner
9bcfc06bd0 fix(library/module): has_sorry: check examples 2017-02-24 21:08:49 +01:00
Leonardo de Moura
552a185e6a feat(frontends/lean): 'let' in 'do' blocks 2017-02-24 09:10:36 -08:00
Sebastian Ullrich
dd379e5b34 refactor(library/equations_compiler/elim_match): simplify is_complete_transition 2017-02-23 01:52:14 +01:00
Sebastian Ullrich
dfe1874365 refactor(frontends/lean/{parser,util}): extract quote functions 
Also fixes ``f when f is private
 2017-02-23 01:52:13 +01:00
Sebastian Ullrich
1c7ca3f20a feat(frontends/lean/parser): ignore implicit arguments in expr patterns 2017-02-23 01:52:13 +01:00
Sebastian Ullrich
44b27a9a44 chore(library/vm/vm): fix comment 2017-02-23 01:52:13 +01:00
Sebastian Ullrich
908a7bd9f3 feat(frontends/lean/parser): expr patterns 2017-02-23 01:52:13 +01:00
Sebastian Ullrich
3aa9e32c5f fix(library/equations_compiler/elim_match): always prefer value transitions over complete transitions 2017-02-23 01:21:14 +01:00
Leonardo de Moura
6b76b65881 feat(library/equations_compiler/elim_match): change default max number of steps to 2048 2017-02-21 21:33:10 -08:00
Leonardo de Moura
d9307413b9 feat(library/compiler): skip bytecode optimization for interactive tactics 
This is just overhead for begin...end blocks.
Moreover, "live variable analysis" is currently a recursive
procedure, and Lean will run of stack space when processing big begin
end blocks (> 1000 tactics).
 2017-02-21 21:23:58 -08:00
Leonardo de Moura
ca31ad0107 perf(library/metavar_util): quadratic behavior when assembling the final proof 
The peformance problem was affecting theorems that contain many `intro`
tactic applications.

@gebner After this optimization, the GAPT benchmark elaboration time went from
1.6 secs to 0.6 secs.
 2017-02-21 21:02:43 -08:00
Leonardo de Moura
321105099f feat(library/type_context): add compilation flag for disabling type inference cache 2017-02-21 20:17:25 -08:00
Leonardo de Moura
09819cb159 feat(library/type_context): add compilation flag for disabling type class resolution flag 2017-02-21 20:04:43 -08:00
Leonardo de Moura
7c57d23b46 feat(kernel): add compilation flags for disabling has_local optimization and abstraction cache 2017-02-21 19:46:31 -08:00
Leonardo de Moura
1c959f6790 feat(kernel): add compilation flag for disabling "free var range" optimization 2017-02-21 19:40:56 -08:00
Leonardo de Moura
622e9a6035 feat(library/type_context): use m_unfold_pred to decide whether macros should be unfolded or not 
see #1394
 2017-02-21 18:07:39 -08:00
Leonardo de Moura
074574be0a feat(library/vm/vm_array): add native array.iterate and array.foreach 2017-02-21 16:22:25 -08:00
Leonardo de Moura
d9dcb4461e perf(library/vm/vm_array): minor optimization 2017-02-21 15:43:48 -08:00
Leonardo de Moura
3f87fd15eb feat(library/vm): add liveness analysis, and support destructive updates for lean arrays 2017-02-21 15:09:37 -08:00
Leonardo de Moura
19cf5e916b chore(script/gen_constants_cpp): generates a warning if automatically generated C++ function is not used in the source code 2017-02-21 12:05:41 -08:00
Sebastian Ullrich
a053175714 refactor(init/meta,library/vm): use structure for position information 2017-02-21 11:06:39 -08:00
Sebastian Ullrich
bea34eb936 fix(emacs/lean-debug): disable undo in lean-debug buffer 2017-02-21 10:55:15 -08:00
Gabriel Ebner
61c380b534 fix(shell/lean): catch all exceptions 
Fixes #1386.
 2017-02-21 10:54:46 -08:00
Gabriel Ebner
4ae9312a5f fix(emacs/lean-input): remove overloading of \b abbreviation 2017-02-21 10:54:19 -08:00
Leonardo de Moura
613a4d6e45 chore(shell/CMakeFiles): enable SMT2 tests 2017-02-21 10:53:03 -08:00