-
Notifications
You must be signed in to change notification settings - Fork 1
/
boot-strap-pass-2-b.lisp
693 lines (564 loc) · 26.7 KB
/
boot-strap-pass-2-b.lisp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
; ACL2 Version 8.4 -- A Computational Logic for Applicative Common Lisp
; Copyright (C) 2021, Regents of the University of Texas
; This version of ACL2 is a descendent of ACL2 Version 1.9, Copyright
; (C) 1997 Computational Logic, Inc. See the documentation topic NOTE-2-0.
; This program is free software; you can redistribute it and/or modify
; it under the terms of the LICENSE file distributed with ACL2.
; This program is distributed in the hope that it will be useful,
; but WITHOUT ANY WARRANTY; without even the implied warranty of
; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
; LICENSE for more details.
; Written by: Matt Kaufmann and J Strother Moore
; email: Kaufmann@cs.utexas.edu and Moore@cs.utexas.edu
; Department of Computer Science
; University of Texas at Austin
; Austin, TX 78712 U.S.A.
(in-package "ACL2")
; This is the second of a pair of files; see boot-strap-pass-2-a.lisp for the
; first of the pair.
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;; Support for system-verify-guards
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; This section supports a mechanism for users to extend the set of
; guard-verified functions. They do so in community books under books/system/,
; which are checked when building with feature :acl2-devel, for example
; building with `make' with ACL2_DEVEL=d. But normal builds will not set that
; feature, and will simply trust that functions marked in
; *system-verify-guards-alist* can be guard-verified. For more about this, see
; the comments in *system-verify-guards-alist*.
; A flaw in our approach is that user-supplied guard verifications may depend
; on package axioms. Thus, we view such verifications as strong hints, rather
; than as ironclad guarantees that the functions can be guard-verified in
; definitional (or even conservative) extensions of the ground-zero theory. We
; consider this sufficient, as the event that some package axiom will cause
; such bogus marking as guard-verified seems much less likely than the event
; that our system has other serious bugs!
(verify-termination-boot-strap safe-access-command-tuple-form) ; and guards
(verify-termination-boot-strap acl2-system-namep) ; and guards
(mutual-recursion
(defun system-pseudo-termp (x w)
(declare (xargs :guard (and (pseudo-termp x)
(plist-worldp w))
:mode :logic))
(cond ((atom x) (symbolp x))
((eq (car x) 'quote)
(and (consp (cdr x))
(null (cdr (cdr x)))))
((not (true-listp x)) nil)
((not (system-pseudo-term-listp (cdr x) w)) nil)
(t (if (symbolp (car x))
(acl2-system-namep (car x) w)
(system-pseudo-termp (caddr (car x)) w)))))
(defun system-pseudo-term-listp (lst w)
(declare (xargs :guard (and (pseudo-term-listp lst)
(plist-worldp w))
:mode :logic))
(cond ((atom lst) (equal lst nil))
(t (and (system-pseudo-termp (car lst) w)
(system-pseudo-term-listp (cdr lst) w)))))
)
(defun pair-fns-with-measured-subsets (fns wrld acc)
(declare (xargs :guard (and (symbol-listp fns)
(plist-worldp wrld)
(true-listp acc))))
(cond ((endp fns) (reverse acc))
(t (pair-fns-with-measured-subsets
(cdr fns)
wrld
(cons (let* ((fn (car fns))
(justification (getpropc fn 'justification nil wrld))
(ms0 (and (weak-justification-p justification) ; for guard
(access justification justification
:measure)))
(ms (and ms0
(if (and (pseudo-termp ms0) ; for guard
(system-pseudo-termp ms0 wrld))
ms0
(cons :? (access justification
justification :subset))))))
(cons fn ms))
acc)))))
(defun new-verify-guards-fns1 (wrld installed-wrld acc)
(declare (xargs :guard (and (plist-worldp wrld)
(plist-worldp installed-wrld)
(symbol-listp acc))))
(cond ((or (endp wrld)
(and (eq (caar wrld) 'command-landmark)
(eq (cadar wrld) 'global-value)
(equal (safe-access-command-tuple-form (cddar wrld))
'(exit-boot-strap-mode))))
(pair-fns-with-measured-subsets
(strict-merge-sort-symbol< acc)
installed-wrld
nil))
((and (eq (cadar wrld) 'symbol-class)
(eq (cddar wrld) :COMMON-LISP-COMPLIANT)
(getpropc (caar wrld) 'predefined nil installed-wrld))
(new-verify-guards-fns1 (cdr wrld)
installed-wrld
(cons (caar wrld) acc)))
(t (new-verify-guards-fns1 (cdr wrld) installed-wrld acc))))
(defun new-verify-guards-fns (state)
; It is important for performance that this function be guard-verified, because
; it is called inside an assert-event form in chk-new-verified-guards, which
; causes evaluation to be in safe-mode and would cause evaluation of
; plist-worldp on behalf of guard-checking for new-verify-guards-fns1.
(declare (xargs :stobjs state))
(let ((wrld (w state)))
(new-verify-guards-fns1 wrld wrld nil)))
; An error occurred when moving this below, because arities-okp is to be put
; into :logic mode and calls logicp.
(verify-termination-boot-strap logicp) ; and guards
(defconst *devel-check-book* "system/devel-check")
(defmacro chk-new-verified-guards ()
`(progn (include-book ,*devel-check-book*
:DIR :SYSTEM
:UNCERTIFIED-OKP nil
:DEFAXIOMS-OKP nil
:SKIP-PROOFS-OKP nil
:TTAGS nil)
(assert-event
(equal *system-verify-guards-alist*
(new-verify-guards-fns state))
:msg (msg "ERROR: The set of newly guard-verified functions from ~
the ACL2 community book ~x0 does not match the expected ~
set from the constant ~
*system-verify-guards-alist*.~|~%From the ~
book:~|~X13~|~%Expected from ~
*system-verify-guards-alist*:~|~X23~|"
,*devel-check-book*
(new-verify-guards-fns state)
*system-verify-guards-alist*
nil))
(value-triple :CHK-NEW-VERIFIED-GUARDS-SUCCESS)))
(defun system-verify-guards-aux (fns-alist acc acc1 old-num)
(declare (xargs :guard
(and (alistp fns-alist)
(nat-listp (strip-cars fns-alist)) ; and ordered by >
(symbol-alistp (strip-cdrs fns-alist))
(true-list-listp acc1))
:ruler-extenders :lambdas))
(let* ((num/pair (car fns-alist))
(num (car num/pair))
(entry (cdr num/pair))
(next (and fns-alist ; optimization
`(,(car entry)
,@(let ((measure (cdr entry)))
(and measure
`((declare (xargs :measure
,measure))))))))
(same-num-p (and fns-alist
(eql num old-num)))
(acc (if (or (null acc1) ; true at the top-level
same-num-p)
acc
(list* `(skip-proofs (verify-termination-boot-strap ,@acc1))
; Normally the following verify-guards will be redundant. However, if the
; original defun specified xargs :verify-guards nil, then this will be
; necessary. (An example as of this writing is remove-lambdas.)
`(skip-proofs (verify-guards ,(caar acc1)))
acc))))
(cond ((endp fns-alist) acc)
(t (system-verify-guards-aux
(cdr fns-alist)
acc
(if same-num-p ; accumulate into mutual-recursion
(cons next acc1)
(list next))
num)))))
(defun cons-absolute-event-numbers (fns-alist wrld acc)
(declare (xargs :guard (and (symbol-alistp fns-alist)
(plist-worldp wrld)
(alistp acc))))
(cond ((endp fns-alist) acc)
(t (cons-absolute-event-numbers
(cdr fns-alist)
wrld
(acons (or (getpropc (caar fns-alist) 'absolute-event-number nil
wrld)
(er hard? 'cons-absolute-event-numbers
"The 'absolute-event-number property is missing ~
for ~x0."
(caar fns-alist)))
(car fns-alist)
acc)))))
(defun sort->-absolute-event-number (fns-alist wrld)
(declare (xargs :mode :program)) ; because of merge-sort-car->
(merge-sort-car->
(cons-absolute-event-numbers fns-alist wrld nil)))
(defmacro system-verify-guards (alist)
`(make-event
(let ((events (system-verify-guards-aux
(sort->-absolute-event-number
,alist
(w state))
nil nil nil)))
(list* 'encapsulate
()
'(set-inhibit-warnings "Skip-proofs")
'(set-verify-guards-eagerness 2)
events))))
; Normally we trust *system-verify-guards-alist*, installing guard-verified
; functions with the two calls below of system-verify-guards. But when feature
; :acl2-devel is set, then we do not do so, as we instead intend to run
; (chk-new-verified-guards) in order to check that the effect of
; system-verify-guards is sound. This check is performed by using `make' with
; target devel-check; see the comments in *system-verify-guards-alist*.
; Note that we have two calls of system-verify-guards below. First, we verify
; termination and guards for other than apply$ scions. That is necessary
; before we verify termination and guards for apply$-prim and badge, which is
; done next. Finally, we verify termination and guards for apply$ scions.
#+(and acl2-loop-only ; Note that make-event can't be called here in raw Lisp.
(not acl2-devel))
(system-verify-guards *system-verify-guards-alist-1*)
#-acl2-devel
(verify-termination-boot-strap apply$-prim) ; and guards
#-acl2-devel
(verify-termination-boot-strap badge) ; and guards
#-acl2-devel
(in-theory (disable badge
(:executable-counterpart badge)))
#+(and acl2-loop-only ; Note that make-event can't be called here in raw Lisp.
(not acl2-devel))
(system-verify-guards *system-verify-guards-alist-2*)
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;; Support for system-events
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(defmacro system-event (event)
; (System-event E) expands to (skip-proofs E) during normal builds. However,
; for acl2-devel builds (see discussion under the section "Support for
; system-verify-guards" earlier in this file), we merely add the event to a
; table, to be checked by Make target devel-check, which invokes function
; system-verify-skip-proofs for that purpose.
#+acl2-devel `(table system-event-table ',event t)
; It is tempting to generate a progn, where the skip-proofs is preceded by:
; (value-triple ,(concatenate 'string "Verified in community book " book-name))
; However, that value-triple event doesn't show up with any of :pe, :pcb, or
; :pcb!, so we won't bother.
#-acl2-devel `(skip-proofs ,event))
(defun system-events-fn (events)
(declare (xargs :guard (true-listp events)))
(cond ((endp events) nil)
(t (cons `(system-event ,(car events))
(system-events-fn (cdr events))))))
(defmacro system-events (&rest events)
(declare (xargs :guard t))
(cons 'progn (system-events-fn events)))
(defmacro check-system-events ()
; Executed by "make devel-check".
'(make-event
(let ((event-book-alist (table-alist 'system-event-table (w state))))
`(progn (include-book ,*devel-check-book* :dir :system)
(set-enforce-redundancy t)
,@(strip-cars event-book-alist)
(value-triple :CHECK-SYSTEM-EVENTS-SUCCESS)))))
(system-events ; "system/termp"
(defthm legal-variable-or-constant-namep-implies-symbolp
(implies (not (symbolp x))
(not (legal-variable-or-constant-namep x))))
(in-theory (disable legal-variable-or-constant-namep))
(defthm termp-implies-pseudo-termp
(implies (termp x w)
(pseudo-termp x))
:rule-classes (:rewrite :forward-chaining))
(defthm term-listp-implies-pseudo-term-listp
(implies (term-listp x w)
(pseudo-term-listp x))
:rule-classes (:rewrite :forward-chaining))
(defthm arities-okp-implies-arity
(implies (and (arities-okp user-table w)
(assoc fn user-table))
(equal (arity fn w) (cdr (assoc fn user-table)))))
(defthm arities-okp-implies-logicp
(implies (and (arities-okp user-table w)
(assoc fn user-table))
(logicp fn w)))
(in-theory (disable arity arities-okp logicp))
)
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;; Finishing up with apply$
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; Events in this section must occur after the calls above of
; system-verify-guards. They would naturally belong in apply.lisp were it not
; for the fact that the relevant functions are in :program mode at that point.
; We disable the executable counterparts of tamep because badge-userfn is
; undefined, so running tamep on constants, such as (tamep '(CONS A B)) fails
; and introduces a HIDE. However, expansion of the definitional axioms allow
; us to use the badge properties of warrants.
; Of course, when acl2-devel is true then these functions are in :program mode,
; so it makes no sense to try to disable them.
#-acl2-devel
(in-theory (disable (:executable-counterpart tamep)
(:executable-counterpart tamep-functionp)
(:executable-counterpart suitably-tamep-listp)
ev$
ev$-list
apply$
(:executable-counterpart apply$)))
;;; Finish up on "7. Functional Equivalence"
#-acl2-loop-only
(progn (declaim (ftype (function (t t)
(values t))
apply$))
(declaim (ftype (function (t)
(values t))
tamep-functionp)))
#-acl2-devel ; because apply$ is in :program mode when #+acl2-devel
(defun-sk apply$-equivalence (fn1 fn2)
(declare (xargs :verify-guards t ; avoid make-event in boot-strap
:guard t))
(forall (args)
; We use ec-call to support guard verification in "make proofs".
(equal (ec-call (apply$ fn1 args))
(ec-call (apply$ fn2 args)))))
#-acl2-devel ; because apply$-equivalence is #-acl2-devel
(defun fn-equal (fn1 fn2)
(declare (xargs :mode :logic :guard t))
(if (equal fn1 fn2)
t
(and (tamep-functionp fn1)
(tamep-functionp fn2)
(apply$-equivalence fn1 fn2))))
(system-events ; "system/apply/apply"
(defequiv fn-equal :package :legacy)
)
(system-events ; "system/apply/loop-scions"
(defthm natp-from-to-by-measure
(natp (from-to-by-measure i j))
:rule-classes :type-prescription)
)
(defun mempos (e lst)
; Even though it is not necessary to define mempos in order to build ACL2, but
; because the book books/projects/apply/loop.lisp introduces it to establish
; the loop$-as-correspondence rule, it is best not to let the user define it.
; It is identically defined in community book books/projects/apply/loop.lisp,
; which should be included anytime the user is serious about using scions.
(declare (xargs :guard (true-listp lst)))
(cond ((endp lst) 0)
((equal e (car lst)) 0)
(t (+ 1 (mempos e (cdr lst))))))
#-acl2-devel
(when-pass-2
; The following function symbols must be in :logic mode, and they are, from the
; system-verify-guards calls above. Moreover, we must restrict to pass 2 since
; defwarrant is defined within when-pass-2. Of course, this source file
; (boot-strap-pass-2-b.lisp) is only given to LD in pass 2 anyhow; but by using
; when-pass-2, we don't rely on that and more importantly, we do not evaluate
; these forms in raw Lisp (important because defwarrant is not defined during
; compilation).
; WARNING. Note that proofs are skipped for forms in the scope of when-pass-2.
; So there is a critical invariant: that all of the defwarrant forms below are
; justified in the books. As of this writing in late January 2019, that is the
; case; see the defun$ forms in community book
; books/system/apply/loop-scions.lisp, whose proof obligations are checked when
; running #+acl2-devel builds. Perhaps we should consider using system-events
; for ensuring this invariant.
(defwarrant empty-loop$-as-tuplep)
(defwarrant car-loop$-as-tuple)
(defwarrant cdr-loop$-as-tuple)
(defwarrant loop$-as-ac)
(defwarrant loop$-as)
(defwarrant from-to-by-measure)
(defwarrant tails-ac)
(defwarrant tails)
(defwarrant from-to-by-ac)
(defwarrant from-to-by)
(defwarrant until$-ac)
(defwarrant until$)
(defwarrant until$+-ac)
(defwarrant until$+)
(defwarrant when$-ac)
(defwarrant when$)
(defwarrant when$+-ac)
(defwarrant when$+)
(defwarrant sum$-ac)
(defwarrant sum$)
(defwarrant sum$+-ac)
(defwarrant sum$+)
(defwarrant always$)
(defwarrant always$+)
(defwarrant thereis$)
(defwarrant thereis$+)
(defwarrant collect$-ac)
(defwarrant collect$)
(defwarrant collect$+-ac)
(defwarrant collect$+)
(defwarrant revappend-true-list-fix)
(defwarrant append$-ac)
(defwarrant append$)
(defwarrant append$+-ac)
(defwarrant append$+)
(defwarrant mempos)
)
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;; Memoization
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
#+(and hons acl2-loop-only)
(progn
; We skip raw Lisp functions here; see *thread-unsafe-builtin-memoizations*.
(memoize 'fchecksum-obj :stats nil :forget t)
; We no longer define pkg-names-memoize (other than to cause an error); see the
; comment there.
; (memoize 'pkg-names-memoize :stats nil :forget t)
; Comment on memoizing a worse-than function:
; In Version_7.0 and several preceding versions, we memoized a "worse-than"
; function as follows.
; (memoize 'worse-than-builtin-memoized :stats nil)
; We now use a clocked version of worse-than and avoid such memoization. See
; worse-than-builtin-clocked for comments about potential memoization.
; To restore such memoization, search for every occurrence of
; "Comment on memoizing a worse-than function".
; Below, we discuss some earlier experiments on memoizing worse-than-builtin or
; the like, on two particular community books:
; - books/coi/dtrees/base.lisp does not benefit from memoization of worse-than
; functions, and can be slowed down by it.
; - books/centaur/esim/stv/stv2c/stv2c.lisp requires memoization of
; worse-than-builtin (or the like) to avoid stalling out in a proof.
; Below we look at some results for the first of these books, using Version_7.0
; except (where indicated below) a development version that was close to
; Version_7.0.
; Except where indicated otherwise, we memoized worse-than-builtin for
; experiments below as follows.
; (memoize 'worse-than-builtin
; :stats nil
; :condition ; Sol Swords suggestion
; '(and (nvariablep term1)
; (not (fquotep term1))
; (nvariablep term2)
; (not (fquotep term2))))
; Specifically, we ran the following commands in the above book's directory.
; (ld "cert.acl2")
; (rebuild "base.lisp" t)
; (in-package "DTREE")
; (ubt! 'aux-domain-of-dtreemapfix)
; (skip-proofs (defthm lemma
; (implies (set::in a (aux-domain (dtreemapfix map)))
; (set::in a (aux-domain map)))))
; In some cases we also ran the following command or the following two commands
; after the commands above but before evaluating the defthm shown below:
; (acl2::unmemoize 'acl2::worse-than-builtin)
; #!acl2(memoize 'worse-than-builtin
; :stats nil
; :forget t
; :condition ; Sol Swords suggestion
; '(and (nvariablep term1)
; (not (fquotep term1))
; (nvariablep term2)
; (not (fquotep term2))))
; Then we submitted the following event.
; (defthm lemma2-for-aux-domain-of-dtreemapfix
; (implies (set::in a (aux-domain map))
; (set::in a (aux-domain (dtreemapfix map)))))
; Times and memory use (last two reports from top) from some experiments are
; shown below. The key is that all runs with worse-than unmemoized were
; significantly faster than all runs with worse-than memoized, regardless of
; various attempts to speed up that memoization.
; With GCL, out of the box:
; Time: 122.34 seconds (prove: 122.33, print: 0.01, other: 0.00)
; 11959 kaufmann 20 0 15.3g 2.8g 53m R 100 9.0 2:05.29 gcl-saved_acl2h
; 11959 kaufmann 20 0 15.3g 2.8g 53m S 21 9.0 2:05.92 gcl-saved_acl2h
; With GCL, after the above unmemoize form:
; Time: 78.72 seconds (prove: 78.72, print: 0.00, other: 0.00)
; 11934 kaufmann 20 0 13.2g 854m 53m R 100 2.7 1:18.68 gcl-saved_acl2h
; 11934 kaufmann 20 0 13.2g 854m 53m S 98 2.7 1:21.64 gcl-saved_acl2h
; With GCL, after the above sequence of unmemoize and memoize:
; Time: 94.45 seconds (prove: 94.44, print: 0.01, other: 0.00)
; 11995 kaufmann 20 0 13.8g 727m 53m R 100 2.3 1:35.62 gcl-saved_acl2h
; 11995 kaufmann 20 0 13.8g 727m 53m S 62 2.3 1:37.47 gcl-saved_acl2h
; With CCL, out of the box:
; Time: 131.46 seconds (prove: 131.42, print: 0.04, other: 0.00)
; 12044 kaufmann 20 0 512g 1.8g 17m S 100 5.6 2:10.31 lx86cl64
; 12044 kaufmann 20 0 512g 1.8g 17m S 81 5.6 2:12.73 lx86cl64
; With CCL, after the above unmemoize form:
; Time: 89.83 seconds (prove: 89.82, print: 0.00, other: 0.00)
; 12068 kaufmann 20 0 512g 1.3g 17m S 99 4.2 1:29.91 lx86cl64
; 12068 kaufmann 20 0 512g 1.4g 17m S 40 4.4 1:31.12 lx86cl64
; With CCL, after the above sequence of unmemoize and memoize:
; Time: 147.46 seconds (prove: 147.44, print: 0.02, other: 0.00)
; 12093 kaufmann 20 0 512g 804m 18m S 100 2.5 2:27.86 lx86cl64
; 12093 kaufmann 20 0 512g 1.0g 18m S 30 3.2 2:28.77 lx86cl64
; All of the above were run with EGC off (the default at the time). Now we
; repeat some of the above tests, but after turning EGC on as follows.
; (acl2::value :q) (ccl::egc t) (acl2::lp)
; With CCL, out of the box:
; Time: 1439.72 seconds (prove: 1439.71, print: 0.01, other: 0.00)
; 12127 kaufmann 20 0 512g 3.0g 35m S 100 9.4 23:58.68 lx86cl64
; 12127 kaufmann 20 0 512g 3.0g 35m S 78 9.4 24:01.03 lx86cl64
; With CCL, after the above unmemoize form:
; Time: 87.27 seconds (prove: 87.26, print: 0.01, other: 0.00)
; 12362 kaufmann 20 0 512g 407m 35m S 100 1.3 1:25.72 lx86cl64
; 12362 kaufmann 20 0 512g 417m 35m S 93 1.3 1:28.51 lx86cl64
; With CCL, after the above sequence of unmemoize and memoize:
; Time: 135.92 seconds (prove: 135.90, print: 0.02, other: 0.00)
; 12384 kaufmann 20 0 512g 705m 36m S 0 2.2 2:17.39 lx86cl64
; 12384 kaufmann 20 0 512g 705m 36m S 0 2.2 2:17.40 lx86cl64
; As just above, but after redefining waterfall1 in raw Lisp so that its
; body is (prog2$ (clear-memoize-table 'worse-than-builtin) <old-body>)
; Time: 134.38 seconds (prove: 134.37, print: 0.02, other: 0.00)
; 12631 kaufmann 20 0 512g 691m 36m S 99 2.1 2:14.64 lx86cl64
; 12631 kaufmann 20 0 512g 698m 36m S 38 2.2 2:15.79 lx86cl64
; All of the above used ACL2 Version_7.0. The tests below were run with a
; development copy as of 1/21/2015 (a mere 9 days after the release of 7.0).
; We continue to turn EGC on at the start, as above.
; With CCL, after the above sequence of unmemoize and memoize:
; Time: 135.80 seconds (prove: 135.79, print: 0.01, other: 0.00)
; 13018 kaufmann 20 0 512g 664m 36m S 99 2.1 2:15.91 lx86cl64
; 13018 kaufmann 20 0 512g 671m 36m S 42 2.1 2:17.16 lx86cl64
; With CCL executable built without start-sol-gc (now called
; set-gc-strategy-builtin-delay) and with EGC on, after the above sequence of
; unmemoize and memoize:
; Time: 136.47 seconds (prove: 136.45, print: 0.02, other: 0.00)
; 13049 kaufmann 20 0 512g 59m 36m S 99 0.2 2:14.93 lx86cl64
; 13049 kaufmann 20 0 512g 59m 36m S 96 0.2 2:17.81 lx86cl64
; With CCL executable built without start-sol-gc (now called
; set-gc-strategy-builtin-delay) and with EGC on, after the above unmemoize
; form:
; Time: 86.33 seconds (prove: 86.33, print: 0.01, other: 0.00)
; 13178 kaufmann 20 0 512g 58m 35m S 100 0.2 1:27.18 lx86cl64
; 13178 kaufmann 20 0 512g 58m 35m S 17 0.2 1:27.70 lx86cl64
; With CCL executable built without start-sol-gc (now called
; set-gc-strategy-builtin-delay) and with EGC on, out of the box except for
; redefining waterfall1 in raw Lisp so that its body is (prog2$
; (clear-memoize-table 'worse-than-builtin) <old-body>); notice that :forget
; remains nil.
; Time: 182.61 seconds (prove: 182.58, print: 0.02, other: 0.00)
; 13135 kaufmann 20 0 512g 137m 17m S 100 0.4 3:02.78 lx86cl64
; 13135 kaufmann 20 0 512g 137m 17m S 37 0.4 3:03.90 lx86cl64
)
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;; Deprecation
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(defmacro defdeprecate (old new version)
; Old is the first parameter so that meta-. can find its definition (based on
; defdeprecate).
(declare (type string version)) ; e.g., "8.4"
`(progn
(defmacro ,old (&rest args)
(prog2$
(cw "~%**DEPRECATED** in ACL2 Version ~s0: ~x1. Use ~x2 instead.~|"
,version ',old ',new)
(cons ',new args)))
(add-macro-alias ,old ,new)
(value-triple ',old)))
(defdeprecate symbol-< symbol< "8.4")
(defdeprecate strict-symbol-<-sortedp strict-symbol<-sortedp "8.4")
(defdeprecate strict-merge-symbol-< strict-merge-symbol< "8.4")
(defdeprecate strict-merge-sort-symbol-< strict-merge-sort-symbol< "8.4")
(defdeprecate strict-merge-sort-symbol-<-cdrs strict-merge-sort-symbol<-cdrs "8.4")
(defdeprecate merge-symbol-< merge-symbol< "8.4")
(defdeprecate merge-sort-symbol-< merge-sort-symbol< "8.4")
(defdeprecate logical-defun get-defun-event "8.4")
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;; End
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(deftheory ground-zero
; We want to keep this near the end of *acl2-pass-2-files* in order for the
; ground-zero theory to be as expected; hence the assert$ below.
#-acl2-loop-only ; *acl2-pass-2-files* is only defined in raw Lisp.
(assert (equal (car (last *acl2-pass-2-files*))
"boot-strap-pass-2-b.lisp"))
(current-theory :here))
(deflabel
; WARNING: Put this at the end of the last file in
; *acl2-pass-2-files*.
end-of-pass-2)