Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] Annotation of /sml/trunk/HISTORY
ViewVC logotype

Annotation of /sml/trunk/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log


Revision 644 - (view) (download)

1 : dbm 570 This is the HISTORY file for the Yale SML/NJ CVS repository.
2 :    
3 :     An entry should be made for _every_ commit to the repository.
4 :     The entries in this file will be used when creating the README
5 :     for new versions, so keep that in mind when writing the
6 :     description.
7 :    
8 :     The form of an entry should be:
9 :    
10 :     Name:
11 :     Date:
12 :     Tag: <post-commit CVS tag>
13 :     Description:
14 : leunga 591 ----------------------------------------------------------------------
15 : blume 642 Name: Matthias Blume
16 : blume 644 Date: 2000/05/12 18:20:00 JST
17 :     Tag: blume-20000512-ml-build
18 :     Description:
19 :    
20 :     Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
21 :     getting installed properly (by config/install.sh).
22 :    
23 :     ----------------------------------------------------------------------
24 :     Name: Matthias Blume
25 : blume 643 Date: 2000/05/12 17:30:00 JST
26 :     Tag: blume-20000512-anchors
27 :     Description:
28 :    
29 :     !!! NEW BOOT FILES !!!
30 :    
31 :     This change is in preparation of fading out support for "implicitly
32 :     anchored path names". I went through all sources and used the
33 :     explicit (and relatively new) $-notation. See system/README and the
34 :     CM manual for more info on this.
35 :    
36 :     I also modified the anchoring scheme for some things such as "smlnj",
37 :     "MLRISC", "cm", etc. to take advantage of the fact that explicit
38 :     anchors are more expressive: anchor name and first arc do not have to
39 :     coincide. This entails the following user-visible change:
40 :    
41 :     You have to write $smlnj/foo/bar instead of smlnj/foo/bar. In
42 :     particular, when you fire up sml with a command-line argument, say,
43 :     e.g.:
44 :    
45 :     sml '$smlnj/cmb.cm'
46 :    
47 :     At the ML toplevel prompt:
48 :    
49 :     CM.autoload "$smlnj/cmb.cm";
50 :    
51 :     There is also a new controller in CM.Control that can be used to turn
52 :     off all remaining support for implicit anchors by saying:
53 :    
54 :     CM.autoload "$smlnj/
55 :     #set CM.Control.implicit_anchors false;
56 :    
57 :     This causes CM to reject implicitly anchored paths. This is (for the
58 :     time being) less permissive than the "final" version where there will
59 :     be no more such implicit anchors and relative paths will be just that:
60 :     relative.
61 :    
62 :     The next step (version after next version?) will be to make the
63 :     default for CM.Control.implicit_anchors false. After the dust has
64 :     settled, I can then produce the "final" version of this...
65 :    
66 :     Note: Since bootstrapping is a bit tricky, I provided new boot files.
67 :    
68 :     ----------------------------------------------------------------------
69 :     Name: Matthias Blume
70 : blume 642 Date: 2000/05/11 16:30:00 JST
71 :     Tag: blume-20000511-sources
72 :     Description:
73 :    
74 :     The main change is that I added function CM.sources as a generalized
75 :     version of the earlier CM.makedepend. This entails the following
76 :     additional changes:
77 :    
78 :     - CM.makedepend has been dropped.
79 :    
80 :     - CM manual has been updated.
81 :    
82 :     - TOOLS signature and API have been changed.
83 :    
84 :     ----------------------------------------------------------------------
85 : leunga 640 Name: Allen Leung
86 : leunga 641 Date: 2000/05/10 21:17 EDT
87 :     Tag: leunga-20000510-moby-c--ssa
88 :     Description:
89 :    
90 :     Various bug fixes and new features for C--, Moby and MLRISC optimizations.
91 :     None of these affect SML/NJ.
92 :    
93 :     1. Register Allocation
94 :    
95 :     a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
96 :     This module tries to remove local (i.e. basic block level) redundancies
97 :     during spilling.
98 :    
99 :     b. A new framework for performing region based register allocation.
100 :     Not yet entirely functional.
101 :    
102 :     2. X86
103 :    
104 :     a. DefUse for POP was missing the stack pointer [found by Lal]
105 :     b. Reload for CALL was incorrect in X86Spill [found by John]
106 :     c. Various fixes in X86Spill so that it can be used correctly for
107 :     the new spilling module.
108 :    
109 :     3. SSA/IR
110 :    
111 :     a. New module ir/dj-dataflow.sml implements elimination based
112 :     data flow analysis.
113 :    
114 :     4. MLRiscGen
115 :    
116 :     a. Fix for gc type annotation
117 :    
118 :     5. MDGen
119 :    
120 :     Various fixes for machine description -> ml code translation. For ssa
121 :     only.
122 :    
123 :     ----------------------------------------------------------------------
124 :     Name: Allen Leung
125 : leunga 640 Date: 2000/05/08 22:17 EDT
126 :     Tag: leunga-20000508-labexp
127 :     Description:
128 :    
129 :     Fermin has found a few assembly problems with constant expressions
130 :     generated in LabelExp. Mostly, the problems involve extra parentheses,
131 :     which choke on dumb assemblers. This is his fix.
132 :    
133 :     ----------------------------------------------------------------------
134 : dbm 635 Name: Dave MacQueen
135 :     Date: 2000/04/09 14:00 EDT
136 :     Tag: dbm-20000502-Version_110_28
137 :     Description:
138 :    
139 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
140 :    
141 :     2. Updated config/version to 110.28
142 :    
143 :     3. Updated config/srcarchiveurl
144 :    
145 :     3. New boot files!
146 :     ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
147 :    
148 :     ----------------------------------------------------------------------
149 : blume 634 Name: Matthias Blume
150 :     Date: 2000/05/01 19:05:00 JST
151 :     Tag: blume-20000501-noweb
152 :     Description:
153 :    
154 :     A new noweb tool has been added. The existing system is entirely
155 :     unaffected by this, but some CM users have asked for renewed noweb
156 :     support. Everything is documented in the CM manual.
157 :    
158 :     New (plugin) libraries:
159 :    
160 :     noweb-tool.cm
161 :     nw-ext.cm
162 :    
163 :     ----------------------------------------------------------------------
164 : dbm 633 Name: Dave MacQueen
165 :     Date: 2000/04/30 12:40PM EDT
166 :     Tag: dbm-20000430-bug_fixes
167 :     Description:
168 :    
169 :     1. Fix for bug 1498
170 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sig
171 :     smlnj/src/system/Basis/Implementation/Unsafe/object.sml
172 :     added toRealArray function
173 :     smlnj/src/compiler/MiscUtil/print/ppobj.sml
174 :     added check for tag Obj.RealArray to array printing case in ppObj
175 :    
176 :     2. Fix for bug 1510
177 :     smlnj/src/compiler/Semant/types/typesutil.sml
178 :     fixed definition of dummyargs (used by equalTycon) so that
179 :     dummy args are distinct types
180 :    
181 :     ----------------------------------------------------------------------
182 : blume 629 Name: Matthias Blume
183 : blume 632 Date: 2000/04/30 01:00:00 JST
184 :     Tag: blume-20000430-versions
185 :     Description:
186 :    
187 :     1. CM version numbering added. This is an implementation of Lal's
188 :     proposal for adding version numbers and version checking to .cm
189 :     files. Lal said that his proposal was just that -- a proposal.
190 :     For the time being I went ahead and implemented it so that people
191 :     can comment on it. Everything is completely backward-compatible
192 :     (except for the stable library format, i.e., new bootfiles!).
193 :    
194 :     As usual, see the CM manual for details.
195 :    
196 :     2. An alternative syntax for anchored paths has been implemented.
197 :     Dave has recently voiced the same concerns that I had when I did
198 :     this, so there should be some support. My take is that eventually
199 :     I will let support for the current syntax (where anchors are
200 :     "implicit") fade out in favor of the new, explicit syntax.
201 :     In order to be backward-compatible, both old and new syntax are
202 :     currently supported.
203 :    
204 :     Again, see the CM manual for details.
205 :    
206 :     3. Parallel make is trying to be slightly smarter: When the master
207 :     process finds a "bottleneck", i.e., when there is only one
208 :     compilation unit that can be compiled and everybody else is
209 :     waiting on it, then it will simply compile it directly instead
210 :     of clumsily telling one of the slaves to do it.
211 :    
212 :     4. Support for "unsharing" added. This is necessary in order to be
213 :     able to have two different versions of the same library running
214 :     at the same time (e.g., for trying out a new MLRISC while still
215 :     having the old MLRISC linked into the current compiler, etc.)
216 :     See the CM manual.
217 :    
218 :     5. Simple "makedepend" functionality added for generating Makefile
219 :     dependency information. (This is rather crude at the moment.
220 :     Expect some changes here in the future.)
221 :    
222 :     6. ".fun" added as a recognized suffix for ML files. Also documented
223 :     explicitly in the manual that the fallback behavior (unknown suffix
224 :     -> ML file) is not an official feature!
225 :    
226 :     7. Small changes to the pickler for stable libraries.
227 :    
228 :     8. Several internal changes to CM (for cleanup/improvement).
229 :    
230 :    
231 :     !!!! NEW BINFILES !!!!
232 :    
233 :     ----------------------------------------------------------------------
234 :     Name: Matthias Blume
235 : blume 631 Date: 2000/04/28 17:30:00 JST
236 :     Tag: blume-20000428-pathconfig
237 :     Description:
238 :    
239 :     1. I changed config/install.sh to remove duplicate entries from the
240 :     lib/pathconfig file at the end. Moreover, the final version of
241 :     lib/pathconfig is sorted alphabetically. The same (sorting) is done
242 :     in src/system/installml.
243 :    
244 :     2. The config/install.sh script now consistently uses relative
245 :     pathnames in lib/pathconfig whenever the anchor is in the lib
246 :     directory. (So far this was true for the libraries that come
247 :     pre-compiled and bundled as part of the bootfiles but not for
248 :     libraries that are compiled by the script itself.)
249 :    
250 :     ----------------------------------------------------------------------
251 :     Name: Matthias Blume
252 : blume 629 Date: 2000/04/26 13:10:00 JST
253 :     Tag: blume-20000426-fun_suffix
254 :     Description:
255 :    
256 :     Added ".fun" as a recognized file name suffix (for ML code).
257 :    
258 :     ----------------------------------------------------------------------
259 : leunga 628 Name: Allen Leung
260 :     Date: 2000/04/25 17:00:00 EST
261 :     Tag: leunga-20000425-alpha-ra
262 :     Description:
263 :    
264 :     1. Alpha
265 :    
266 :     PSEUDOARITH was missing in AlphaRewrite. This causes an endless loop
267 :     in C--.
268 :    
269 :     2. RA
270 :    
271 :     Added a flag "ra-dump-size" to print out the size of the flowgraph
272 :     and the interference graph.
273 :    
274 :     ----------------------------------------------------------------------
275 : dbm 627 Name: Dave MacQueen
276 :     Date: 2000/04/25/
277 :     Tag: dbm-20000425-mlyacc_doc_examples
278 :     Description:
279 :     Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
280 :     Updated all three examples in src/ml-yacc/examples to run
281 :     under 110.* using CM.make.
282 :    
283 :     ----------------------------------------------------------------------
284 : leunga 624 Name: Allen Leung
285 :     Date: 2000/04/20 23:04:00 EST
286 :     Tag: leunga-20000420-ssa-c---stuff
287 :     Description:
288 :    
289 :     This update synchronizes my repository with Yale's. Most of these
290 :     changes, however, do not affect SML/NJ at all (the RA is an exception).
291 :    
292 :     1. Register Allocator
293 :    
294 :     a. An improvement in the interference graph construction:
295 :     Given a copy
296 :    
297 :     s <- t
298 :    
299 : leunga 628 no interference edge between s and t is added for this definition of s.
300 : leunga 624
301 :     b. I've added two new spill heuristic modules that Fermin and I developed
302 :     (in the new library RA.cm). These are unused in SML/NJ but maybe
303 :     useful for others (Moby?)
304 :    
305 :     2. X86
306 :    
307 :     a. Various fixes in the backend provided by Fermin [C--] and Lal.
308 :    
309 :     3. Alpha
310 :    
311 :     a. Added the BSR instruction and code generation that goes with it [C--]
312 :     b. Other fixes too numerous to recount provided by Fermin [C--]
313 :    
314 :     4. Regmaps
315 :    
316 :     a. The regmaps are not initialized with the identity physical bindings
317 :     at creation time. This is unneeded.
318 :    
319 :     5. MLRISC Optimizations
320 :    
321 :     a. The DJ-Graph module can now compute the iterated dominance frontiers
322 :     intersects with liveness incrementally in linear time! Woohoo!
323 :     This is now used in my new SSA construction algorithm.
324 :    
325 :     b. THe branch reorganization module is now smarter about linear chains of
326 :     basic blocks.
327 :    
328 :    
329 :     ----------------------------------------------------------------------
330 : blume 615 Name: Matthias Blume
331 :     Date: 2000/04/12 13:52:00 JST
332 :     Tag: blume_main_v110p27_1
333 :     Description:
334 :    
335 :     Changed install.sh script to handle archive files without version number
336 :     and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
337 :     name of the boot file archive.
338 :    
339 :     ----------------------------------------------------------------------
340 : dbm 607 Name: Dave MacQueen
341 :     Date: 2000/04/09 14:00 EDT
342 :     Tag: dbm-20000410-Version_110_27
343 :     Description:
344 :    
345 :     1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
346 :    
347 :     2. Updated src/config/version to 110.27
348 :    
349 :     3. New boot files!
350 :    
351 :     ----------------------------------------------------------------------
352 : leunga 606 Name: Allen Leung
353 :     Date: 2000/04/09 19:09:00 EST
354 :     Tag: leunga-20000409-misc
355 :     Description:
356 :    
357 :     1. Yet another fix for x86 assembly for idivl, imull, mull and friends.
358 :    
359 :     2. Miscellaneous improvements to MLRISC (unused in sml/nj)
360 :    
361 :     ----------------------------------------------------------------------
362 : monnier 605 Name: Stefan
363 :     Date: 2000/04/07 10:00:00 EDT
364 :     Tag: monnier-20000406-branch-handling
365 :     Description:
366 :    
367 :     Improved handling of branches (mostly those generated from
368 :     polymorphic equality), removed switchoff and changed the
369 :     default optimization settings (more cpsopt and less flintopt).
370 :    
371 :     ----------------------------------------------------------------------
372 : leunga 601 Name: Allen Leung
373 : leunga 602 Date: 2000/04/06 01:30:00 EST
374 :     Tag: leunga-20000406-peephole-x86-SSA-2
375 :     Description:
376 :    
377 :     Forgot a few files.
378 :    
379 :     ----------------------------------------------------------------------
380 :     Name: Allen Leung
381 : leunga 601 Date: 2000/04/06 00:36:00 EST
382 :     Tag: leunga-20000406-peephole-x86-SSA
383 :     Description:
384 :    
385 :     1. New Peephole code
386 :    
387 :     2. Minor improvement to X86 instruction selection
388 :    
389 :     3. Various fixes to SSA and machine description -> code translator
390 :    
391 :     ----------------------------------------------------------------------
392 : blume 595 Name: Matthias Blume
393 :     Date: 2000/04/05 12:30:00 JST
394 :     Tag: blume_main_v110p26p2_3
395 :     Description:
396 :    
397 :     This update just merges three minor cosmetic updates to CM's sources
398 :     to get ready for the 110.27 code freeze on Friday. No functionality
399 :     has changed.
400 :    
401 :     ----------------------------------------------------------------------
402 : leunga 593 Name: Allen Leung
403 : leunga 594 Date: 2000/04/04 19:39:00 EST
404 :     Tag: leunga-20000404-x86-asm
405 :     Description:
406 :    
407 :     1. Fixed a problem in X86 assembly.
408 :    
409 :     Things like
410 :    
411 :     jmp %eax
412 :     jmp (%eax)
413 :    
414 :     should be output as
415 :    
416 :     jmp *%eax
417 :     jmp *(%eax)
418 :    
419 :     2. Assembly output
420 :    
421 :     Added a new flag
422 :    
423 :     "asm-indent-copies" (default to false)
424 :    
425 :     When this flag is on, parallel copies will be indented an extra level.
426 :    
427 :     ----------------------------------------------------------------------
428 :     Name: Allen Leung
429 : leunga 593 Date: 2000/04/04 03:18:00 EST
430 :     Tag: leunga-20000404-C--Moby
431 :     Description:
432 :    
433 :     All of these fixes are related to C--, Moby, and my own optimization
434 :     stuff; so they shouldn't affect SML/NJ.
435 :    
436 :     1. X86
437 :    
438 :     Various fixes related floating point, and extensions.
439 :    
440 :     2. Alpha
441 :    
442 :     Some extra patterns related to loads with signed/zero extension
443 :     provided by Fermin.
444 :    
445 :     3. Assembly
446 :    
447 :     When generating assemby, resolve the value of client defined constants,
448 :     instead of generating symbolic values. This is controlled by the
449 :     new flag "asm-resolve-constants", which is default to true.
450 :    
451 :     4. Machine Descriptions
452 :    
453 :     a. The precedence parser was slightly broken when parsing infixr symbols.
454 :     b. The type generalizing code had the bound variables reversed, resulting
455 :     in a problem during arity raising.
456 :     c. Various fixes in machine descriptions.
457 :    
458 :     ----------------------------------------------------------------------
459 : blume 592 Name: Matthias Blume
460 :     Date: 2000/04/03 16:05:00 JST
461 :     Tag: blume_main_v110p26p2_2
462 :     Description:
463 :    
464 :     I eliminated coreEnv from compInfo. Access to the "Core" structure is
465 :     now done via the ordinary static environment that is context to each
466 :     compilation unit.
467 :    
468 :     To this end, I arranged that instead of "structure Core" as "structure
469 :     _Core" is bound in the pervasive environment. Core access is done via
470 :     _Core (which can never be accidentially rebound because _Core is not a
471 :     legal surface-syntax symbol).
472 :    
473 :     The current solution is much cleaner because the core environment is
474 :     now simply part of the pervasive environment which is part of every
475 :     compilation unit's context anyway. In particular, this eliminates all
476 :     special-case handling that was necessary until now in order to deal
477 :     with dynamic and symbolic parts of the core environment.
478 :    
479 :     Remaining hackery (to bind the "magic" symbol _Core) is localized in the
480 :     compilation mananger's bootstrap compiler (actually: in the "init group"
481 :     handling). See the comments in src/system/smlnj/init/init.cmi for
482 :     more details.
483 :    
484 :     I also tried to track down all mentions of "Core" (as string argument
485 :     to Symbol.strSymbol) in the compiler and replaced them with a
486 :     reference to the new CoreSym.coreSym. Seems cleaner since the actual
487 :     name appears in one place only.
488 :    
489 :     Binfile and bootfile format have not changed, but the switchover from
490 :     the old "init.cmi" to the new one is a bit tricky, so I supplied new
491 :     bootfiles anyway.
492 :    
493 :     ----------------------------------------------------------------------
494 : leunga 591 Name: Allen Leung
495 :     Date: 2000/04/02 21:17:00 EST
496 :     Tag: leunga-20000402-mltree
497 :     Description:
498 : leunga 585
499 : leunga 591 1. Renamed the constructor CALL in MLTREE by popular demand.
500 :     2. Added a bunch of files from my repository. These are currently
501 :     used by other non-SMLNJ backends.
502 :    
503 : leunga 576 ----------------------------------------------------------------------
504 : leunga 590 Name: Allen Leung
505 :     Date: 2000/03/31 21:15:00 EST
506 :     Tag: leunga-20000331-aliasing
507 :     Description:
508 :    
509 :     This update contains a rewritten (and hopefully more correct) module
510 :     for extracting aliasing information from CPS.
511 :    
512 :     To turn on this feature:
513 :    
514 :     Compiler.Control.CG.memDisambiguate := true
515 :    
516 :     To pretty print the region information with assembly
517 :    
518 :     Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
519 :    
520 :     To control how many levels of aliasing information are printed, use:
521 :    
522 :     Compiler.Control.MLRISC.getInt "points-to-show-level" := n
523 :    
524 :     The default of n is 3.
525 :    
526 :     ----------------------------------------------------------------------
527 : dbm 589 Name: David MacQueen
528 :     Date: 2000/03/31 11:15:00 EST
529 :     Tag: dbm-20000331-runtime_fix
530 :     Description:
531 :    
532 :     This update contains:
533 :    
534 :     1. runtime/c-lib/c-libraries.c
535 :     includes added in revision 1.2 caused compilation errors on hppa-hpux
536 :    
537 :     2. fix for bug 1556
538 :     system/Basis/Implementation/NJ/internal-signals.sml
539 :    
540 :     ----------------------------------------------------------------------
541 : blume 587 Name: Matthias Blume
542 : blume 588 Date: 2000/03/31 18:00:00 JST
543 :     Tag: blume_main_v110p26p2_1
544 :     Description:
545 :    
546 :     This update contains:
547 :    
548 :     1. A small change to CM's handling of stable libraries:
549 :     CM now maintains one "global" modmap that is used for all stable
550 :     libraries. The use of such a global modmap maximizes sharing and
551 :     minimizes the need for re-traversing parts of environments during
552 :     modmap construction. (However, this has minor impact since modmap
553 :     construction seems to account for just one percent or less of total
554 :     compile time.)
555 :    
556 :     2. I added a "genmap" phase to the statistics. This is where I got the
557 :     "one percent" number (see above).
558 :    
559 :     3. CM's new tool parameter mechanism just became _even_ better. :)
560 :     - The parser understands named parameters and recursive options.
561 :     - The "make" and "shell" tools use these new features.
562 :     (This makes it a lot easier to cascade these tools.)
563 :     - There is a small syntax change: named parameters use a
564 :    
565 :     <name> : ( <option> ... ) or
566 :     <name> : <string>
567 :    
568 :     syntax. Previously, named parameters were implemented in an
569 :     ad-hoc fashion by each tool individually (by parsing strings)
570 :     and had the form
571 :    
572 :     <name>=<string>
573 :    
574 :     See the CM manual for a full description of these issues.
575 :    
576 :     ----------------------------------------------------------------------
577 :     Name: Matthias Blume
578 : blume 587 Date: 2000/03/30 18:00:00 JST
579 :     Tag: blume_main_v110p26p2_0
580 :     Description:
581 :    
582 :     !!!!! WARNING !!!!!!
583 :     !! New binfiles !!
584 :     !!!!!!!!!!!!!!!!!!!!
585 :    
586 :     This update contains:
587 :    
588 :     1. Moderate changes to CM:
589 :    
590 :     - Changes to CM's tools mechanism. In particular, it is now possible
591 :     to have tools that accept additional "command line" parameters
592 :     (specified in the .cm file at each instance where the tool's class is
593 :     used).
594 :    
595 :     This was done to accomodate the new "make" and "shell" tools which
596 :     facilitate fairly seemless hookup to portions of code managed using
597 :     Makefiles or Shell scripts.
598 :    
599 :     There are no classes "shared" or "private" anymore. Instead, the
600 :     sharing annotation is now a parameter to the "sml" class.
601 :    
602 :     There is a bit of generic machinery for implementing one's own
603 :     tools that accept command-line parameters. However, I am not yet fully
604 :     satisfied with that part, so expect changes here in the future.
605 :    
606 :     All existing tools are described in the CM manual.
607 :    
608 :     - Slightly better error handling. (CM now surpresses many followup
609 :     error messages that tended to be more annoying than helpful.)
610 :    
611 :     2. Major changes to the compiler's static environment data structures.
612 :    
613 :     - no CMStaticEnv anymore.
614 :     - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
615 :     but it is called Environment), no conversions between different
616 :     kinds of static environments
617 :    
618 :     - There is still a notion of a "modmap", but such modmaps are generated
619 :     on demand at the time when they are needed. This sounds slow, but I
620 :     sped up the code that generates modmaps enough for this not to lead to
621 :     a slowdown of the compiler (at least I didn't detect any).
622 :    
623 :     - To facilitate rapid modmap generation, static environments now
624 :     contain an (optional) "modtree" structure. Modtree annotations are
625 :     constructed by the unpickler during unpickling. (This means that
626 :     the elaborator does not have to worry about modtrees at all.)
627 :     Modtrees have the advantage that they are compositional in the same
628 :     way as the environment data structure itself is compositional.
629 :     As a result, modtrees never hang on to parts of an environment that
630 :     has already been rendered "stale" by filtering or rebinding.
631 :    
632 :     - I went through many, many trials and errors before arriving at the
633 :     current solution. (The initial idea of "linkpaths" did not work.)
634 :     But the result of all this is that I have touched a lot of files that
635 :     depend on the "modules" and "types" data structures (most of the
636 :     elaborator). There were a lot of changes during my "linkpath" trials
637 :     that could have been reverted to their original state but weren't.
638 :     Please, don't be too harsh on me for messing with this code a bit more
639 :     than what was strictly necessary... (I _did_ resist the tempation
640 :     of doing any "global reformatting" to avoid an untimely death at
641 :     Dave's hands. :)
642 :    
643 :     - One positive aspect of the previous point: At least I made sure that
644 :     all files that I touched now compile without warnings (other than
645 :     "polyEqual").
646 :    
647 :     - compiler now tends to run "leaner" (i.e., ties up less memory in
648 :     redundant modmaps)
649 :    
650 :     ----------------------------------------------------------------------
651 : leunga 580 Name: Allen Leung
652 : leunga 585 Date: 2000/03/29 18:00:00
653 :     Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
654 : leunga 586 Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
655 : leunga 585 Description:
656 :    
657 :     This update contains *MAJOR* changes to the way code is generated from CPS
658 :     in the module mlriscGen, and in various backend modules.
659 :    
660 :     CHANGES
661 :     =======
662 :    
663 :     1. MLRiscGen: forward propagation fix.
664 :    
665 :     There was a bug in forward propagation introduced at about the same time
666 :     as the MLRISC x86 backend, which prohibits coalescing to be
667 :     performed effectively in loops.
668 :    
669 :     Effect: speed up of loops in RISC architectures.
670 :     By itself, this actually slowed down certain benchmarks on the x86.
671 :    
672 :     2. MLRiscGen: forward propagating addresses from consing.
673 :    
674 :     I've changed the way consing code is generated. Basically I separated
675 :     out the initialization part:
676 :    
677 :     store tag, offset(allocptr)
678 :     store elem1, offset+4(allocptr)
679 :     store elem2, offset+8(allocptr)
680 :     ...
681 :     store elemn, offset+4n(allocptr)
682 :    
683 :     and the address computation part:
684 :    
685 :     celladdr <- offset+4+alloctpr
686 :    
687 :     and move the address computation part
688 :    
689 :     Effect: register pressure is generally lower as a result. This
690 :     makes compilation of certain expressions much faster, such as
691 :     long lists with non-trivial elements.
692 :    
693 :     [(0,0), (0,0), .... (0,0)]
694 :    
695 :     3. MLRiscGen: base pointer elimination.
696 :    
697 :     As part of the linkage mechanism, we generate the sequence:
698 :    
699 :     L: ... <- start of the code fragment
700 :    
701 :     L1:
702 :     base pointer <- linkreg - L1 + L
703 :    
704 :     The base pointer was then used for computing relocatable addresses
705 :     in the code fragment. Frequently (such as in lots of continuations)
706 :     this is not needed. We now eliminate this sequence whenever possible.
707 :    
708 :     For compile time efficiency, I'm using a very stupid local heuristic.
709 :     But in general, this should be done as a control flow analysis.
710 :    
711 :     Effect: Smaller code size. Speed up of most programs.
712 :    
713 :     4. Hppa back end
714 :    
715 :     Long jumps in span dependence resolution used to depend on the existence
716 :     of the base pointer.
717 :    
718 :     A jump to a long label L was expanded into the following sequence:
719 :    
720 :     LDIL %hi(L-8192), %r29
721 :     LDO %lo(L-8192)(%r29), %r29
722 :     ADD %r29, baseptr, %r29
723 :     BV,n %r0(%r29)
724 :    
725 :     In the presence of change (3) above, this will not work. I've changed
726 :     it so that the following sequence of instructions are generated, which
727 :     doesn't mention the base pointer at all:
728 :    
729 :     BL,n L', %r29 /* branch and link, L' + 4 -> %r29 */
730 :     L': ADDIL L-(L'+4), %r29 /* Compute address of L */
731 :     BV,n %r0(%r29) /* Jump */
732 :    
733 :     5. Alpha back end
734 :    
735 :     New alpha instructions LDB/LDW have been added, as per Fermin's
736 :     suggestions. This is unrelated to all other changes.
737 :    
738 :     6. X86 back end
739 :    
740 :     I've changed andl to testl in the floating point test sequence
741 :     whenever appropriate. The Intel optimization guide states that
742 :     testl is perferable to andl.
743 :    
744 :     7. RA (x86 only)
745 :    
746 :     I've improved the spill propagation algorithm, using an approximation
747 :     of maximal weighted independent sets. This seems to be necessary to
748 :     alleviate the negative effect in light of the slow down in (1).
749 :    
750 :     I'll write down the algorithm one of these days.
751 :    
752 :     8. MLRiscGen: frequencies
753 :    
754 :     I've added an annotation that states that all call gc blocks have zero
755 :     execution frequencies. This improves register allocation on the x86.
756 :    
757 :     BENCHMARKS
758 :     ==========
759 :    
760 :     I've only perform the comparison on 110.25.
761 :    
762 :     The platforms are:
763 :    
764 :     HPPA A four processor HP machine (E9000) with 5G of memory.
765 :     X86 A 300Hhz Pentium II with 128M of memory, and
766 :     SPARC An Ultra sparc 2 with 512M of memory.
767 :    
768 :     I used the following parameters for the SML benchmarks:
769 :    
770 :     @SMLalloc
771 :     HPPA 256k
772 :     SPARC 512k
773 :     X86 256k
774 :    
775 :     COMPILATION TIME
776 :     ----------------
777 :     Here are the numbers comparing the compilation times of the compilers.
778 :     I've only compared 110.25 compiling the new sources versus
779 :     a fixpoint version of the new compiler compiling the same.
780 :    
781 :     110.25 New
782 :     Total Time in RA Spill+Reload Total Time In RA Spill+Reload
783 :     HPPA 627s 116s 2684+3584 599s 95s 1003+1879
784 :     SPARC 892s 173s 2891+3870 708s 116s 1004+1880
785 :     X86 999s 315s 94006+130691 987s 296s 108877+141957
786 :    
787 :     110.25 New
788 :     Code Size Code Size
789 :     HPPA 8596736 8561421
790 :     SPARC 8974299 8785143
791 :     X86 9029180 8716783
792 :    
793 :     So in summary, things are at least as good as before. Dramatic
794 :     reduction in compilation is obtained on the Sparc; I can't explain it,
795 :     but it is reproducible. Perhaps someone should try to reproduce this
796 :     on their own machines.
797 :    
798 :     SML BENCHMARKS
799 :     --------------
800 :    
801 :     On the average, all benchmarks perform at least as well as before.
802 :    
803 :     HPPA Compilation Time Spill+Reload Run Time
804 :     110.25 New 110.25 New 110.25 New
805 :    
806 :     barnesHut 3.158 3.015 4.75% 1+1 0+0 2.980 2.922 2.00%
807 :     boyer 6.152 5.708 7.77% 0+0 0+0 0.218 0.213 2.34%
808 :     count-graphs 1.168 1.120 4.32% 0+0 0+0 22.705 23.073 -1.60%
809 :     fft 0.877 0.792 10.74% 1+3 1+3 0.602 0.587 2.56%
810 :     knuthBendix 3.180 2.857 11.32% 0+0 0+0 0.675 0.662 2.02%
811 :     lexgen 6.190 5.290 17.01% 0+0 0+0 0.913 0.788 15.86%
812 :     life 0.803 0.703 14.22% 25+25 0+0 0.153 0.140 9.52%
813 :     logic 2.048 2.007 2.08% 6+6 1+1 4.133 4.008 3.12%
814 :     mandelbrot 0.077 0.080 -4.17% 0+0 0+0 0.765 0.712 7.49%
815 :     mlyacc 22.932 20.937 9.53% 154+181 32+57 0.468 0.430 8.91%
816 :     nucleic 5.183 5.060 2.44% 2+2 0+0 0.125 0.120 4.17%
817 :     ratio-regions 3.357 3.142 6.84% 0+0 0+0 116.225 113.173 2.70%
818 :     ray 1.283 1.290 -0.52% 0+0 0+0 2.887 2.855 1.11%
819 :     simple 6.307 6.032 4.56% 28+30 5+7 3.705 3.658 1.28%
820 :     tsp 0.888 0.862 3.09% 0+0 0+0 7.040 6.893 2.13%
821 :     vliw 24.378 23.455 3.94% 106+127 25+45 2.758 2.707 1.91%
822 :     --------------------------------------------------------------------------
823 :     Average 6.12% 4.09%
824 :    
825 :     SPARC Compilation Time Spill+Reload Run Time
826 :     110.25 New 110.25 New 110.25 New
827 :    
828 :     barnesHut 3.778 3.592 5.20% 2+2 0+0 3.648 3.453 5.65%
829 :     boyer 6.632 6.110 8.54% 0+0 0+0 0.258 0.242 6.90%
830 :     count-graphs 1.435 1.325 8.30% 0+0 0+0 33.672 34.737 -3.07%
831 :     fft 0.980 0.940 4.26% 3+9 2+6 0.838 0.827 1.41%
832 :     knuthBendix 3.590 3.138 14.39% 0+0 0+0 0.962 0.967 -0.52%
833 :     lexgen 6.593 6.072 8.59% 1+1 0+0 1.077 1.078 -0.15%
834 :     life 0.972 0.868 11.90% 26+26 0+0 0.143 0.140 2.38%
835 :     logic 2.525 2.387 5.80% 7+7 1+1 5.625 5.158 9.05%
836 :     mandelbrot 0.090 0.093 -3.57% 0+0 0+0 0.855 0.728 17.39%
837 :     mlyacc 26.732 23.827 12.19% 162+189 32+57 0.550 0.560 -1.79%
838 :     nucleic 6.233 6.197 0.59% 3+3 0+0 0.163 0.173 -5.77%
839 :     ratio-regions 3.780 3.507 7.79% 0+0 0+0 133.993 131.035 2.26%
840 :     ray 1.595 1.550 2.90% 1+1 0+0 3.440 3.418 0.63%
841 :     simple 6.972 6.487 7.48% 29+32 5+7 3.523 3.525 -0.05%
842 :     tsp 1.115 1.063 4.86% 0+0 0+0 7.393 7.265 1.77%
843 :     vliw 27.765 24.818 11.87% 110+135 25+45 2.265 2.135 6.09%
844 :     ----------------------------------------------------------------------------
845 :     Average 6.94% 2.64%
846 :    
847 :     X86 Compilation Time Spill+Reload Run Time
848 :     110.25 New 110.25 New 110.25 New
849 :    
850 :     barnesHut 5.530 5.420 2.03% 593+893 597+915 3.532 3.440 2.66%
851 :     boyer 8.768 7.747 13.19% 493+199 301+289 0.327 0.297 10.11%
852 :     count-graphs 2.040 2.010 1.49% 298+394 315+457 26.578 28.660 -7.26%
853 :     fft 1.327 1.302 1.92% 112+209 115+210 1.055 0.962 9.71%
854 :     knuthBendix 5.218 5.475 -4.69% 451+598 510+650 0.928 0.932 -0.36%
855 :     lexgen 9.970 9.623 3.60% 1014+841 1157+885 0.947 0.928 1.97%
856 :     life 1.183 1.183 0.00% 162+182 145+148 0.127 0.103 22.58%
857 :     logic 3.285 3.512 -6.45% 514+684 591+836 5.682 5.577 1.88%
858 :     mandelbrot 0.147 0.143 2.33% 38+41 33+54 0.703 0.690 1.93%
859 :     mlyacc 35.457 32.763 8.22% 3496+4564 3611+4860 0.552 0.550 0.30%
860 :     nucleic 7.100 6.888 3.07% 239+168 201+158 0.175 0.173 0.96%
861 :     ratio-regions 6.388 6.843 -6.65% 1182+257 981+300 120.142 120.345 -0.17%
862 :     ray 2.332 2.338 -0.29% 346+398 402+494 3.593 3.540 1.51%
863 :     simple 9.912 9.903 0.08% 1475+941 1579+1168 3.057 3.178 -3.83%
864 :     tsp 1.623 1.532 5.98% 266+200 250+211 8.045 7.878 2.12%
865 :     vliw 33.947 35.470 -4.29% 2629+2774 2877+3171 2.072 1.890 9.61%
866 :     ----------------------------------------------------------------------------
867 :     Average 1.22% 3.36%
868 :    
869 :     ----------------------------------------------------------------------
870 :     Name: Allen Leung
871 : leunga 583 Date: 2000/03/23 16:25:00
872 :     Tag: leunga-20000323-fix_x86_alpha
873 :     Description:
874 :    
875 :     1. X86 fixes/changes
876 :    
877 :     a. The old code generated for SETcc was completely wrong.
878 :     The Intel optimization guide is VERY misleading.
879 :    
880 :     2. ALPHA fixes/changes
881 :    
882 :     a. Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
883 :     b. Added a new mode byteWordLoadStores to the functor parameter to Alpha()
884 :     c. Added reassociation code for address computation.
885 :    
886 :     ----------------------------------------------------------------------
887 :     Name: Allen Leung
888 : leunga 580 Date: 2000/03/22 01:23:00
889 :     Tag: leunga-20000322-fix_x86_hppa_ra
890 :     Description:
891 :    
892 :     1. X86 fixes/changes
893 :    
894 :     a. x86Rewrite bug with MUL3 (found by Lal)
895 :     b. Added the instructions FSTS, FSTL
896 :    
897 :     2. PA-RISC fixes/changes
898 :    
899 :     a. B label should not be a delay slot candidate! Why did this work?
900 :     b. ADDT(32, REG(32, r), LI n) now generates one instruction instead of two,
901 :     as it should be.
902 :     c. The assembly syntax for fstds and fstdd was wrong.
903 :     d. Added the composite instruction COMICLR/LDO, which is the immediate
904 :     operand variant of COMCLR/LDO.
905 :    
906 :     3. Generic MLRISC
907 :    
908 :     a. shuffle.sml rewritten to be slightly more efficient
909 :     b. DIV bug in mltree-simplify fixed (found by Fermin)
910 :    
911 :     4. Register Allocator
912 :    
913 :     a. I now release the interference graph earlier during spilling.
914 :     May improve memory usage.
915 :    
916 :     ----------------------------------------------------------------------
917 : blume 577 Name: Matthias Blume
918 : blume 578 Date: 2000/03/14 14:15:32
919 :     Tag: blume_main_v110p26p1_2
920 :     Description:
921 :    
922 :     1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
923 :     additional argument called "template" which is an optional string that
924 :     specifiel the layout of the tool command line. See the CM manual for
925 :     explanation.
926 :    
927 :     2. A special-purpose tool can be "regisitered" by simply dropping the
928 :     corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
929 :     directory where the .cm file lives that uses this tool. (The
930 :     behavior/misfeature until now was to look for the tool description
931 :     files in the current working directory.) As before, tool description
932 :     files could also be anchored -- in which case they can live anywhere
933 :     they like. Following the recent e-mail discussion, this change should
934 :     make it easier to have special-purpose tools that are shipped together
935 :     with the sources of the program that uses them.
936 :    
937 :     ----------------------------------------------------------------------
938 :     Name: Matthias Blume
939 : blume 577 Date: 2000/03/10 07:48:34
940 :     Tag: blume_main_v110p26p1_1
941 :     Description:
942 :    
943 :     I added a re-written version of Dave's fixpt script to src/system.
944 :     Changes relative to the original version:
945 :     - sh-ified (not everybody has ksh)
946 :     - automatically figures out which architecture it runs on
947 :     - uses ./makeml a bit more cleverly
948 :     - never invokes ./installml (and, thus, does not clobber your
949 :     good and working installation of sml in case something goes wrong)
950 :     - accepts max iteration count using option "-iter <n>"
951 :     - accepts a "base" name using option "-base <base>"
952 :    
953 :     It does not build any extraneous heap images but directly rebuilds
954 :     bin- and boot-hierarchies using makeml's "-rebuild" switch. Finally,
955 :     it can incorporate existing bin- and boot- hierarchies. For example,
956 :     suppose the base is set to "sml" (which is the default). Then it
957 :     successively builds
958 :    
959 :     sml.bin.<arch>-unix and sml.boot.<arch>-unix
960 :     then sml1.bin.<arch>-unix and sml1.boot.<arch>-unix
961 :     then sml2.bin.<arch>-unix and sml2.boot.<arch>-unix
962 :     ...
963 :     then sml<n>.bin.<arch>-unix and sml<n>.boot.<arch>-unix
964 :    
965 :     and so on. If any of these already exist, it will just use what's
966 :     there. In particular, many people will have the initial set of bin
967 :     and boot files around, so this saves time for at least one full
968 :     rebuild. Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
969 :     <k>=1,2,... is normally not a good idea when invoking fixpt. However,
970 :     they might be the result of an earlier partial run of fixpt (which
971 :     perhaps got accidentially killed). In this case, fixpt will quickly
972 :     move through what exists before continuing where it left off earlier,
973 :     and, thus, saves a lot of time.
974 :    
975 :     ----------------------------------------------------------------------
976 : leunga 576 Name: Allen Leung
977 :     Date: 00/03/10 02:20:00
978 :     Tag: leunga-20000310-fix_x86_asm_ra
979 :     Description:
980 : dbm 570
981 : leunga 576 More assembly output problems involving the indexed addressing mode
982 :     on the x86 have been found and corrected. Thanks to Fermin Reig for the
983 :     fix.
984 :    
985 :     The interface and implementation of the register allocator have been changed
986 :     slightly to accommodate the possibility to skip the register allocation
987 :     phases completely and go directly to memory allocation. This is needed
988 :     for C-- use.
989 :    
990 : dbm 570 ----------------------------------------------------------------------
991 : blume 572 Name: Matthias Blume
992 : blume 575 Date: 00/03/09 10:23:53
993 :     Tag: blume_main_v110p26p1_0
994 :     Description:
995 :    
996 :     * Complete re-organization of library names. Many libraries have been
997 :     consolidated so that they share the same path anchor. For example,
998 :     all MLRISC-related libraries are anchored at MLRISC, most libraries that
999 :     are SML/NJ-specific are under "smlnj". Notice that names like
1000 :     host-cmb.cm or host-compiler.cm no longer exist. See system/README
1001 :     for a complete description of the new naming scheme. Quick reference:
1002 :    
1003 :     host-cmb.cm -> smlnj/cmb.cm
1004 :     host-compiler.cm -> smlnj/compiler.cm
1005 :     full-cm.cm -> smlnj/cm.cm
1006 :     <arch>-<os>.cm -> smlnj/cmb/<arch>-<os>.cm
1007 :     <arch>-compiler.cm -> smlnj/compiler/<arch>.cm
1008 :    
1009 :     * Bug fixes in CM.
1010 :     - exceptions in user code are being passed through (i.e., reach top level)
1011 :     - more bugs in paranoia mode fixed
1012 :     - bug related to checking group owners fixed
1013 :    
1014 :     * New install.sh script that automagically fetches archive files:
1015 :     The new file config/srcarchiveurl must contain the URL of the
1016 :     (remote) directory that contains bin files (or other source archives).
1017 :     If install.sh does not find the archive locally, it tries to get
1018 :     it from that remote directory.
1019 :     This should simplify installation further: For machines that have
1020 :     access to the internet, just fetch <version>-config.tgz, unpack it,
1021 :     edit config/targets, and go (run config/install.sh). The scipt will
1022 :     fetch everything else that it might need all by itself.
1023 :    
1024 :     For CVS users, this mechanism is not relevant for source archives, but
1025 :     it is convenient for getting new sets of binfiles.
1026 :    
1027 :     Archives should be tar files compressed with either gzip, compress, or
1028 :     bzip2. The script recognizes .tgz, .tar, tar.gz, tz, .tar.Z, and .tar.bz2.
1029 :    
1030 :     ----------------------------------------------------------------------
1031 :     Name: Matthias Blume
1032 : blume 572 Date: 2000/03/07 04:01:04
1033 :     Tag: blume_main_v110_26_2
1034 : dbm 570 Description:
1035 : blume 572 - size info in BOOTLIST
1036 :     * no fixed upper limits for number of bootfiles or length of
1037 :     bootfile names in runtime
1038 :     * falling back to old behavior if no BOOTLIST size info found
1039 :     - allocation size heuristics in .run-sml
1040 :     * tries to read cache size from /proc/cpuinfo (this is important for
1041 :     small-cache Celeron systems!)
1042 :     - install.sh robustified
1043 :     - CM manual updates
1044 :     - paranoid mode
1045 :     * no more CMB.deliver() (i.e., all done by CMB.make())
1046 :     * can re-use existing sml.boot.* files
1047 :     * init.cmi now treated as library
1048 :     * library stamps for consistency checks
1049 :     - sml.boot.<arch>-<os>/PIDMAP file
1050 :     * This file is read by the CM startup code. This is used to minimize
1051 :     the amount of dynamic state that needs to be stowed away for the
1052 :     purpose of sharing between interactive system and user code.
1053 :     - CM.Anchor.anchor instead of CM.Anchor.{set,cancel}
1054 :     * Upon request by Elsa. Anchors now controlled by get-set-pair
1055 :     like most other CM state variables.
1056 :     - Compiler.CMSA eliminated
1057 :     * No longer supported by CM anyway.
1058 :     - fixed bugs in pickler that kept biting Stefan
1059 :     * past refs to past refs (was caused by the possibility that
1060 :     ad-hoc sharing is more discriminating than hash-cons sharing)
1061 :     * integer overflow on LargeInt.minInt
1062 :     - ml-{lex,yacc} build scripts now use new mechanism
1063 :     for building standalone programs
1064 :     - fixed several gcc -Wall warnings that were caused by missing header
1065 :     files, missing initializations, etc., in runtime (not all warnings
1066 :     eliminated, though)

root@smlnj-gforge.cs.uchicago.edu
ViewVC Help
Powered by ViewVC 1.0.0