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

SCM Repository

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

Diff of /sml/trunk/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 583, Thu Mar 23 21:52:30 2000 UTC revision 744, Fri Dec 8 04:11:42 2000 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Allen Leung
17    Date: Thu Dec  7 22:01:04 EST 2000
18    Tag:  leunga-20001207-cell-monster-hack
19    Description:
20    
21    Major MLRISC internal changes.  Affect all clients.
22    Summary:
23    
24    1.  Type CELLS.cell = int is now replaced by a datatype.
25        As a result, the old regmap is now gone.  Almost all interfaces
26        in MLRISC change as a consequence.
27    
28    2.  A new brand version of machine description tool (v3.0) that generates
29        modules expecting the new interface.  The old version is removed.
30    
31    3.  The RA interface has been further abstracted into two new functors.
32        RISC_RA and X86RA.  These functors have much simpler interfaces.
33        [See also directory MLRISC/demo.]
34    
35    4.  Some other new source->source code generation tools are available:
36    
37        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
38        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
39           I use this tool to generate the peephole optimizers---with the new
40           cell type changes, peephole rules are becoming difficult to write
41           without conditional pattern matching.
42    
43    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
44        cover the entire MLRISC source tree so many things broke.
45    
46    6.  CM files have been moved to the subdirectory MLRISC/cm.
47        They are moved because there are a lot of them and they clutter up the
48        root dir.
49    
50    7.  More detailed documentation to come...
51    
52        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
53        a makeml -rebuild first.  This is because of other other
54        changes that Matthias has made (see below).
55    
56    
57    ----------------------------------------------------------------------
58    Name: Matthias Blume
59    Date: 2000/11/30 23:12:00 JST
60    Tag: blume-20001130-filereorg
61    Description:
62    
63    Some manual updates and some file reorganizations in CM.
64    
65    ----------------------------------------------------------------------
66    Name: Matthias Blume
67    Date: 2000/11/24 17:45:00 JST
68    Tag: blume-20001124-link
69    Description:
70    
71    Drastically improved link traversal code for the case that the dynamic
72    value was already loaded at bootstrap time.  As a result, CM and CMB
73    now both load blazingly fast -- even on a very slow machine.  Also,
74    memory consumption has been further reduced by this.
75    
76    Warning: The format of the PIDMAP file has changed.  THerefore, to
77    bootstrap you have to do this:
78    
79    1. Run CMB.make
80    2. Make a symbolic link for the boot directory:
81         ln -s sml.boot.ARCH-OS xxx
82    3. "Rebuild" the boot directory:
83         ./makeml -boot xxx -rebuild sml ; rm xxx
84    4. Boot normally:
85          ./makeml
86    
87    ----------------------------------------------------------------------
88    Name: Matthias Blume
89    Date: 2000/11/21 21:20:00 JST
90    Tag: blume-20001121-tools
91    Description:
92    
93    Continued hacking on autoloading problem -- with success this time.
94    Also changed tool-plugin mechanism.  See new CM manual.
95    
96    ----------------------------------------------------------------------
97    Name: Matthias Blume
98    Date: 2000/11/19 14:30:00 JST
99    Tag:  blume-20001119-autoload
100    Description:
101    
102    Some hacking to make autoloading faster.  Success for CMB, no success
103    so far for CM.  There is a reduced structure CM' that autoloads faster.
104    (This is a temporary, non-documented hack to be eliminated again when
105    the general problem is solved.)
106    
107    ----------------------------------------------------------------------
108    Name: Matthias Blume
109    Date: 2000/11/17 14:10:00 JST
110    Tag: blume-20001117-pickle-lib
111    Description:
112    
113    1. Eliminated comp-lib.cm
114    2. Made pickle-lib.cm
115    3. Eliminated all uses of intset.sml (from comp-lib.cm)
116    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
117       equivalent constructs from smlnj-lib.cm (INtHashTable).
118    5. Point 4. also goes for those uses of intmap.* in MLRISC.
119       Duplicated intmap modules thrown out.
120    6. Hunted down all duplicated SCC code and replaced it with
121       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
122    7. Rewrote Feedback module.
123    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
124       should be thrown out and equivalent modules from smlnj-lib.cm
125       should be used (IntRedBlackSet, IntListSet, ...).
126    
127    Confirmed that compiler compiles to fixpoint.
128    
129    ----------------------------------------------------------------------
130    Name: Allen Leung
131    Date: 2000/11/10 18:00:00
132    Tag: leunga-20001110-new-x86-fp
133    
134    A new x86 floating point code generator has been added.
135    By default this is turned off.  To turn this on, do:
136    
137        CM.autoload "$smlnj/compiler.cm";
138        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
139    
140    Changes:
141    
142    1.  Changed FTAN to FPTAN so that the assembly output is correct.
143    2.  Changed the extension callback for FTANGENT to generate:
144    
145              fptan
146              fstp  %st(0)
147        instead of
148              fptan
149              fstpl ftempmem
150    
151    3.  Numerous assembly fixes for x86.
152    
153    5.  Cleaned up the machine code output module x86/x86MC.sml and added
154        support for a whole bunch of instructions and addressing modes:
155    
156          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
157          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
158          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
159          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
160          fxch %st(n)
161          fld %st(n)
162          fst %st(n)
163          fst mem
164          fstp %st(n)
165          fucom %st(n)
166          fucomp %st(n)
167    
168        All these are now generated when the fast fp mode is turned on.
169    
170    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
171    
172    ----------------------------------------------------------------------
173    Name: Matthias Blume
174    Date: 2000/11/09 11:20:00 JST
175    Tag: blume-20001109-scc
176    Description:
177    
178    Eliminated some code duplication:
179    
180    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
181       (Otherwise the functor is useless.)
182    2. Used GraphSCCFn where SCCUtilFun was used previously.
183    3. Got rid of SCCUtilFun (in comp-lib.cm).
184    
185    ----------------------------------------------------------------------
186    Name: Lal George
187    Date: 2000/11/06 09:02:21 EST 2000
188    Tag: Release_110_30
189    Description:
190    
191    - Version 110.30
192    ----------------------------------------------------------------------
193    Name: Matthias Blume
194    Date: 2000/11/04 14:45:00
195    Tag: blume-20001104-mlbuild
196    Description:
197    
198    - Made ml-build faster on startup.
199    - Documentation fixes.
200    
201    ----------------------------------------------------------------------
202    Name: Matthias Blume
203    Date: 2000/11/02 17:00:00 JST
204    Tag: blume-20001102-condcomp
205    Description:
206    
207    - Small tweaks to pickler -- new BOOTFILES!
208    - Version bumped to 110.29.2.
209    - Added conditional compilation facility to init.cmi (see comment there).
210    ----------------------------------------------------------------------
211    Name: Allen Leung
212    Date: 2000/10/23 19:31:00
213    Tag: leunga-20001023-demo-ra
214    
215    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
216    2. Test programs for the graph library updated
217    3. Some new MLRISC demo programs added
218    
219    ----------------------------------------------------------------------
220    Name: Matthias Blume
221    Date: 2000/08/31 22:15:00 JST
222    Tag: blume-20001017-errmsg
223    Description:
224    
225    More error message grief: Where there used to be no messages, there
226    now were some that had bogus error regions.  Fixed.
227    
228    ----------------------------------------------------------------------
229    Name: Matthias Blume
230    Date: 2000/08/31 17:30:00 JST
231    Tag: blume-20001017-v110p29p1
232    Description:
233    
234    I made a version 110.29.1 with new bootfiles.
235    
236    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
237              CM documentation changes and a small bugfix in CM's error reporting.
238    
239    ----------------------------------------------------------------------
240    Name: Lal George
241    Date: 2000/09/27 14:42:35 EDT
242    Tag: george-20000927-nodestatus
243    Description:
244    
245    Changed the type of the nodestatus, so that:
246    
247            SPILLED(~1)             is now SPILLED
248            SPILLED(m) where m>=0   is now MEMREG(m)
249            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
250    
251    ----------------------------------------------------------------------
252    Name: Matthias Blume
253    Date: 2000/09/07 14:45:00 JST
254    Tag: blume-20000907-cmerrmsg
255    Description:
256    
257    Small tweak to CM to avoid getting ML syntax error messages twice.
258    
259    ----------------------------------------------------------------------
260    Name: Matthias Blume
261    Date: 2000/08/31 18:00:00 JST
262    Tag: blume-20000831-cvsbootfiles
263    Description:
264    
265    New URL for boot files (because the 110.29 files on the BL server do
266    now work correctly with my updated install scripts for yacc and lex).
267    
268    ----------------------------------------------------------------------
269    Name: Matthias Blume
270    Date: 2000/08/08 12:33:00 JST
271    Tag: blume-20000808-manual
272    Description:
273    
274    Tiny update to CM manual.
275    
276    ----------------------------------------------------------------------
277    Name: Allen Leung
278    Date: 2000/08/7 19:31:00
279    Tag: leunga-20000807-a-whole-bunch-of-stuff
280    
281      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
282    and Mobdy.
283    
284    1.  x86
285    
286       a.  Fixes to peephole module by John and Dan.
287       b.  Assembly fix to SETcc by Allen.
288       c.  Fix to c-call by John.
289       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
290       e.  Instruction selection optimization to SETcc as suggested by John.
291    
292           For example,
293    
294            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
295    
296           should generate:
297    
298            MOVL a, x
299            SUBL b, x
300            SHRL 31, x
301    
302    2.  IR stuff
303    
304         A bunch of new DJ-graph related algorithms added.  These
305         speed up SSA construction.
306    
307    3.  SSA + Scheduling
308    
309         Added code for SSA and scheduling to the repository
310    
311    ----------------------------------------------------------------------
312    Name: Lal George
313    Date: 2000/07/27 11:53:14 EDT
314    
315    Tag: lal-20000727-linux-ppc
316    Description:
317    
318     Made changes to support Linux PPC.
319     p.s. I have confirmation that the 110.29 boot files work fine.
320    
321    ----------------------------------------------------------------------
322    Name: Matthias Blume
323    Date: 2000/07/27 17:40:00 JST
324    Tag: blume-20000727-scripts
325    Description:
326    
327    !!!! WARNING !!!!
328    You must recompile the runtime system!
329    !!!! WARNING !!!!
330    
331    This is basically another round of script-enhancements:
332    
333    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
334       and undefine CM preprocessor symbols.
335    
336    2. ml-build avoids generating a new heap image if it finds that the
337       existing one is still ok.  (The condition is that no ML file had to
338       be recompiled and all ML files are found to be older that the heap
339       file.)
340    
341       To make this work smoothly, I also hacked the runtime system as
342       well as SMLofNJ.SysInfo to get access to the heap image suffix
343       (.sparc-solaris, ...) that is currently being used.
344    
345       Moreover, the signature of CM.mk_standalone has changed.  See the
346       CM manual.
347    
348    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
349       CM manual for details.)
350    
351    4. More CM manual updates:
352        - all of the above has been documented.
353        - there is now a section describing the (CM-related) command line
354          arguments that are accepted by the "sml" command
355    
356    ----------------------------------------------------------------------
357    Name: Matthias Blume
358    Date: 2000/07/25 16:20:00 JST
359    Tag: blume-20000725-makedepend
360    Description:
361    
362    Added a script called ml-makedepend.  This can be used in makefiles
363    for Unix' make in a way very similar to the "makedepend" command for
364    C.
365    
366    The script internally uses function CM.sources.
367    
368    Synopsis:
369    
370        ml-makedepend [-f makefile] cmfile targetname
371    
372    The default for the makefile is "makefile" (or "Makefile" should
373    "makefile" not exist).
374    
375    ml-makedepend adds a cmfile/targetname-specific section to this
376    makefile (after removing the previous version of this section).  The
377    section contains a single dependency specification with targetname on
378    the LHS (targetname is an arbitrary name), and a list of files derived
379    from the cmfile on the RHS.  Some of the files on the RHS are
380    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
381    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
382    path names.  The makefile writer is responsible for making sure that
383    these variables have correct at the time "make" is invoked.
384    
385    ----------------------------------------------------------------------
386    Name: Matthias Blume
387    Date: 2000/07/22 23:30:00 JST
388    Tag: blume-20000722-urlupdate
389    Description:
390    
391    Changed BOOT and config/srcarchiveurl to point to BL server:
392    
393        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
394    
395    ----------------------------------------------------------------------
396    Name: Matthias Blume
397    Date: 2000/07/18 18:00:00 JST
398    Tag: blume-20000718-Version_110_29
399    Description:
400    
401    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
402    
403    2. Updated config/version to 110.29
404    
405    3. Updated config/srcarchiveurl
406    
407    3. New boot files!
408       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
409    
410    ----------------------------------------------------------------------
411    Name: Matthias Blume
412    Date: 2000/07/11 13:58:00 JST
413    Tag: blume-20000711-doctypo
414    Description:
415    
416    Fixed a few typos in CM manual.
417    
418    ----------------------------------------------------------------------
419    Name: Allen Leung
420    Date: 2000/06/15 00:38:00
421    Tag: leunga-20000704-sparc-x86
422    
423    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
424    2. fix to x86 RET bug [found by Dan Grossman]
425    3. sparc assembly bug fix for ticc instructions [found by Fermin]
426    
427       Affects c-- and moby only
428    
429    ----------------------------------------------------------------------
430    Name: Matthias Blume
431    Date: 2000/07/04 15:26:00
432    Tag: blume-20000704-trigger
433    Description:
434    
435    1. Improvements to CM manual.
436    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
437       of getting a back-trace.  The function, when called, raises an
438       internal exception which explicitly carries the full back-trace history,
439       so it is unaffected by any intervening handle-raise pairs ("trivial"
440       or not).  The interactive loop will print that history once it arrives
441       at top level.
442       Short of having all exceptions implicitly carry the full history, the
443       recommended way of using this facility is:
444         - compile your program with instrumentation "on"
445         - run it, when it raises an exception, look at the history
446         - if the history is "cut off" because of some handler, go and modify
447           your program so that it explicitly calls BTrace.trigger
448         - recompile (still instrumented), and rerun; look at the full history
449    
450    ----------------------------------------------------------------------
451    Name: Matthias Blume
452    Date: 2000/07/03 15:36:00 JST
453    Tag: blume-20000702-manual
454    Description:
455    
456    Small corrections and updates to CM manual.
457    
458    ----------------------------------------------------------------------
459    Name: Matthias Blume
460    Date: 2000/06/29 16:04:00 JST
461    Tag: blume-20000629-yacctool
462    Description:
463    
464    Changes:
465    
466    1. Class "mlyacc" now takes separate arguments to pass options to
467       generated .sml- and .sig-files independently.
468    2. Corresponding CM manual updates.
469    3. BTrace module now also reports call sites.  (However, for loop clusters
470       it only shows from where the cluster was entered.)  There are associated
471       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
472    
473    ----------------------------------------------------------------------
474    Name: Matthias Blume
475    Date: 2000/06/27 16:51:00 JST
476    Tag: blume-20000627-noweb
477    Description:
478    
479    Changes:
480    
481     1. Implemented "subdir" and "witness" options for noweb tool.
482        This caused some slight internal changes in CM's tool implementation.
483     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
484        some remaining issues from earlier path anchor changes.
485     3. Updated CM manual accordingly.
486    
487     4. Changed implementation of back-tracing so that I now consider it
488        ready for prime-time.
489    
490        In particular, you don't have to explicitly trigger the back-trace
491        anymore.  Instead, if you are running BTrace-instrumented code and
492        there is an uncaught exception (regardless of whether or not it was
493        raised in instrumented code), the top-level evalloop will print
494        the back-trace.
495    
496        Features:
497    
498          - Instrumented and uninstrumented code work together seemlessly.
499            (Of course, uninstrumented code is never mentioned in actual
500             back-traces.)
501    
502          - Asymptotic time- and space-complexity of instrumented code is
503            equal to that of uninstrumented code.  (This means that
504            tail-recursion is preserved by the instrumentation phase.)
505    
506          - Modules whose code has been instrumented in different sessions
507            work together without problem.
508    
509          - There is no penalty whatsoever on uninstrumented code.
510    
511          - There is no penalty on "raise" expressions, even in
512            instrumented code.
513    
514        A potential bug (or perhaps it is a feature, too):
515    
516          A back-trace reaches no further than the outermost instrumented
517          non-trivial "raise".  Here, a "trivial" raise is one that is the
518          sole RHS of a "handle" rule.  Thus, back-traces reach trough
519    
520               <exp> handle e => raise e
521    
522          and even
523    
524               <exp> handle Foo => raise Bar
525    
526          and, of course, through
527    
528               <exp> handle Foo => ...
529    
530         if the exception was not Foo.
531    
532         Back-traces always reach right through any un-instrumented code
533         including any of its "handle" expressions, trivial or not.
534    
535       To try this out, do the following:
536    
537         - Erase all existing binfiles for your program.
538           (You may keep binfiles for those modules where you think you
539            definitely don't need back-tracing.)
540         - Turn on back-trace instrumentation:
541              SMLofNJ.Internals.BTrace.mode (SOME true);
542         - Recompile your program.  (I.e., run "CM.make" or "use".)
543         - You may now turn instrumentation off again (if you want):
544              SMLofNJ.Internals.BTrace.mode (SOME false);
545         - Run your program as usual.  If it raises an exception that
546           reaches the interactive toplevel, then a back-trace will
547           automatically be printed.  After that, the toplevel loop
548           will print the exception history as usual.
549    
550    ----------------------------------------------------------------------
551    Name: Matthias Blume
552    Date: 2000/06/26 09:56:46 JST
553    Tag: blume-20000626-setup
554    Description:
555    
556    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
557          ML code before and after compiling a file (e.g., to set compiler
558          flags)
559    
560    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
561              - associated changes to btrace.sml (BTrace instrumentation pass)
562              - cleaner implementation of btimp.sml (BTrace tracing and report
563                module)
564    
565    CM manual: * new path encoding documented
566               * description of setup-parameter to "sml" added
567    
568    The biggest user-visible change to back-tracing is that it is no
569    longer necessary to compile all traced modules within the same
570    session.  (This was a real limitation.)
571    
572    ----------------------------------------------------------------------
573    Name: Matthias Blume
574    Date: 2000/06/24 12:40:00 JST
575    Tag: blume-20000624-startup
576    Description:
577    
578    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
579    bit too often -- to put it mildly. :)
580    
581    ----------------------------------------------------------------------
582    Name: Matthias Blume
583    Date: 2000/06/23 18:20:00 JST
584    Tag: blume-20000623-btrace
585    Description:
586    
587    This updates adds a backtrace facility to aid programmers in debugging
588    their programs.  This involves the following changes:
589    
590    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
591       keeping track of the current call stack.  When programs are compiled
592       in a special mode, the compiler will insert calls to these hooks
593       into the user program.
594       "Hook" means that it is possible for different implementations of
595       back-tracing to register themselves (at different times).
596    
597    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
598       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
599       It can be turned on using this call:
600         SMLofNJ.Internals.BTrace.mode (SOME true);
601       Turning it off again:
602         SMLofNJ.Internals.BTrace.mode (SOME false);
603       Querying the current status:
604         SMLofNJ.Internals.BTrace.mode NONE;
605       Annotated programs are about twice as big as normal ones, and they
606       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
607       where all hooks do nothing).  The slowdown with a plugin that is
608       actually useful (such as the one supplied by default) is even greater,
609       but in the case of the default plugin it is still only an constant
610       factor (amortized).
611    
612    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
613       with a sub-structure BTrace for controlling back-tracing.  In particular,
614       the above-mentioned function "mode" controls whether the annotation
615       phase is invoked by the compiler.  Another important function is
616       "trigger": when called it aborts the current execution and causes
617       the top-level loop to print a full back-trace.
618    
619    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
620       for back-tracing.  It keeps track of the dynamic call stack and in
621       addition to that it keeps a partial history at each "level" of that
622       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
623       d and b (at separate times, dynamically), then the report will show:
624    
625       GOTO   d
626             /c
627       GOTO  \b
628       CALL   a
629    
630       This shows that there was an initial non-tail call of a, then a
631       tail-call to b or c, looping behavior in a cluster of functions that
632       consist of b and c, and then a goto from that cluster (i.e., either from
633       b or from c) to d.
634    
635       Note that (depending on the user program) the amount of information
636       that the back-trace module has to keep track of at each level is bounded
637       by a constant.  Thus, the whole implementation has the same asymptotical
638       complexity as the original program (both in space and in time).
639    
640    5. compiler/TopLevel/interact/evalloop.sml has been modified to
641       handle the special exception SMLofNJ.Internals.BTrace.BTrace
642       which is raised by the "trigger" function mentioned above.
643    
644    Notes on usage:
645    
646    - Annotated code works well together with unannotated code:
647    Unannotated calls simply do not show up at all in the backtrace.
648    
649    - It is not a good idea to let modules that were annotated during
650    different sessions run at the same time.  This is because the compiler
651    chooses small integers to identify individual functions, and there
652    will be clashes if different modules were compiled in separate sessions.
653    (Nothing will crash, and you will even be told about the clashes, but
654    back-trace information will in general not be useful.)
655    
656    - Back-tracing can be confused by callcc and capture.
657    
658    - The only way of getting a back-trace right now is to explicitly
659    invoke the "trigger" function from your user program.  Eventually, we
660    should make every exception carry back-trace information (if
661    available).  But since this creates more overhead at "raise"-time
662    (similar to the current exnHistory overhead), I have not yet
663    implemented this.  (The implementation will be rather easy.)  With
664    exceptions carrying back-trace information, this facility will be even
665    more useful because users don't need to modify their programs...
666    
667    - While it is possible to compile the compiler with back-trace
668    annotations turned on (I did it to get some confidence in
669    correctness), you must make absolutely sure that core.sml and
670    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
671    be compiled with annotation because there is no core access yet, but
672    if you compile btimp.sml with annotation, then the system will go into
673    an infinite recursion and crash.)
674    Since CM currently does not know about BTrace, the only way to turn
675    annotations on and off for different modules of the compiler is to
676    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
677    this is awkward and clumsy.
678    
679    Sample sessions:
680    
681    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
682    - SMLofNJ.Internals.BTrace.mode (SOME true);
683    [autoloading]
684    [autoloading done]
685    val it = false : bool
686    - structure X = struct
687    -     fun main n = let
688    -         fun a (x, 0) = d x
689    -           | a (x, n) = b (x, n - 1)
690    -         and b (x, n) = c (x, n)
691    -         and c (x, n) = a (x, n)
692    -         and d x = e (x, 3)
693    -         and e (x, 0) = f x
694    -           | e (x, n) = e (x, n - 1)
695    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
696    -           | f n = n * g (n - 1)
697    -         and g n = a (n, 3)
698    -     in
699    -         f n
700    -     end
701    - end;
702    structure X : sig val main : int -> int end
703    - X.main 3;
704    *** BACK-TRACE ***
705    GOTO   stdIn:4.2-13.20: X.main[2].f
706    GOTO-( stdIn:4.2-13.20: X.main[2].e
707    GOTO   stdIn:4.2-13.20: X.main[2].d
708         / stdIn:4.2-13.20: X.main[2].a
709         | stdIn:4.2-13.20: X.main[2].b
710    GOTO-\ stdIn:4.2-13.20: X.main[2].c
711    CALL   stdIn:4.2-13.20: X.main[2].g
712    GOTO   stdIn:4.2-13.20: X.main[2].f
713    GOTO-( stdIn:4.2-13.20: X.main[2].e
714    GOTO   stdIn:4.2-13.20: X.main[2].d
715         / stdIn:4.2-13.20: X.main[2].a
716         | stdIn:4.2-13.20: X.main[2].b
717    GOTO-\ stdIn:4.2-13.20: X.main[2].c
718    CALL   stdIn:4.2-13.20: X.main[2].g
719    GOTO   stdIn:4.2-13.20: X.main[2].f
720    GOTO-( stdIn:4.2-13.20: X.main[2].e
721    GOTO   stdIn:4.2-13.20: X.main[2].d
722         / stdIn:4.2-13.20: X.main[2].a
723         | stdIn:4.2-13.20: X.main[2].b
724    GOTO-\ stdIn:4.2-13.20: X.main[2].c
725    CALL   stdIn:4.2-13.20: X.main[2].g
726    GOTO   stdIn:4.2-13.20: X.main[2].f
727    CALL   stdIn:2.15-17.4: X.main[2]
728    -
729    
730    (Note that because of a FLINt bug the above code currently does not
731    compile without BTrace turned on.)
732    
733    Here is another example, using my modified Tiger compiler:
734    
735    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
736    - SMLofNJ.Internals.BTrace.mode (SOME true);
737    [autoloading]
738    [autoloading done]
739    val it = false : bool
740    - CM.make "sources.cm";
741    [autoloading]
742    ...
743    [autoloading done]
744    [scanning sources.cm]
745    [parsing (sources.cm):parse.sml]
746    [creating directory CM/SKEL ...]
747    [parsing (sources.cm):tiger.lex.sml]
748    ...
749    [wrote CM/sparc-unix/semant.sml]
750    [compiling (sources.cm):main.sml]
751    [wrote CM/sparc-unix/main.sml]
752    [New bindings added.]
753    val it = true : bool
754    - Main.compile ("../testcases/merge.tig", "foo.out");
755    *** BACK-TRACE ***
756    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
757    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
758    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
759    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
760    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
761    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
762    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
763         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
764    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
765    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
766    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
767    CALL   main.sml:1.18-118.4: Main.compile[2]
768    -
769    
770    ----------------------------------------------------------------------
771    Name: Matthias Blumen
772    Date: 2000/06/21 18:00:00 JST
773    Tag: blume-20000621-manual
774    Description:
775    
776    CM manual update: Path environments documented.
777    
778    ----------------------------------------------------------------------
779    Name: Matthias Blume
780    Date: 2000/06/19 13:40:00
781    Tag: blume-20000619-manual
782    Description:
783    
784    CM manual and system/README update.  This only covers the fact that
785    there are no more implicit anchors.  (Path environments and the "bind"
786    option to "cm" have yet to be documented.)
787    
788    ----------------------------------------------------------------------
789    Name: Matthias Blume
790    Date: 2000/06/19 11:05:00 JST
791    Tag: blume-20000619-chdir-bugfix
792    Description:
793    
794    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
795    
796    ----------------------------------------------------------------------
797    Name: Matthias Blume
798    Date: 2000/06/18 22:00:10 JST
799    Tag: blume-20000618-implicit-anchors-really-gone
800    Description:
801    
802    I updates the previous HISTORY entry where I forgot to mention that
803    implicit anchors are no longer with us.
804    
805    The current update also gets rid of the (now useless) controller
806    CM.Control.implicit_anchors.
807    
808    ----------------------------------------------------------------------
809    Name: Matthias Blume
810    Date: 2000/06/16 17:30:00 JST
811    Tag: blume-20000616-anchorenv
812    Description:
813    
814    This patch implements the long anticipated (just kidding :) "anchor
815    environment" mechanism.  In the course of doing this, I also
816    re-implemented CM's internal "SrcPath" module from scratch.  The new
817    one should be more robust in certain boundary cases.  In any case, it
818    is a lot cleaner than its predecessor (IMHO).
819    
820    This time, although there is yet another boot file format change, I
821    kept the unpickler backward-compatible.  As a result, no new bootfiles
822    are necessary and bootstrapping is straightforward.  (You cannot read
823    new bootfiles into an old system, but the other way around is no
824    problem.)
825    
826    Visible changes:
827    
828    ** 0. Implicit path anchors (without the leading $-symbol) are no
829    longer recognized at all. This means that such path names are not
830    illegal either.  For example, the name basis.cm simply refers to a
831    local file called "basis.cm" (i.e, the name is an ordinary path
832    relative to .cm-files directory).  Or, to put it differently, only
833    names that start with $ are anchored paths.
834    
835    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
836    vanished.
837    
838    John (Reppy) had critizised this as soon as I originally proposed and
839    implemented it, but at that time I did not really deeply believe
840    him. :) Now I came full-circle because I need the $<singlearc> syntax
841    in another place where it cannot be seen as an abbreviation for
842    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
843    seems to mean (i.e., it "expands" into the corresponding anchor
844    value).
845    
846    However, when paths are used as members in CM description files, it
847    continues to be true that there must be at least another arc after the
848    anchor.  This is now enforced separately during semantic analysis
849    (i.e., from a lexical/syntactical point of view, the notation is ok.)
850    
851    ** 2. The "cm" class now accepts an option "bind".  The option's value
852    is a sub-option list of precisely two items -- one labeled "anchor"
853    and the other one labeled "value".  As you might expect, "anchor" is
854    used to specify an anchor name to be bound, and "value" specifies what
855    the anchor is being bound to.
856    
857    The value must be a directory name and can be given in either standard
858    syntax (including the possibility that it is itself an anchored path)
859    or native syntax.
860    
861    Examples:
862    
863       foo.cm (bind:(anchor:bar value:$mystuff/bar))
864       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
865    
866    and so on.
867    
868    The meaning of this is that the .cm-file will be processed with an
869    augmented anchor environment where the given anchor(s) is/are bound to
870    the given values(s).
871    
872    The rationale for having this feature is this: Suppose you are trying
873    to use two different (already stable) libraries a.cm and b.cm (that
874    you perhaps didn't write yourself).  Further, suppose each of these
875    two libraries internally uses its own auxiliary library $aux/lib.cm.
876    Normally you would now have a problem because the anchor "lib" can not
877    be bound to more than one value globally.  Therefore, the project that
878    uses both a.cm and b.cm must locally redirect the anchor to some other
879    place:
880    
881       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
882       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
883    
884    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
885    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
886    
887    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
888    you when you do so at the time of CM.stabilize).  Therefore, you can
889    also use an anchored path as the value:
890    
891      a.cm (bind:(anchor:lib value:$a-lib))
892      b.cm (bind:(anchor:lib value:$b-lib))
893    
894    Now you can globally configure (using the usual CM.Anchor.anchor or
895    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
896    itself is always locally bound, setting it globally is no longer
897    meaningful or necessary (but it does not hurt either).  In fact, "lib"
898    can still be used as a global anchor for separate purposes.  As a
899    matter of fact, one can locally define "lib" in terms of a global
900    "lib":
901    
902      a.cm (bind:(anchor:lib value:$lib/a))
903      b.cm (bind:(anchor:lib value:$lib/b))
904    
905    ** 3: The encoding of path names has changed.  This affects the way
906    path names are shown in CM's progress report and also the internal
907    protocol encoding used for parallel make.
908    
909    The encoding now uses one or more ':'-separated segments.  Each
910    segments corresponds to a file that has been specified relative to the
911    file given by its preceding segment.  The first segment is either
912    relative to the CWD, absolute, or anchored.  Each segment itself is
913    basically a Unix pathname; all segments but the first are relative.
914    
915    Example:
916    
917       $foo/bar/baz.cm:a/b/c.sml
918    
919    This path denotes the file bar/a/b/c.sml relative to the directory
920    denoted by anchor "foo".  Notice that the encoding also includes
921    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
922    paths are resolved relative to the .cm-files directory, so baz.cm must
923    be ignored to get the "real" pathname.
924    
925    To make this fact more obvious, CM puts the names of such "virtual
926    arcs" into parentheses when they appear in progress reports. (No
927    parentheses will appear in the internal protocol encoding.)  Thus,
928    what you really see is:
929    
930      $foo/bar/(baz.cm):a/b/c.sml
931    
932    I find this notation to be much more informative than before.
933    
934    Another new feature of the encoding is that special characters
935    including parentheses, colons, (back)slashes, and white space are
936    written as \ddd (where ddd is the decimal encoding of the character).
937    
938    *** The CM manual still needs to be updated.
939    
940    ----------------------------------------------------------------------
941    Name: Allen Leung
942    Date: 2000/06/15 00:38:00
943    Tag: leunga-20000615-x86-peephole
944    
945    x86 Peephole fix by Fermin.  Affects c-- and moby only.
946    
947    ----------------------------------------------------------------------
948    Name: Matthias Blume
949    Date: 2000/06/12 11:40:00
950    Tag: blume-20000612-parmakefix
951    Description:
952    
953    More cleanup after changing the file naming scheme: This time I
954    repaired the parallel make mechanism for CMB.make which I broke earlier.
955    
956    ----------------------------------------------------------------------
957    Name: Allen Leung
958    Date: 2000/06/09 01:25:00
959    Tag: leunga-20000609-various
960    
961    None of these things should affect normal SML/NJ operations
962    
963    1. Peephole improvements provided by Fermin (c--)
964    2. New annotation DEFUSE for adding extra dependence (moby)
965    3. New X86 LOCK instructions (moby)
966    4. New machine description language for reservation tables (scheduling)
967    5. Fixes to various optimization/analysis modules (branch chaining, dominator
968       trees etc.)
969    6. I've changed the CM files so that they can work with versions
970       110.0.6, 110.25 and 110.28
971    
972    ----------------------------------------------------------------------
973    Name: Matthias Blume
974    Date: 2000/06/09 12:40:00
975    Tag: blume-20000609-log
976    Description:
977    
978    - Removed all(?) remaining RCS Log entries from sources.
979    
980    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
981      anchored paths).
982    
983    ----------------------------------------------------------------------
984    Name: Matthias Blume
985    Date: 2000/06/07 17:00:00 JST
986    Tag: blume-20000607-no-implicit-anchors
987    Description:
988    
989    1. This update changes the default setting for
990    CM.Control.implicit_anchors from true to false.  This means that
991    implicit anchors are no longer permitted by default.  I also tried to
992    make sure that nothing else still relies on implicit anchors.
993    (This is the next step on the schedule towards a CM that does not even
994    have the notion of implicit anchors anymore.)
995    
996    2. More CM manual updates.
997    
998    3. I managed to track down and fix the pickling bug I mentioned last
999    time.  Because of the previously existing workaround, this entails no
1000    immediate practical changes.
1001    
1002    ----------------------------------------------------------------------
1003    Name: Matthias Blume
1004    Date: 2000/06/06 11:15:00 JST
1005    Tag: blume-20000606-lazierpickle
1006    Description:
1007    
1008    !!!! NEW BOOT FILES !!!!
1009    
1010    * The main purpose of this update is to make library pickles lazier in
1011    order to reduce the initial space penalty for autoloading a library.
1012    As a result, it is now possible to have $smlnj/compiler.cm
1013    pre-registered.  This should take care of the many complaints or
1014    inquiries about missing structure Compiler.  This required changes to
1015    CM's internal data structures and small tweaks to some algorithms.
1016    
1017    As a neat additional effect, it is no longer necessary (for the sake
1018    of lean heap image files) to distinguish between a "minimal" CM and a
1019    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1020    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1021    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1022    
1023    To make the life of compiler-hackers easier, "makeml" now also
1024    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1025    words, after you bootstrap a new sml for the first time, you will not
1026    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1027    around you will still have to do it, though.)
1028    
1029    * A second change consists of major updates to the CM manual.  There
1030    are now several appendices with summary information and also a full
1031    specification of the CM description file syntax.
1032    
1033    * In directory src/system I added the script "allcross".  This script
1034    invokes sml and cross-compiles the compiler for all supported
1035    architectures.  (Useful when providing a new set of boot files.)
1036    
1037    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1038    added a small tweak to pickle-util.sml to work around this problem,
1039    but it is not a proper fix yet.  I will investigate further.  (The
1040    effect of the bug was an inflation of library pickle size.)
1041    
1042    * Version number increased to 110.28.1 (to avoid compatibility problems).
1043    
1044    ----------------------------------------------------------------------
1045    Name: Allen Leung
1046    Date: 2000/05/25 17:28 EDT
1047    Tag: leunga-20000525-ra
1048    Description:
1049    
1050      Fixed a bug in freezing phase of the register allocator.
1051    
1052    ----------------------------------------------------------------------
1053    Name: Allen Leung
1054    Date: 2000/05/15 22:53 EDT
1055    Tag: leunga-20000515-alpha-x86-ra
1056    Description:
1057    
1058      1. Alpha
1059    
1060          Slight cleanup.  Removed the instruction SGNXL
1061    
1062      2. X86
1063    
1064          Added the following instructions to the instruction set:
1065    
1066            ROLx, RORx,
1067            BTx, BTSx, BTLx, BTRx,
1068            XCHGx, and variants with the LOCK prefix
1069    
1070      3. Register Allocation
1071    
1072          The module ra-rewrite-with-renaming has been improved.
1073    
1074      These have no effect on SML/NJ.
1075    
1076    ----------------------------------------------------------------------
1077    Name: Matthias Blume
1078    Date: 2000/05/15 16:20:00 JST
1079    Tag: blume-20000515-lightrebuild
1080    Description:
1081    
1082    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1083       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1084       defined.  In effect, the command will not build any cross-compiler
1085       backends and therefore finish more quickly.
1086    
1087       The "fixpt" script also takes a "-light" switch to be able to use
1088       this new facility while compiling for a fixpoint.
1089    
1090    2. I replaced all mentions of anchored paths in group owner specifications
1091       with simple relative paths (usually starting with "..").
1092       The rationale is that a library's internal workings should not be
1093       compromised by the lack of some anchor.  (An anchor is necessary
1094       for someone who wants to refer to the library by an anchored path,
1095       but it should not be necessary to build the same library in the first
1096       place.)
1097    
1098    3. I changed the way CM's tool mechanism determines the shell command
1099       string used for things like ml-yacc etc. so that it does not break
1100       when CM.Control.implicit_anchors is turned off.
1101    
1102    ----------------------------------------------------------------------
1103    Name: Matthias Blume
1104    Date: 2000/05/12 18:20:00 JST
1105    Tag: blume-20000512-ml-build
1106    Description:
1107    
1108    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1109    getting installed properly (by config/install.sh).
1110    
1111    ----------------------------------------------------------------------
1112    Name: Matthias Blume
1113    Date: 2000/05/12 17:30:00 JST
1114    Tag: blume-20000512-anchors
1115    Description:
1116    
1117    !!! NEW BOOT FILES !!!
1118    
1119    This change is in preparation of fading out support for "implicitly
1120    anchored path names".  I went through all sources and used the
1121    explicit (and relatively new) $-notation.  See system/README and the
1122    CM manual for more info on this.
1123    
1124    I also modified the anchoring scheme for some things such as "smlnj",
1125    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1126    anchors are more expressive: anchor name and first arc do not have to
1127    coincide.  This entails the following user-visible change:
1128    
1129    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1130    particular, when you fire up sml with a command-line argument, say,
1131    e.g.:
1132    
1133       sml '$smlnj/cmb.cm'
1134    
1135    At the ML toplevel prompt:
1136    
1137       CM.autoload "$smlnj/cmb.cm";
1138    
1139    There is also a new controller in CM.Control that can be used to turn
1140    off all remaining support for implicit anchors by saying:
1141    
1142        CM.autoload "$smlnj/
1143        #set CM.Control.implicit_anchors false;
1144    
1145    This causes CM to reject implicitly anchored paths.  This is (for the
1146    time being) less permissive than the "final" version where there will
1147    be no more such implicit anchors and relative paths will be just that:
1148    relative.
1149    
1150    The next step (version after next version?) will be to make the
1151    default for CM.Control.implicit_anchors false.  After the dust has
1152    settled, I can then produce the "final" version of this...
1153    
1154    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1155    
1156    ----------------------------------------------------------------------
1157    Name: Matthias Blume
1158    Date: 2000/05/11 16:30:00 JST
1159    Tag: blume-20000511-sources
1160    Description:
1161    
1162    The main change is that I added function CM.sources as a generalized
1163    version of the earlier CM.makedepend.  This entails the following
1164    additional changes:
1165    
1166      - CM.makedepend has been dropped.
1167    
1168      - CM manual has been updated.
1169    
1170      - TOOLS signature and API have been changed.
1171    
1172    ----------------------------------------------------------------------
1173    Name: Allen Leung
1174    Date: 2000/05/10 21:17 EDT
1175    Tag: leunga-20000510-moby-c--ssa
1176    Description:
1177    
1178      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1179    None of these affect SML/NJ.
1180    
1181    1. Register Allocation
1182    
1183        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1184           This module tries to remove local (i.e. basic block level) redundancies
1185           during spilling.
1186    
1187        b. A new framework for performing region based register allocation.
1188           Not yet entirely functional.
1189    
1190    2. X86
1191    
1192       a. DefUse for POP was missing the stack pointer [found by Lal]
1193       b. Reload for CALL was incorrect in X86Spill [found by John]
1194       c. Various fixes in X86Spill so that it can be used correctly for
1195          the new spilling module.
1196    
1197    3. SSA/IR
1198    
1199       a. New module ir/dj-dataflow.sml implements elimination based
1200          data flow analysis.
1201    
1202    4. MLRiscGen
1203    
1204       a. Fix for gc type annotation
1205    
1206    5. MDGen
1207    
1208       Various fixes for machine description -> ml code translation.  For ssa
1209       only.
1210    
1211    ----------------------------------------------------------------------
1212    Name: Allen Leung
1213    Date: 2000/05/08 22:17 EDT
1214    Tag: leunga-20000508-labexp
1215    Description:
1216    
1217      Fermin has found a few assembly problems with constant expressions
1218      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1219      which choke on dumb assemblers.  This is his fix.
1220    
1221    ----------------------------------------------------------------------
1222    Name: Dave MacQueen
1223    Date: 2000/04/09 14:00 EDT
1224    Tag: dbm-20000502-Version_110_28
1225    Description:
1226    
1227    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1228    
1229    2. Updated config/version to 110.28
1230    
1231    3. Updated config/srcarchiveurl
1232    
1233    3. New boot files!
1234       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1235    
1236    ----------------------------------------------------------------------
1237    Name: Matthias Blume
1238    Date: 2000/05/01 19:05:00 JST
1239    Tag: blume-20000501-noweb
1240    Description:
1241    
1242    A new noweb tool has been added.  The existing system is entirely
1243    unaffected by this, but some CM users have asked for renewed noweb
1244    support.  Everything is documented in the CM manual.
1245    
1246    New (plugin) libraries:
1247    
1248       noweb-tool.cm
1249       nw-ext.cm
1250    
1251    ----------------------------------------------------------------------
1252    Name: Dave MacQueen
1253    Date: 2000/04/30 12:40PM EDT
1254    Tag: dbm-20000430-bug_fixes
1255    Description:
1256    
1257    1. Fix for bug 1498
1258       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1259       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1260         added toRealArray function
1261       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1262         added check for tag Obj.RealArray to array printing case in ppObj
1263    
1264    2. Fix for bug 1510
1265       smlnj/src/compiler/Semant/types/typesutil.sml
1266         fixed definition of dummyargs (used by equalTycon) so that
1267         dummy args are distinct types
1268    
1269    ----------------------------------------------------------------------
1270    Name: Matthias Blume
1271    Date: 2000/04/30 01:00:00 JST
1272    Tag: blume-20000430-versions
1273    Description:
1274    
1275    1. CM version numbering added.  This is an implementation of Lal's
1276       proposal for adding version numbers and version checking to .cm
1277       files.  Lal said that his proposal was just that -- a proposal.
1278       For the time being I went ahead and implemented it so that people
1279       can comment on it.  Everything is completely backward-compatible
1280       (except for the stable library format, i.e., new bootfiles!).
1281    
1282       As usual, see the CM manual for details.
1283    
1284    2. An alternative syntax for anchored paths has been implemented.
1285       Dave has recently voiced the same concerns that I had when I did
1286       this, so there should be some support.  My take is that eventually
1287       I will let support for the current syntax (where anchors are
1288       "implicit") fade out in favor of the new, explicit syntax.
1289       In order to be backward-compatible, both old and new syntax are
1290       currently supported.
1291    
1292       Again, see the CM manual for details.
1293    
1294    3. Parallel make is trying to be slightly smarter:  When the master
1295       process finds a "bottleneck", i.e., when there is only one
1296       compilation unit that can be compiled and everybody else is
1297       waiting on it, then it will simply compile it directly instead
1298       of clumsily telling one of the slaves to do it.
1299    
1300    4. Support for "unsharing" added.  This is necessary in order to be
1301       able to have two different versions of the same library running
1302       at the same time (e.g., for trying out a new MLRISC while still
1303       having the old MLRISC linked into the current compiler, etc.)
1304       See the CM manual.
1305    
1306    5. Simple "makedepend" functionality added for generating Makefile
1307       dependency information.  (This is rather crude at the moment.
1308       Expect some changes here in the future.)
1309    
1310    6. ".fun" added as a recognized suffix for ML files. Also documented
1311       explicitly in the manual that the fallback behavior (unknown suffix
1312       -> ML file) is not an official feature!
1313    
1314    7. Small changes to the pickler for stable libraries.
1315    
1316    8. Several internal changes to CM (for cleanup/improvement).
1317    
1318    
1319    !!!! NEW BINFILES !!!!
1320    
1321    ----------------------------------------------------------------------
1322    Name: Matthias Blume
1323    Date: 2000/04/28 17:30:00 JST
1324    Tag: blume-20000428-pathconfig
1325    Description:
1326    
1327    1. I changed config/install.sh to remove duplicate entries from the
1328       lib/pathconfig file at the end.  Moreover, the final version of
1329       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1330       in src/system/installml.
1331    
1332    2. The config/install.sh script now consistently uses relative
1333       pathnames in lib/pathconfig whenever the anchor is in the lib
1334       directory.  (So far this was true for the libraries that come
1335       pre-compiled and bundled as part of the bootfiles but not for
1336       libraries that are compiled by the script itself.)
1337    
1338    ----------------------------------------------------------------------
1339    Name: Matthias Blume
1340    Date: 2000/04/26 13:10:00 JST
1341    Tag: blume-20000426-fun_suffix
1342    Description:
1343    
1344    Added ".fun" as a recognized file name suffix (for ML code).
1345    
1346    ----------------------------------------------------------------------
1347    Name: Allen Leung
1348    Date: 2000/04/25 17:00:00 EST
1349    Tag: leunga-20000425-alpha-ra
1350    Description:
1351    
1352    1. Alpha
1353    
1354        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1355    in C--.
1356    
1357    2. RA
1358    
1359       Added a flag "ra-dump-size" to print out the size of the flowgraph
1360       and the interference graph.
1361    
1362    ----------------------------------------------------------------------
1363    Name: Dave MacQueen
1364    Date: 2000/04/25/
1365    Tag: dbm-20000425-mlyacc_doc_examples
1366    Description:
1367      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1368      Updated all three examples in src/ml-yacc/examples to run
1369      under 110.* using CM.make.
1370    
1371    ----------------------------------------------------------------------
1372    Name: Allen Leung
1373    Date: 2000/04/20 23:04:00 EST
1374    Tag: leunga-20000420-ssa-c---stuff
1375    Description:
1376    
1377      This update synchronizes my repository with Yale's.  Most of these
1378    changes, however, do not affect SML/NJ at all (the RA is an exception).
1379    
1380    1. Register Allocator
1381    
1382       a. An improvement in the interference graph construction:
1383          Given a copy
1384    
1385                s <- t
1386    
1387          no interference edge between s and t is added for this definition of s.
1388    
1389       b. I've added two new spill heuristic modules that Fermin and I developed
1390          (in the new library RA.cm). These are unused in SML/NJ but maybe
1391          useful for others (Moby?)
1392    
1393    2. X86
1394    
1395       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1396    
1397    3. Alpha
1398    
1399       a. Added the BSR instruction and code generation that goes with it [C--]
1400       b. Other fixes too numerous to recount provided by Fermin [C--]
1401    
1402    4. Regmaps
1403    
1404       a. The regmaps are not initialized with the identity physical bindings
1405          at creation time.  This is unneeded.
1406    
1407    5. MLRISC Optimizations
1408    
1409       a. The DJ-Graph module can now compute the iterated dominance frontiers
1410          intersects with liveness incrementally in linear time! Woohoo!
1411          This is now used in my new SSA construction algorithm.
1412    
1413       b. THe branch reorganization module is now smarter about linear chains of
1414          basic blocks.
1415    
1416    
1417    ----------------------------------------------------------------------
1418    Name: Matthias Blume
1419    Date: 2000/04/12 13:52:00 JST
1420    Tag: blume_main_v110p27_1
1421    Description:
1422    
1423    Changed install.sh script to handle archive files without version number
1424    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1425    name of the boot file archive.
1426    
1427    ----------------------------------------------------------------------
1428    Name: Dave MacQueen
1429    Date: 2000/04/09 14:00 EDT
1430    Tag: dbm-20000410-Version_110_27
1431    Description:
1432    
1433    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1434    
1435    2. Updated src/config/version to 110.27
1436    
1437    3. New boot files!
1438    
1439    ----------------------------------------------------------------------
1440    Name: Allen Leung
1441    Date: 2000/04/09 19:09:00 EST
1442    Tag: leunga-20000409-misc
1443    Description:
1444    
1445    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1446    
1447    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1448    
1449    ----------------------------------------------------------------------
1450    Name: Stefan
1451    Date: 2000/04/07 10:00:00 EDT
1452    Tag: monnier-20000406-branch-handling
1453    Description:
1454    
1455    Improved handling of branches (mostly those generated from
1456    polymorphic equality), removed switchoff and changed the
1457    default optimization settings (more cpsopt and less flintopt).
1458    
1459    ----------------------------------------------------------------------
1460    Name: Allen Leung
1461    Date: 2000/04/06 01:30:00 EST
1462    Tag: leunga-20000406-peephole-x86-SSA-2
1463    Description:
1464    
1465       Forgot a few files.
1466    
1467    ----------------------------------------------------------------------
1468    Name: Allen Leung
1469    Date: 2000/04/06 00:36:00 EST
1470    Tag: leunga-20000406-peephole-x86-SSA
1471    Description:
1472    
1473    1.  New Peephole code
1474    
1475    2.  Minor improvement to X86 instruction selection
1476    
1477    3.  Various fixes to SSA and machine description -> code translator
1478    
1479    ----------------------------------------------------------------------
1480    Name: Matthias Blume
1481    Date: 2000/04/05 12:30:00 JST
1482    Tag: blume_main_v110p26p2_3
1483    Description:
1484    
1485    This update just merges three minor cosmetic updates to CM's sources
1486    to get ready for the 110.27 code freeze on Friday.  No functionality
1487    has changed.
1488    
1489    ----------------------------------------------------------------------
1490    Name: Allen Leung
1491    Date: 2000/04/04 19:39:00 EST
1492    Tag: leunga-20000404-x86-asm
1493    Description:
1494    
1495    1.  Fixed a problem in X86 assembly.
1496    
1497        Things like
1498    
1499           jmp %eax
1500           jmp (%eax)
1501    
1502        should be output as
1503    
1504           jmp *%eax
1505           jmp *(%eax)
1506    
1507    2.  Assembly output
1508    
1509          Added a new flag
1510    
1511              "asm-indent-copies" (default to false)
1512    
1513          When this flag is on, parallel copies will be indented an extra level.
1514    
1515    ----------------------------------------------------------------------
1516    Name: Allen Leung
1517    Date: 2000/04/04 03:18:00 EST
1518    Tag: leunga-20000404-C--Moby
1519    Description:
1520    
1521        All of these fixes are related to C--, Moby, and my own optimization
1522        stuff; so they shouldn't affect SML/NJ.
1523    
1524    1.  X86
1525    
1526        Various fixes related floating point, and extensions.
1527    
1528    2.  Alpha
1529    
1530        Some extra patterns related to loads with signed/zero extension
1531        provided by Fermin.
1532    
1533    3.  Assembly
1534    
1535        When generating assembly, resolve the value of client defined constants,
1536        instead of generating symbolic values.  This is controlled by the
1537        new flag "asm-resolve-constants", which is default to true.
1538    
1539    4.  Machine Descriptions
1540    
1541        a. The precedence parser was slightly broken when parsing infixr symbols.
1542        b. The type generalizing code had the bound variables reversed, resulting
1543           in a problem during arity raising.
1544        c. Various fixes in machine descriptions.
1545    
1546    ----------------------------------------------------------------------
1547    Name: Matthias Blume
1548    Date: 2000/04/03 16:05:00 JST
1549    Tag: blume_main_v110p26p2_2
1550    Description:
1551    
1552    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1553    now done via the ordinary static environment that is context to each
1554    compilation unit.
1555    
1556    To this end, I arranged that instead of "structure Core" as "structure
1557    _Core" is bound in the pervasive environment.  Core access is done via
1558    _Core (which can never be accidentally rebound because _Core is not a
1559    legal surface-syntax symbol).
1560    
1561    The current solution is much cleaner because the core environment is
1562    now simply part of the pervasive environment which is part of every
1563    compilation unit's context anyway.  In particular, this eliminates all
1564    special-case handling that was necessary until now in order to deal
1565    with dynamic and symbolic parts of the core environment.
1566    
1567    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1568    compilation manager's bootstrap compiler (actually: in the "init group"
1569    handling).  See the comments in src/system/smlnj/init/init.cmi for
1570    more details.
1571    
1572    I also tried to track down all mentions of "Core" (as string argument
1573    to Symbol.strSymbol) in the compiler and replaced them with a
1574    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1575    name appears in one place only.
1576    
1577    Binfile and bootfile format have not changed, but the switchover from
1578    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1579    bootfiles anyway.
1580    
1581    ----------------------------------------------------------------------
1582    Name: Allen Leung
1583    Date: 2000/04/02 21:17:00 EST
1584    Tag: leunga-20000402-mltree
1585    Description:
1586    
1587       1. Renamed the constructor CALL in MLTREE by popular demand.
1588       2. Added a bunch of files from my repository.  These are currently
1589          used by other non-SMLNJ backends.
1590    
1591    ----------------------------------------------------------------------
1592    Name: Allen Leung
1593    Date: 2000/03/31 21:15:00 EST
1594    Tag: leunga-20000331-aliasing
1595    Description:
1596    
1597    This update contains a rewritten (and hopefully more correct) module
1598    for extracting aliasing information from CPS.
1599    
1600       To turn on this feature:
1601    
1602            Compiler.Control.CG.memDisambiguate := true
1603    
1604       To pretty print the region information with assembly
1605    
1606           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1607    
1608       To control how many levels of aliasing information are printed, use:
1609    
1610           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1611    
1612       The default of n is 3.
1613    
1614    ----------------------------------------------------------------------
1615    Name: David MacQueen
1616    Date: 2000/03/31 11:15:00 EST
1617    Tag: dbm-20000331-runtime_fix
1618    Description:
1619    
1620    This update contains:
1621    
1622    1. runtime/c-lib/c-libraries.c
1623       includes added in revision 1.2 caused compilation errors on hppa-hpux
1624    
1625    2. fix for bug 1556
1626       system/Basis/Implementation/NJ/internal-signals.sml
1627    
1628    ----------------------------------------------------------------------
1629    Name: Matthias Blume
1630    Date: 2000/03/31 18:00:00 JST
1631    Tag: blume_main_v110p26p2_1
1632    Description:
1633    
1634    This update contains:
1635    
1636    1. A small change to CM's handling of stable libraries:
1637       CM now maintains one "global" modmap that is used for all stable
1638       libraries.  The use of such a global modmap maximizes sharing and
1639       minimizes the need for re-traversing parts of environments during
1640       modmap construction.  (However, this has minor impact since modmap
1641       construction seems to account for just one percent or less of total
1642       compile time.)
1643    
1644    2. I added a "genmap" phase to the statistics.  This is where I got the
1645       "one percent" number (see above).
1646    
1647    3. CM's new tool parameter mechanism just became _even_ better. :)
1648       - The parser understands named parameters and recursive options.
1649       - The "make" and "shell" tools use these new features.
1650         (This makes it a lot easier to cascade these tools.)
1651       - There is a small syntax change: named parameters use a
1652    
1653           <name> : ( <option> ... )            or
1654           <name> : <string>
1655    
1656         syntax.  Previously, named parameters were implemented in an
1657         ad-hoc fashion by each tool individually (by parsing strings)
1658         and had the form
1659    
1660           <name>=<string>
1661    
1662       See the CM manual for a full description of these issues.
1663    
1664    ----------------------------------------------------------------------
1665    Name: Matthias Blume
1666    Date: 2000/03/30 18:00:00 JST
1667    Tag: blume_main_v110p26p2_0
1668    Description:
1669    
1670    !!!!! WARNING !!!!!!
1671    !!  New binfiles  !!
1672    !!!!!!!!!!!!!!!!!!!!
1673    
1674    This update contains:
1675    
1676    1. Moderate changes to CM:
1677    
1678       - Changes to CM's tools mechanism.  In particular, it is now possible
1679       to have tools that accept additional "command line" parameters
1680       (specified in the .cm file at each instance where the tool's class is
1681       used).
1682    
1683       This was done to accommodate the new "make" and "shell" tools which
1684       facilitate fairly seamless hookup to portions of code managed using
1685       Makefiles or Shell scripts.
1686    
1687       There are no classes "shared" or "private" anymore.  Instead, the
1688       sharing annotation is now a parameter to the "sml" class.
1689    
1690       There is a bit of generic machinery for implementing one's own
1691       tools that accept command-line parameters.  However, I am not yet fully
1692       satisfied with that part, so expect changes here in the future.
1693    
1694       All existing tools are described in the CM manual.
1695    
1696       - Slightly better error handling.  (CM now suppresses many followup
1697       error messages that tended to be more annoying than helpful.)
1698    
1699    2. Major changes to the compiler's static environment data structures.
1700    
1701       - no CMStaticEnv anymore.
1702            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
1703              but it is called Environment), no conversions between different
1704              kinds of static environments
1705    
1706       - There is still a notion of a "modmap", but such modmaps are generated
1707         on demand at the time when they are needed.  This sounds slow, but I
1708         sped up the code that generates modmaps enough for this not to lead to
1709         a slowdown of the compiler (at least I didn't detect any).
1710    
1711       - To facilitate rapid modmap generation, static environments now
1712         contain an (optional) "modtree" structure.  Modtree annotations are
1713         constructed by the unpickler during unpickling.  (This means that
1714         the elaborator does not have to worry about modtrees at all.)
1715         Modtrees have the advantage that they are compositional in the same
1716         way as the environment data structure itself is compositional.
1717         As a result, modtrees never hang on to parts of an environment that
1718         has already been rendered "stale" by filtering or rebinding.
1719    
1720       - I went through many, many trials and errors before arriving at the
1721         current solution.  (The initial idea of "linkpaths" did not work.)
1722         But the result of all this is that I have touched a lot of files that
1723         depend on the "modules" and "types" data structures (most of the
1724         elaborator). There were a lot of changes during my "linkpath" trials
1725         that could have been reverted to their original state but weren't.
1726         Please, don't be too harsh on me for messing with this code a bit more
1727         than what was strictly necessary...  (I _did_ resist the tempation
1728         of doing any "global reformatting" to avoid an untimely death at
1729         Dave's hands. :)
1730    
1731       - One positive aspect of the previous point:  At least I made sure that
1732         all files that I touched now compile without warnings (other than
1733         "polyEqual").
1734    
1735       - compiler now tends to run "leaner" (i.e., ties up less memory in
1736         redundant modmaps)
1737    
1738    ----------------------------------------------------------------------
1739    Name: Allen Leung
1740    Date: 2000/03/29 18:00:00
1741    Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
1742    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
1743    Description:
1744    
1745       This update contains *MAJOR* changes to the way code is generated from CPS
1746    in the module mlriscGen, and in various backend modules.
1747    
1748    CHANGES
1749    =======
1750    
1751    1. MLRiscGen: forward propagation fix.
1752    
1753       There was a bug in forward propagation introduced at about the same time
1754       as the MLRISC x86 backend, which prohibits coalescing to be
1755       performed effectively in loops.
1756    
1757       Effect: speed up of loops in RISC architectures.
1758               By itself, this actually slowed down certain benchmarks on the x86.
1759    
1760    2. MLRiscGen:  forward propagating addresses from consing.
1761    
1762       I've changed the way consing code is generated.  Basically I separated
1763       out the initialization part:
1764    
1765            store tag,   offset(allocptr)
1766            store elem1, offset+4(allocptr)
1767            store elem2, offset+8(allocptr)
1768            ...
1769            store elemn, offset+4n(allocptr)
1770    
1771       and the address computation part:
1772    
1773            celladdr <- offset+4+alloctpr
1774    
1775       and move the address computation part
1776    
1777       Effect:  register pressure is generally lower as a result.  This
1778                makes compilation of certain expressions much faster, such as
1779                long lists with non-trivial elements.
1780    
1781                 [(0,0), (0,0), .... (0,0)]
1782    
1783    3. MLRiscGen: base pointer elimination.
1784    
1785        As part of the linkage mechanism, we generate the sequence:
1786    
1787         L:  ...  <- start of the code fragment
1788    
1789         L1:
1790             base pointer <- linkreg - L1 + L
1791    
1792         The base pointer was then used for computing relocatable addresses
1793       in the code fragment.  Frequently (such as in lots of continuations)
1794       this is not needed.  We now eliminate this sequence whenever possible.
1795    
1796         For compile time efficiency, I'm using a very stupid local heuristic.
1797       But in general, this should be done as a control flow analysis.
1798    
1799       Effect:  Smaller code size.  Speed up of most programs.
1800    
1801    4. Hppa back end
1802    
1803         Long jumps in span dependence resolution used to depend on the existence
1804      of the base pointer.
1805    
1806         A jump to a long label L was expanded into the following sequence:
1807    
1808          LDIL %hi(L-8192), %r29
1809          LDO  %lo(L-8192)(%r29), %r29
1810          ADD  %r29, baseptr, %r29
1811          BV,n %r0(%r29)
1812    
1813         In the presence of change (3) above, this will not work.  I've changed
1814       it so that the following sequence of instructions are generated, which
1815       doesn't mention the base pointer at all:
1816    
1817             BL,n  L', %r29           /* branch and link, L' + 4 -> %r29 */
1818        L':  ADDIL L-(L'+4), %r29     /* Compute address of L */
1819             BV,n  %r0(%r29)          /* Jump */
1820    
1821    5. Alpha back end
1822    
1823          New alpha instructions LDB/LDW have been added, as per Fermin's
1824       suggestions.   This is unrelated to all other changes.
1825    
1826    6. X86 back end
1827    
1828         I've changed andl to testl in the floating point test sequence
1829         whenever appropriate.  The Intel optimization guide states that
1830         testl is preferable to andl.
1831    
1832    7. RA (x86 only)
1833    
1834         I've improved the spill propagation algorithm, using an approximation
1835       of maximal weighted independent sets.   This seems to be necessary to
1836       alleviate the negative effect in light of the slow down in (1).
1837    
1838         I'll write down the algorithm one of these days.
1839    
1840    8. MLRiscGen: frequencies
1841    
1842         I've added an annotation that states that all call gc blocks have zero
1843       execution frequencies.  This improves register allocation on the x86.
1844    
1845    BENCHMARKS
1846    ==========
1847    
1848       I've only perform the comparison on 110.25.
1849    
1850       The platforms are:
1851    
1852        HPPA  A four processor HP machine (E9000) with 5G of memory.
1853        X86   A 300Hhz Pentium II with 128M of memory, and
1854        SPARC An Ultra sparc 2 with 512M of memory.
1855    
1856       I used the following parameters for the SML benchmarks:
1857    
1858                 @SMLalloc
1859         HPPA    256k
1860         SPARC   512k
1861         X86     256k
1862    
1863    COMPILATION TIME
1864    ----------------
1865       Here are the numbers comparing the compilation times of the compilers.
1866       I've only compared 110.25 compiling the new sources versus
1867       a fixpoint version of the new compiler compiling the same.
1868    
1869                     110.25                                  New
1870               Total  Time in RA  Spill+Reload   Total  Time In RA Spill+Reload
1871         HPPA   627s    116s        2684+3584     599s    95s       1003+1879
1872         SPARC  892s    173s        2891+3870     708s    116s      1004+1880
1873         X86    999s    315s       94006+130691   987s    296s    108877+141957
1874    
1875                   110.25         New
1876                Code Size      Code Size
1877         HPPA   8596736         8561421
1878         SPARC  8974299         8785143
1879         X86    9029180         8716783
1880    
1881       So in summary, things are at least as good as before.   Dramatic
1882       reduction in compilation is obtained on the Sparc; I can't explain it,
1883       but it is reproducible.  Perhaps someone should try to reproduce this
1884       on their own machines.
1885    
1886    SML BENCHMARKS
1887    --------------
1888    
1889        On the average, all benchmarks perform at least as well as before.
1890    
1891          HPPA         Compilation Time     Spill+Reload      Run Time
1892                     110.25  New            110.25    New   110.25  New
1893    
1894          barnesHut  3.158  3.015  4.75%    1+1       0+0   2.980  2.922   2.00%
1895              boyer  6.152  5.708  7.77%    0+0       0+0   0.218  0.213   2.34%
1896       count-graphs  1.168  1.120  4.32%    0+0       0+0  22.705 23.073  -1.60%
1897                fft  0.877  0.792 10.74%    1+3       1+3   0.602  0.587   2.56%
1898        knuthBendix  3.180  2.857 11.32%    0+0       0+0   0.675  0.662   2.02%
1899             lexgen  6.190  5.290 17.01%    0+0       0+0   0.913  0.788  15.86%
1900               life  0.803  0.703 14.22%   25+25      0+0   0.153  0.140   9.52%
1901              logic  2.048  2.007  2.08%    6+6       1+1   4.133  4.008   3.12%
1902         mandelbrot  0.077  0.080 -4.17%    0+0       0+0   0.765  0.712   7.49%
1903             mlyacc 22.932 20.937  9.53%  154+181    32+57  0.468  0.430   8.91%
1904            nucleic  5.183  5.060  2.44%    2+2       0+0   0.125  0.120   4.17%
1905      ratio-regions  3.357  3.142  6.84%    0+0       0+0  116.225 113.173 2.70%
1906                ray  1.283  1.290 -0.52%    0+0       0+0   2.887  2.855   1.11%
1907             simple  6.307  6.032  4.56%   28+30      5+7   3.705  3.658   1.28%
1908                tsp  0.888  0.862  3.09%    0+0       0+0   7.040  6.893   2.13%
1909               vliw 24.378 23.455  3.94%  106+127    25+45  2.758  2.707   1.91%
1910      --------------------------------------------------------------------------
1911       Average                     6.12%                                   4.09%
1912    
1913          SPARC        Compilation Time     Spill+Reload      Run Time
1914                     110.25  New            110.25    New   110.25  New
1915    
1916          barnesHut  3.778  3.592  5.20%    2+2       0+0   3.648  3.453    5.65%
1917              boyer  6.632  6.110  8.54%    0+0       0+0   0.258  0.242    6.90%
1918       count-graphs  1.435  1.325  8.30%    0+0       0+0  33.672 34.737   -3.07%
1919                fft  0.980  0.940  4.26%    3+9       2+6   0.838  0.827    1.41%
1920        knuthBendix  3.590  3.138 14.39%    0+0       0+0   0.962  0.967   -0.52%
1921             lexgen  6.593  6.072  8.59%    1+1       0+0   1.077  1.078   -0.15%
1922               life  0.972  0.868 11.90%   26+26      0+0   0.143  0.140    2.38%
1923              logic  2.525  2.387  5.80%    7+7       1+1   5.625  5.158    9.05%
1924         mandelbrot  0.090  0.093 -3.57%    0+0       0+0   0.855  0.728   17.39%
1925             mlyacc 26.732 23.827 12.19%  162+189    32+57  0.550  0.560   -1.79%
1926            nucleic  6.233  6.197  0.59%    3+3       0+0   0.163  0.173   -5.77%
1927      ratio-regions  3.780  3.507  7.79%    0+0       0+0 133.993 131.035   2.26%
1928                ray  1.595  1.550  2.90%    1+1       0+0   3.440  3.418    0.63%
1929             simple  6.972  6.487  7.48%   29+32      5+7   3.523  3.525   -0.05%
1930                tsp  1.115  1.063  4.86%    0+0       0+0   7.393  7.265    1.77%
1931               vliw 27.765 24.818 11.87%  110+135    25+45  2.265  2.135    6.09%
1932      ----------------------------------------------------------------------------
1933       Average                     6.94%                                    2.64%
1934    
1935          X86          Compilation Time     Spill+Reload      Run Time
1936                     110.25  New            110.25    New   110.25  New
1937    
1938          barnesHut  5.530  5.420  2.03%  593+893   597+915   3.532  3.440   2.66%
1939              boyer  8.768  7.747 13.19%  493+199   301+289   0.327  0.297  10.11%
1940       count-graphs  2.040  2.010  1.49%  298+394   315+457  26.578 28.660  -7.26%
1941                fft  1.327  1.302  1.92%  112+209   115+210   1.055  0.962   9.71%
1942        knuthBendix  5.218  5.475 -4.69%  451+598   510+650   0.928  0.932  -0.36%
1943             lexgen  9.970  9.623  3.60% 1014+841  1157+885   0.947  0.928   1.97%
1944               life  1.183  1.183  0.00%  162+182   145+148   0.127  0.103  22.58%
1945              logic  3.285  3.512 -6.45%  514+684   591+836   5.682  5.577   1.88%
1946         mandelbrot  0.147  0.143  2.33%   38+41     33+54    0.703  0.690   1.93%
1947             mlyacc 35.457 32.763  8.22% 3496+4564 3611+4860  0.552  0.550   0.30%
1948            nucleic  7.100  6.888  3.07%  239+168   201+158   0.175  0.173   0.96%
1949      ratio-regions  6.388  6.843 -6.65% 1182+257   981+300  120.142 120.345 -0.17%
1950                ray  2.332  2.338 -0.29%  346+398   402+494   3.593  3.540   1.51%
1951             simple  9.912  9.903  0.08% 1475+941  1579+1168  3.057  3.178  -3.83%
1952                tsp  1.623  1.532  5.98%  266+200   250+211   8.045  7.878   2.12%
1953               vliw 33.947 35.470 -4.29% 2629+2774 2877+3171  2.072  1.890   9.61%
1954      ----------------------------------------------------------------------------
1955       Average                     1.22%                                     3.36%
1956    
1957  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1958  Name: Allen Leung  Name: Allen Leung
1959  Date: 2000/03/23 16:25:00  Date: 2000/03/23 16:25:00
# Line 66  Line 2009 
2009    
2010  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2011  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2012  specifiel the layout of the tool command line.  See the CM manual for  specifies the layout of the tool command line.  See the CM manual for
2013  explanation.  explanation.
2014    
2015  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2016  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2017  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2018  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 113  Line 2056 
2056  rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for  rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
2057  <k>=1,2,... is normally not a good idea when invoking fixpt.  However,  <k>=1,2,... is normally not a good idea when invoking fixpt.  However,
2058  they might be the result of an earlier partial run of fixpt (which  they might be the result of an earlier partial run of fixpt (which
2059  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2060  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2061  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2062    
# Line 163  Line 2106 
2106    it from that remote directory.    it from that remote directory.
2107    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2108    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2109    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2110    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2111    
2112    For CVS users, this mechanism is not relevant for source archives, but    For CVS users, this mechanism is not relevant for source archives, but

Legend:
Removed from v.583  
changed lines
  Added in v.744

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