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 580, Wed Mar 22 06:33:52 2000 UTC revision 752, Fri Dec 8 23:32:37 2000 UTC
# Line 13  Line 13 
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15  Name: Allen Leung  Name: Allen Leung
16    Date: Fri Dec  8 19:23:26 EST 2000
17    Description:
18    Tag:  leunga-20001208-nowhere
19    Description:
20    
21      Slight improvements to the 'nowhere' tool to handle OR-patterns,
22    to generate better error messages etc.  Plus a brief manual.
23    
24    ----------------------------------------------------------------------
25    Name: Lal George
26    Date: 2000/12/08 09:54:02 EST 2000
27    Tag: Release_110_31
28    Description:
29    
30    - Version 110.31
31    ----------------------------------------------------------------------
32    Name: Allen Leung
33    Date: Thu Dec  7 22:01:04 EST 2000
34    Tag:  leunga-20001207-cell-monster-hack
35    Description:
36    
37    Major MLRISC internal changes.  Affect all clients.
38    Summary:
39    
40    1.  Type CELLS.cell = int is now replaced by a datatype.
41        As a result, the old regmap is now gone.  Almost all interfaces
42        in MLRISC change as a consequence.
43    
44    2.  A new brand version of machine description tool (v3.0) that generates
45        modules expecting the new interface.  The old version is removed.
46    
47    3.  The RA interface has been further abstracted into two new functors.
48        RISC_RA and X86RA.  These functors have much simpler interfaces.
49        [See also directory MLRISC/demo.]
50    
51    4.  Some other new source->source code generation tools are available:
52    
53        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
54        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
55           I use this tool to generate the peephole optimizers---with the new
56           cell type changes, peephole rules are becoming difficult to write
57           without conditional pattern matching.
58    
59    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
60        cover the entire MLRISC source tree so many things broke.
61    
62    6.  CM files have been moved to the subdirectory MLRISC/cm.
63        They are moved because there are a lot of them and they clutter up the
64        root dir.
65    
66    7.  More detailed documentation to come...
67    
68        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
69        a makeml -rebuild first.  This is because of other other
70        changes that Matthias has made (see below).
71    
72    
73    ----------------------------------------------------------------------
74    Name: Matthias Blume
75    Date: 2000/11/30 23:12:00 JST
76    Tag: blume-20001130-filereorg
77    Description:
78    
79    Some manual updates and some file reorganizations in CM.
80    
81    ----------------------------------------------------------------------
82    Name: Matthias Blume
83    Date: 2000/11/24 17:45:00 JST
84    Tag: blume-20001124-link
85    Description:
86    
87    Drastically improved link traversal code for the case that the dynamic
88    value was already loaded at bootstrap time.  As a result, CM and CMB
89    now both load blazingly fast -- even on a very slow machine.  Also,
90    memory consumption has been further reduced by this.
91    
92    Warning: The format of the PIDMAP file has changed.  THerefore, to
93    bootstrap you have to do this:
94    
95    1. Run CMB.make
96    2. Make a symbolic link for the boot directory:
97         ln -s sml.boot.ARCH-OS xxx
98    3. "Rebuild" the boot directory:
99         ./makeml -boot xxx -rebuild sml ; rm xxx
100    4. Boot normally:
101          ./makeml
102    
103    ----------------------------------------------------------------------
104    Name: Matthias Blume
105    Date: 2000/11/21 21:20:00 JST
106    Tag: blume-20001121-tools
107    Description:
108    
109    Continued hacking on autoloading problem -- with success this time.
110    Also changed tool-plugin mechanism.  See new CM manual.
111    
112    ----------------------------------------------------------------------
113    Name: Matthias Blume
114    Date: 2000/11/19 14:30:00 JST
115    Tag:  blume-20001119-autoload
116    Description:
117    
118    Some hacking to make autoloading faster.  Success for CMB, no success
119    so far for CM.  There is a reduced structure CM' that autoloads faster.
120    (This is a temporary, non-documented hack to be eliminated again when
121    the general problem is solved.)
122    
123    ----------------------------------------------------------------------
124    Name: Matthias Blume
125    Date: 2000/11/17 14:10:00 JST
126    Tag: blume-20001117-pickle-lib
127    Description:
128    
129    1. Eliminated comp-lib.cm
130    2. Made pickle-lib.cm
131    3. Eliminated all uses of intset.sml (from comp-lib.cm)
132    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
133       equivalent constructs from smlnj-lib.cm (INtHashTable).
134    5. Point 4. also goes for those uses of intmap.* in MLRISC.
135       Duplicated intmap modules thrown out.
136    6. Hunted down all duplicated SCC code and replaced it with
137       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
138    7. Rewrote Feedback module.
139    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
140       should be thrown out and equivalent modules from smlnj-lib.cm
141       should be used (IntRedBlackSet, IntListSet, ...).
142    
143    Confirmed that compiler compiles to fixpoint.
144    
145    ----------------------------------------------------------------------
146    Name: Allen Leung
147    Date: 2000/11/10 18:00:00
148    Tag: leunga-20001110-new-x86-fp
149    
150    A new x86 floating point code generator has been added.
151    By default this is turned off.  To turn this on, do:
152    
153        CM.autoload "$smlnj/compiler.cm";
154        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
155    
156    Changes:
157    
158    1.  Changed FTAN to FPTAN so that the assembly output is correct.
159    2.  Changed the extension callback for FTANGENT to generate:
160    
161              fptan
162              fstp  %st(0)
163        instead of
164              fptan
165              fstpl ftempmem
166    
167    3.  Numerous assembly fixes for x86.
168    
169    5.  Cleaned up the machine code output module x86/x86MC.sml and added
170        support for a whole bunch of instructions and addressing modes:
171    
172          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
173          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
174          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
175          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
176          fxch %st(n)
177          fld %st(n)
178          fst %st(n)
179          fst mem
180          fstp %st(n)
181          fucom %st(n)
182          fucomp %st(n)
183    
184        All these are now generated when the fast fp mode is turned on.
185    
186    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
187    
188    ----------------------------------------------------------------------
189    Name: Matthias Blume
190    Date: 2000/11/09 11:20:00 JST
191    Tag: blume-20001109-scc
192    Description:
193    
194    Eliminated some code duplication:
195    
196    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
197       (Otherwise the functor is useless.)
198    2. Used GraphSCCFn where SCCUtilFun was used previously.
199    3. Got rid of SCCUtilFun (in comp-lib.cm).
200    
201    ----------------------------------------------------------------------
202    Name: Lal George
203    Date: 2000/11/06 09:02:21 EST 2000
204    Tag: Release_110_30
205    Description:
206    
207    - Version 110.30
208    ----------------------------------------------------------------------
209    Name: Matthias Blume
210    Date: 2000/11/04 14:45:00
211    Tag: blume-20001104-mlbuild
212    Description:
213    
214    - Made ml-build faster on startup.
215    - Documentation fixes.
216    
217    ----------------------------------------------------------------------
218    Name: Matthias Blume
219    Date: 2000/11/02 17:00:00 JST
220    Tag: blume-20001102-condcomp
221    Description:
222    
223    - Small tweaks to pickler -- new BOOTFILES!
224    - Version bumped to 110.29.2.
225    - Added conditional compilation facility to init.cmi (see comment there).
226    ----------------------------------------------------------------------
227    Name: Allen Leung
228    Date: 2000/10/23 19:31:00
229    Tag: leunga-20001023-demo-ra
230    
231    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
232    2. Test programs for the graph library updated
233    3. Some new MLRISC demo programs added
234    
235    ----------------------------------------------------------------------
236    Name: Matthias Blume
237    Date: 2000/08/31 22:15:00 JST
238    Tag: blume-20001017-errmsg
239    Description:
240    
241    More error message grief: Where there used to be no messages, there
242    now were some that had bogus error regions.  Fixed.
243    
244    ----------------------------------------------------------------------
245    Name: Matthias Blume
246    Date: 2000/08/31 17:30:00 JST
247    Tag: blume-20001017-v110p29p1
248    Description:
249    
250    I made a version 110.29.1 with new bootfiles.
251    
252    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
253              CM documentation changes and a small bugfix in CM's error reporting.
254    
255    ----------------------------------------------------------------------
256    Name: Lal George
257    Date: 2000/09/27 14:42:35 EDT
258    Tag: george-20000927-nodestatus
259    Description:
260    
261    Changed the type of the nodestatus, so that:
262    
263            SPILLED(~1)             is now SPILLED
264            SPILLED(m) where m>=0   is now MEMREG(m)
265            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
266    
267    ----------------------------------------------------------------------
268    Name: Matthias Blume
269    Date: 2000/09/07 14:45:00 JST
270    Tag: blume-20000907-cmerrmsg
271    Description:
272    
273    Small tweak to CM to avoid getting ML syntax error messages twice.
274    
275    ----------------------------------------------------------------------
276    Name: Matthias Blume
277    Date: 2000/08/31 18:00:00 JST
278    Tag: blume-20000831-cvsbootfiles
279    Description:
280    
281    New URL for boot files (because the 110.29 files on the BL server do
282    now work correctly with my updated install scripts for yacc and lex).
283    
284    ----------------------------------------------------------------------
285    Name: Matthias Blume
286    Date: 2000/08/08 12:33:00 JST
287    Tag: blume-20000808-manual
288    Description:
289    
290    Tiny update to CM manual.
291    
292    ----------------------------------------------------------------------
293    Name: Allen Leung
294    Date: 2000/08/7 19:31:00
295    Tag: leunga-20000807-a-whole-bunch-of-stuff
296    
297      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
298    and Mobdy.
299    
300    1.  x86
301    
302       a.  Fixes to peephole module by John and Dan.
303       b.  Assembly fix to SETcc by Allen.
304       c.  Fix to c-call by John.
305       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
306       e.  Instruction selection optimization to SETcc as suggested by John.
307    
308           For example,
309    
310            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
311    
312           should generate:
313    
314            MOVL a, x
315            SUBL b, x
316            SHRL 31, x
317    
318    2.  IR stuff
319    
320         A bunch of new DJ-graph related algorithms added.  These
321         speed up SSA construction.
322    
323    3.  SSA + Scheduling
324    
325         Added code for SSA and scheduling to the repository
326    
327    ----------------------------------------------------------------------
328    Name: Lal George
329    Date: 2000/07/27 11:53:14 EDT
330    
331    Tag: lal-20000727-linux-ppc
332    Description:
333    
334     Made changes to support Linux PPC.
335     p.s. I have confirmation that the 110.29 boot files work fine.
336    
337    ----------------------------------------------------------------------
338    Name: Matthias Blume
339    Date: 2000/07/27 17:40:00 JST
340    Tag: blume-20000727-scripts
341    Description:
342    
343    !!!! WARNING !!!!
344    You must recompile the runtime system!
345    !!!! WARNING !!!!
346    
347    This is basically another round of script-enhancements:
348    
349    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
350       and undefine CM preprocessor symbols.
351    
352    2. ml-build avoids generating a new heap image if it finds that the
353       existing one is still ok.  (The condition is that no ML file had to
354       be recompiled and all ML files are found to be older that the heap
355       file.)
356    
357       To make this work smoothly, I also hacked the runtime system as
358       well as SMLofNJ.SysInfo to get access to the heap image suffix
359       (.sparc-solaris, ...) that is currently being used.
360    
361       Moreover, the signature of CM.mk_standalone has changed.  See the
362       CM manual.
363    
364    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
365       CM manual for details.)
366    
367    4. More CM manual updates:
368        - all of the above has been documented.
369        - there is now a section describing the (CM-related) command line
370          arguments that are accepted by the "sml" command
371    
372    ----------------------------------------------------------------------
373    Name: Matthias Blume
374    Date: 2000/07/25 16:20:00 JST
375    Tag: blume-20000725-makedepend
376    Description:
377    
378    Added a script called ml-makedepend.  This can be used in makefiles
379    for Unix' make in a way very similar to the "makedepend" command for
380    C.
381    
382    The script internally uses function CM.sources.
383    
384    Synopsis:
385    
386        ml-makedepend [-f makefile] cmfile targetname
387    
388    The default for the makefile is "makefile" (or "Makefile" should
389    "makefile" not exist).
390    
391    ml-makedepend adds a cmfile/targetname-specific section to this
392    makefile (after removing the previous version of this section).  The
393    section contains a single dependency specification with targetname on
394    the LHS (targetname is an arbitrary name), and a list of files derived
395    from the cmfile on the RHS.  Some of the files on the RHS are
396    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
397    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
398    path names.  The makefile writer is responsible for making sure that
399    these variables have correct at the time "make" is invoked.
400    
401    ----------------------------------------------------------------------
402    Name: Matthias Blume
403    Date: 2000/07/22 23:30:00 JST
404    Tag: blume-20000722-urlupdate
405    Description:
406    
407    Changed BOOT and config/srcarchiveurl to point to BL server:
408    
409        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
410    
411    ----------------------------------------------------------------------
412    Name: Matthias Blume
413    Date: 2000/07/18 18:00:00 JST
414    Tag: blume-20000718-Version_110_29
415    Description:
416    
417    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
418    
419    2. Updated config/version to 110.29
420    
421    3. Updated config/srcarchiveurl
422    
423    3. New boot files!
424       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
425    
426    ----------------------------------------------------------------------
427    Name: Matthias Blume
428    Date: 2000/07/11 13:58:00 JST
429    Tag: blume-20000711-doctypo
430    Description:
431    
432    Fixed a few typos in CM manual.
433    
434    ----------------------------------------------------------------------
435    Name: Allen Leung
436    Date: 2000/06/15 00:38:00
437    Tag: leunga-20000704-sparc-x86
438    
439    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
440    2. fix to x86 RET bug [found by Dan Grossman]
441    3. sparc assembly bug fix for ticc instructions [found by Fermin]
442    
443       Affects c-- and moby only
444    
445    ----------------------------------------------------------------------
446    Name: Matthias Blume
447    Date: 2000/07/04 15:26:00
448    Tag: blume-20000704-trigger
449    Description:
450    
451    1. Improvements to CM manual.
452    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
453       of getting a back-trace.  The function, when called, raises an
454       internal exception which explicitly carries the full back-trace history,
455       so it is unaffected by any intervening handle-raise pairs ("trivial"
456       or not).  The interactive loop will print that history once it arrives
457       at top level.
458       Short of having all exceptions implicitly carry the full history, the
459       recommended way of using this facility is:
460         - compile your program with instrumentation "on"
461         - run it, when it raises an exception, look at the history
462         - if the history is "cut off" because of some handler, go and modify
463           your program so that it explicitly calls BTrace.trigger
464         - recompile (still instrumented), and rerun; look at the full history
465    
466    ----------------------------------------------------------------------
467    Name: Matthias Blume
468    Date: 2000/07/03 15:36:00 JST
469    Tag: blume-20000702-manual
470    Description:
471    
472    Small corrections and updates to CM manual.
473    
474    ----------------------------------------------------------------------
475    Name: Matthias Blume
476    Date: 2000/06/29 16:04:00 JST
477    Tag: blume-20000629-yacctool
478    Description:
479    
480    Changes:
481    
482    1. Class "mlyacc" now takes separate arguments to pass options to
483       generated .sml- and .sig-files independently.
484    2. Corresponding CM manual updates.
485    3. BTrace module now also reports call sites.  (However, for loop clusters
486       it only shows from where the cluster was entered.)  There are associated
487       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
488    
489    ----------------------------------------------------------------------
490    Name: Matthias Blume
491    Date: 2000/06/27 16:51:00 JST
492    Tag: blume-20000627-noweb
493    Description:
494    
495    Changes:
496    
497     1. Implemented "subdir" and "witness" options for noweb tool.
498        This caused some slight internal changes in CM's tool implementation.
499     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
500        some remaining issues from earlier path anchor changes.
501     3. Updated CM manual accordingly.
502    
503     4. Changed implementation of back-tracing so that I now consider it
504        ready for prime-time.
505    
506        In particular, you don't have to explicitly trigger the back-trace
507        anymore.  Instead, if you are running BTrace-instrumented code and
508        there is an uncaught exception (regardless of whether or not it was
509        raised in instrumented code), the top-level evalloop will print
510        the back-trace.
511    
512        Features:
513    
514          - Instrumented and uninstrumented code work together seemlessly.
515            (Of course, uninstrumented code is never mentioned in actual
516             back-traces.)
517    
518          - Asymptotic time- and space-complexity of instrumented code is
519            equal to that of uninstrumented code.  (This means that
520            tail-recursion is preserved by the instrumentation phase.)
521    
522          - Modules whose code has been instrumented in different sessions
523            work together without problem.
524    
525          - There is no penalty whatsoever on uninstrumented code.
526    
527          - There is no penalty on "raise" expressions, even in
528            instrumented code.
529    
530        A potential bug (or perhaps it is a feature, too):
531    
532          A back-trace reaches no further than the outermost instrumented
533          non-trivial "raise".  Here, a "trivial" raise is one that is the
534          sole RHS of a "handle" rule.  Thus, back-traces reach trough
535    
536               <exp> handle e => raise e
537    
538          and even
539    
540               <exp> handle Foo => raise Bar
541    
542          and, of course, through
543    
544               <exp> handle Foo => ...
545    
546         if the exception was not Foo.
547    
548         Back-traces always reach right through any un-instrumented code
549         including any of its "handle" expressions, trivial or not.
550    
551       To try this out, do the following:
552    
553         - Erase all existing binfiles for your program.
554           (You may keep binfiles for those modules where you think you
555            definitely don't need back-tracing.)
556         - Turn on back-trace instrumentation:
557              SMLofNJ.Internals.BTrace.mode (SOME true);
558         - Recompile your program.  (I.e., run "CM.make" or "use".)
559         - You may now turn instrumentation off again (if you want):
560              SMLofNJ.Internals.BTrace.mode (SOME false);
561         - Run your program as usual.  If it raises an exception that
562           reaches the interactive toplevel, then a back-trace will
563           automatically be printed.  After that, the toplevel loop
564           will print the exception history as usual.
565    
566    ----------------------------------------------------------------------
567    Name: Matthias Blume
568    Date: 2000/06/26 09:56:46 JST
569    Tag: blume-20000626-setup
570    Description:
571    
572    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
573          ML code before and after compiling a file (e.g., to set compiler
574          flags)
575    
576    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
577              - associated changes to btrace.sml (BTrace instrumentation pass)
578              - cleaner implementation of btimp.sml (BTrace tracing and report
579                module)
580    
581    CM manual: * new path encoding documented
582               * description of setup-parameter to "sml" added
583    
584    The biggest user-visible change to back-tracing is that it is no
585    longer necessary to compile all traced modules within the same
586    session.  (This was a real limitation.)
587    
588    ----------------------------------------------------------------------
589    Name: Matthias Blume
590    Date: 2000/06/24 12:40:00 JST
591    Tag: blume-20000624-startup
592    Description:
593    
594    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
595    bit too often -- to put it mildly. :)
596    
597    ----------------------------------------------------------------------
598    Name: Matthias Blume
599    Date: 2000/06/23 18:20:00 JST
600    Tag: blume-20000623-btrace
601    Description:
602    
603    This updates adds a backtrace facility to aid programmers in debugging
604    their programs.  This involves the following changes:
605    
606    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
607       keeping track of the current call stack.  When programs are compiled
608       in a special mode, the compiler will insert calls to these hooks
609       into the user program.
610       "Hook" means that it is possible for different implementations of
611       back-tracing to register themselves (at different times).
612    
613    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
614       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
615       It can be turned on using this call:
616         SMLofNJ.Internals.BTrace.mode (SOME true);
617       Turning it off again:
618         SMLofNJ.Internals.BTrace.mode (SOME false);
619       Querying the current status:
620         SMLofNJ.Internals.BTrace.mode NONE;
621       Annotated programs are about twice as big as normal ones, and they
622       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
623       where all hooks do nothing).  The slowdown with a plugin that is
624       actually useful (such as the one supplied by default) is even greater,
625       but in the case of the default plugin it is still only an constant
626       factor (amortized).
627    
628    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
629       with a sub-structure BTrace for controlling back-tracing.  In particular,
630       the above-mentioned function "mode" controls whether the annotation
631       phase is invoked by the compiler.  Another important function is
632       "trigger": when called it aborts the current execution and causes
633       the top-level loop to print a full back-trace.
634    
635    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
636       for back-tracing.  It keeps track of the dynamic call stack and in
637       addition to that it keeps a partial history at each "level" of that
638       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
639       d and b (at separate times, dynamically), then the report will show:
640    
641       GOTO   d
642             /c
643       GOTO  \b
644       CALL   a
645    
646       This shows that there was an initial non-tail call of a, then a
647       tail-call to b or c, looping behavior in a cluster of functions that
648       consist of b and c, and then a goto from that cluster (i.e., either from
649       b or from c) to d.
650    
651       Note that (depending on the user program) the amount of information
652       that the back-trace module has to keep track of at each level is bounded
653       by a constant.  Thus, the whole implementation has the same asymptotical
654       complexity as the original program (both in space and in time).
655    
656    5. compiler/TopLevel/interact/evalloop.sml has been modified to
657       handle the special exception SMLofNJ.Internals.BTrace.BTrace
658       which is raised by the "trigger" function mentioned above.
659    
660    Notes on usage:
661    
662    - Annotated code works well together with unannotated code:
663    Unannotated calls simply do not show up at all in the backtrace.
664    
665    - It is not a good idea to let modules that were annotated during
666    different sessions run at the same time.  This is because the compiler
667    chooses small integers to identify individual functions, and there
668    will be clashes if different modules were compiled in separate sessions.
669    (Nothing will crash, and you will even be told about the clashes, but
670    back-trace information will in general not be useful.)
671    
672    - Back-tracing can be confused by callcc and capture.
673    
674    - The only way of getting a back-trace right now is to explicitly
675    invoke the "trigger" function from your user program.  Eventually, we
676    should make every exception carry back-trace information (if
677    available).  But since this creates more overhead at "raise"-time
678    (similar to the current exnHistory overhead), I have not yet
679    implemented this.  (The implementation will be rather easy.)  With
680    exceptions carrying back-trace information, this facility will be even
681    more useful because users don't need to modify their programs...
682    
683    - While it is possible to compile the compiler with back-trace
684    annotations turned on (I did it to get some confidence in
685    correctness), you must make absolutely sure that core.sml and
686    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
687    be compiled with annotation because there is no core access yet, but
688    if you compile btimp.sml with annotation, then the system will go into
689    an infinite recursion and crash.)
690    Since CM currently does not know about BTrace, the only way to turn
691    annotations on and off for different modules of the compiler is to
692    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
693    this is awkward and clumsy.
694    
695    Sample sessions:
696    
697    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
698    - SMLofNJ.Internals.BTrace.mode (SOME true);
699    [autoloading]
700    [autoloading done]
701    val it = false : bool
702    - structure X = struct
703    -     fun main n = let
704    -         fun a (x, 0) = d x
705    -           | a (x, n) = b (x, n - 1)
706    -         and b (x, n) = c (x, n)
707    -         and c (x, n) = a (x, n)
708    -         and d x = e (x, 3)
709    -         and e (x, 0) = f x
710    -           | e (x, n) = e (x, n - 1)
711    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
712    -           | f n = n * g (n - 1)
713    -         and g n = a (n, 3)
714    -     in
715    -         f n
716    -     end
717    - end;
718    structure X : sig val main : int -> int end
719    - X.main 3;
720    *** BACK-TRACE ***
721    GOTO   stdIn:4.2-13.20: X.main[2].f
722    GOTO-( stdIn:4.2-13.20: X.main[2].e
723    GOTO   stdIn:4.2-13.20: X.main[2].d
724         / stdIn:4.2-13.20: X.main[2].a
725         | stdIn:4.2-13.20: X.main[2].b
726    GOTO-\ stdIn:4.2-13.20: X.main[2].c
727    CALL   stdIn:4.2-13.20: X.main[2].g
728    GOTO   stdIn:4.2-13.20: X.main[2].f
729    GOTO-( stdIn:4.2-13.20: X.main[2].e
730    GOTO   stdIn:4.2-13.20: X.main[2].d
731         / stdIn:4.2-13.20: X.main[2].a
732         | stdIn:4.2-13.20: X.main[2].b
733    GOTO-\ stdIn:4.2-13.20: X.main[2].c
734    CALL   stdIn:4.2-13.20: X.main[2].g
735    GOTO   stdIn:4.2-13.20: X.main[2].f
736    GOTO-( stdIn:4.2-13.20: X.main[2].e
737    GOTO   stdIn:4.2-13.20: X.main[2].d
738         / stdIn:4.2-13.20: X.main[2].a
739         | stdIn:4.2-13.20: X.main[2].b
740    GOTO-\ stdIn:4.2-13.20: X.main[2].c
741    CALL   stdIn:4.2-13.20: X.main[2].g
742    GOTO   stdIn:4.2-13.20: X.main[2].f
743    CALL   stdIn:2.15-17.4: X.main[2]
744    -
745    
746    (Note that because of a FLINt bug the above code currently does not
747    compile without BTrace turned on.)
748    
749    Here is another example, using my modified Tiger compiler:
750    
751    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
752    - SMLofNJ.Internals.BTrace.mode (SOME true);
753    [autoloading]
754    [autoloading done]
755    val it = false : bool
756    - CM.make "sources.cm";
757    [autoloading]
758    ...
759    [autoloading done]
760    [scanning sources.cm]
761    [parsing (sources.cm):parse.sml]
762    [creating directory CM/SKEL ...]
763    [parsing (sources.cm):tiger.lex.sml]
764    ...
765    [wrote CM/sparc-unix/semant.sml]
766    [compiling (sources.cm):main.sml]
767    [wrote CM/sparc-unix/main.sml]
768    [New bindings added.]
769    val it = true : bool
770    - Main.compile ("../testcases/merge.tig", "foo.out");
771    *** BACK-TRACE ***
772    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
773    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
774    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
775    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
776    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
777    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
778    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
779         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
780    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
781    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
782    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
783    CALL   main.sml:1.18-118.4: Main.compile[2]
784    -
785    
786    ----------------------------------------------------------------------
787    Name: Matthias Blumen
788    Date: 2000/06/21 18:00:00 JST
789    Tag: blume-20000621-manual
790    Description:
791    
792    CM manual update: Path environments documented.
793    
794    ----------------------------------------------------------------------
795    Name: Matthias Blume
796    Date: 2000/06/19 13:40:00
797    Tag: blume-20000619-manual
798    Description:
799    
800    CM manual and system/README update.  This only covers the fact that
801    there are no more implicit anchors.  (Path environments and the "bind"
802    option to "cm" have yet to be documented.)
803    
804    ----------------------------------------------------------------------
805    Name: Matthias Blume
806    Date: 2000/06/19 11:05:00 JST
807    Tag: blume-20000619-chdir-bugfix
808    Description:
809    
810    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
811    
812    ----------------------------------------------------------------------
813    Name: Matthias Blume
814    Date: 2000/06/18 22:00:10 JST
815    Tag: blume-20000618-implicit-anchors-really-gone
816    Description:
817    
818    I updates the previous HISTORY entry where I forgot to mention that
819    implicit anchors are no longer with us.
820    
821    The current update also gets rid of the (now useless) controller
822    CM.Control.implicit_anchors.
823    
824    ----------------------------------------------------------------------
825    Name: Matthias Blume
826    Date: 2000/06/16 17:30:00 JST
827    Tag: blume-20000616-anchorenv
828    Description:
829    
830    This patch implements the long anticipated (just kidding :) "anchor
831    environment" mechanism.  In the course of doing this, I also
832    re-implemented CM's internal "SrcPath" module from scratch.  The new
833    one should be more robust in certain boundary cases.  In any case, it
834    is a lot cleaner than its predecessor (IMHO).
835    
836    This time, although there is yet another boot file format change, I
837    kept the unpickler backward-compatible.  As a result, no new bootfiles
838    are necessary and bootstrapping is straightforward.  (You cannot read
839    new bootfiles into an old system, but the other way around is no
840    problem.)
841    
842    Visible changes:
843    
844    ** 0. Implicit path anchors (without the leading $-symbol) are no
845    longer recognized at all. This means that such path names are not
846    illegal either.  For example, the name basis.cm simply refers to a
847    local file called "basis.cm" (i.e, the name is an ordinary path
848    relative to .cm-files directory).  Or, to put it differently, only
849    names that start with $ are anchored paths.
850    
851    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
852    vanished.
853    
854    John (Reppy) had critizised this as soon as I originally proposed and
855    implemented it, but at that time I did not really deeply believe
856    him. :) Now I came full-circle because I need the $<singlearc> syntax
857    in another place where it cannot be seen as an abbreviation for
858    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
859    seems to mean (i.e., it "expands" into the corresponding anchor
860    value).
861    
862    However, when paths are used as members in CM description files, it
863    continues to be true that there must be at least another arc after the
864    anchor.  This is now enforced separately during semantic analysis
865    (i.e., from a lexical/syntactical point of view, the notation is ok.)
866    
867    ** 2. The "cm" class now accepts an option "bind".  The option's value
868    is a sub-option list of precisely two items -- one labeled "anchor"
869    and the other one labeled "value".  As you might expect, "anchor" is
870    used to specify an anchor name to be bound, and "value" specifies what
871    the anchor is being bound to.
872    
873    The value must be a directory name and can be given in either standard
874    syntax (including the possibility that it is itself an anchored path)
875    or native syntax.
876    
877    Examples:
878    
879       foo.cm (bind:(anchor:bar value:$mystuff/bar))
880       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
881    
882    and so on.
883    
884    The meaning of this is that the .cm-file will be processed with an
885    augmented anchor environment where the given anchor(s) is/are bound to
886    the given values(s).
887    
888    The rationale for having this feature is this: Suppose you are trying
889    to use two different (already stable) libraries a.cm and b.cm (that
890    you perhaps didn't write yourself).  Further, suppose each of these
891    two libraries internally uses its own auxiliary library $aux/lib.cm.
892    Normally you would now have a problem because the anchor "lib" can not
893    be bound to more than one value globally.  Therefore, the project that
894    uses both a.cm and b.cm must locally redirect the anchor to some other
895    place:
896    
897       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
898       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
899    
900    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
901    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
902    
903    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
904    you when you do so at the time of CM.stabilize).  Therefore, you can
905    also use an anchored path as the value:
906    
907      a.cm (bind:(anchor:lib value:$a-lib))
908      b.cm (bind:(anchor:lib value:$b-lib))
909    
910    Now you can globally configure (using the usual CM.Anchor.anchor or
911    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
912    itself is always locally bound, setting it globally is no longer
913    meaningful or necessary (but it does not hurt either).  In fact, "lib"
914    can still be used as a global anchor for separate purposes.  As a
915    matter of fact, one can locally define "lib" in terms of a global
916    "lib":
917    
918      a.cm (bind:(anchor:lib value:$lib/a))
919      b.cm (bind:(anchor:lib value:$lib/b))
920    
921    ** 3: The encoding of path names has changed.  This affects the way
922    path names are shown in CM's progress report and also the internal
923    protocol encoding used for parallel make.
924    
925    The encoding now uses one or more ':'-separated segments.  Each
926    segments corresponds to a file that has been specified relative to the
927    file given by its preceding segment.  The first segment is either
928    relative to the CWD, absolute, or anchored.  Each segment itself is
929    basically a Unix pathname; all segments but the first are relative.
930    
931    Example:
932    
933       $foo/bar/baz.cm:a/b/c.sml
934    
935    This path denotes the file bar/a/b/c.sml relative to the directory
936    denoted by anchor "foo".  Notice that the encoding also includes
937    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
938    paths are resolved relative to the .cm-files directory, so baz.cm must
939    be ignored to get the "real" pathname.
940    
941    To make this fact more obvious, CM puts the names of such "virtual
942    arcs" into parentheses when they appear in progress reports. (No
943    parentheses will appear in the internal protocol encoding.)  Thus,
944    what you really see is:
945    
946      $foo/bar/(baz.cm):a/b/c.sml
947    
948    I find this notation to be much more informative than before.
949    
950    Another new feature of the encoding is that special characters
951    including parentheses, colons, (back)slashes, and white space are
952    written as \ddd (where ddd is the decimal encoding of the character).
953    
954    *** The CM manual still needs to be updated.
955    
956    ----------------------------------------------------------------------
957    Name: Allen Leung
958    Date: 2000/06/15 00:38:00
959    Tag: leunga-20000615-x86-peephole
960    
961    x86 Peephole fix by Fermin.  Affects c-- and moby only.
962    
963    ----------------------------------------------------------------------
964    Name: Matthias Blume
965    Date: 2000/06/12 11:40:00
966    Tag: blume-20000612-parmakefix
967    Description:
968    
969    More cleanup after changing the file naming scheme: This time I
970    repaired the parallel make mechanism for CMB.make which I broke earlier.
971    
972    ----------------------------------------------------------------------
973    Name: Allen Leung
974    Date: 2000/06/09 01:25:00
975    Tag: leunga-20000609-various
976    
977    None of these things should affect normal SML/NJ operations
978    
979    1. Peephole improvements provided by Fermin (c--)
980    2. New annotation DEFUSE for adding extra dependence (moby)
981    3. New X86 LOCK instructions (moby)
982    4. New machine description language for reservation tables (scheduling)
983    5. Fixes to various optimization/analysis modules (branch chaining, dominator
984       trees etc.)
985    6. I've changed the CM files so that they can work with versions
986       110.0.6, 110.25 and 110.28
987    
988    ----------------------------------------------------------------------
989    Name: Matthias Blume
990    Date: 2000/06/09 12:40:00
991    Tag: blume-20000609-log
992    Description:
993    
994    - Removed all(?) remaining RCS Log entries from sources.
995    
996    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
997      anchored paths).
998    
999    ----------------------------------------------------------------------
1000    Name: Matthias Blume
1001    Date: 2000/06/07 17:00:00 JST
1002    Tag: blume-20000607-no-implicit-anchors
1003    Description:
1004    
1005    1. This update changes the default setting for
1006    CM.Control.implicit_anchors from true to false.  This means that
1007    implicit anchors are no longer permitted by default.  I also tried to
1008    make sure that nothing else still relies on implicit anchors.
1009    (This is the next step on the schedule towards a CM that does not even
1010    have the notion of implicit anchors anymore.)
1011    
1012    2. More CM manual updates.
1013    
1014    3. I managed to track down and fix the pickling bug I mentioned last
1015    time.  Because of the previously existing workaround, this entails no
1016    immediate practical changes.
1017    
1018    ----------------------------------------------------------------------
1019    Name: Matthias Blume
1020    Date: 2000/06/06 11:15:00 JST
1021    Tag: blume-20000606-lazierpickle
1022    Description:
1023    
1024    !!!! NEW BOOT FILES !!!!
1025    
1026    * The main purpose of this update is to make library pickles lazier in
1027    order to reduce the initial space penalty for autoloading a library.
1028    As a result, it is now possible to have $smlnj/compiler.cm
1029    pre-registered.  This should take care of the many complaints or
1030    inquiries about missing structure Compiler.  This required changes to
1031    CM's internal data structures and small tweaks to some algorithms.
1032    
1033    As a neat additional effect, it is no longer necessary (for the sake
1034    of lean heap image files) to distinguish between a "minimal" CM and a
1035    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1036    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1037    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1038    
1039    To make the life of compiler-hackers easier, "makeml" now also
1040    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1041    words, after you bootstrap a new sml for the first time, you will not
1042    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1043    around you will still have to do it, though.)
1044    
1045    * A second change consists of major updates to the CM manual.  There
1046    are now several appendices with summary information and also a full
1047    specification of the CM description file syntax.
1048    
1049    * In directory src/system I added the script "allcross".  This script
1050    invokes sml and cross-compiles the compiler for all supported
1051    architectures.  (Useful when providing a new set of boot files.)
1052    
1053    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1054    added a small tweak to pickle-util.sml to work around this problem,
1055    but it is not a proper fix yet.  I will investigate further.  (The
1056    effect of the bug was an inflation of library pickle size.)
1057    
1058    * Version number increased to 110.28.1 (to avoid compatibility problems).
1059    
1060    ----------------------------------------------------------------------
1061    Name: Allen Leung
1062    Date: 2000/05/25 17:28 EDT
1063    Tag: leunga-20000525-ra
1064    Description:
1065    
1066      Fixed a bug in freezing phase of the register allocator.
1067    
1068    ----------------------------------------------------------------------
1069    Name: Allen Leung
1070    Date: 2000/05/15 22:53 EDT
1071    Tag: leunga-20000515-alpha-x86-ra
1072    Description:
1073    
1074      1. Alpha
1075    
1076          Slight cleanup.  Removed the instruction SGNXL
1077    
1078      2. X86
1079    
1080          Added the following instructions to the instruction set:
1081    
1082            ROLx, RORx,
1083            BTx, BTSx, BTLx, BTRx,
1084            XCHGx, and variants with the LOCK prefix
1085    
1086      3. Register Allocation
1087    
1088          The module ra-rewrite-with-renaming has been improved.
1089    
1090      These have no effect on SML/NJ.
1091    
1092    ----------------------------------------------------------------------
1093    Name: Matthias Blume
1094    Date: 2000/05/15 16:20:00 JST
1095    Tag: blume-20000515-lightrebuild
1096    Description:
1097    
1098    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1099       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1100       defined.  In effect, the command will not build any cross-compiler
1101       backends and therefore finish more quickly.
1102    
1103       The "fixpt" script also takes a "-light" switch to be able to use
1104       this new facility while compiling for a fixpoint.
1105    
1106    2. I replaced all mentions of anchored paths in group owner specifications
1107       with simple relative paths (usually starting with "..").
1108       The rationale is that a library's internal workings should not be
1109       compromised by the lack of some anchor.  (An anchor is necessary
1110       for someone who wants to refer to the library by an anchored path,
1111       but it should not be necessary to build the same library in the first
1112       place.)
1113    
1114    3. I changed the way CM's tool mechanism determines the shell command
1115       string used for things like ml-yacc etc. so that it does not break
1116       when CM.Control.implicit_anchors is turned off.
1117    
1118    ----------------------------------------------------------------------
1119    Name: Matthias Blume
1120    Date: 2000/05/12 18:20:00 JST
1121    Tag: blume-20000512-ml-build
1122    Description:
1123    
1124    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1125    getting installed properly (by config/install.sh).
1126    
1127    ----------------------------------------------------------------------
1128    Name: Matthias Blume
1129    Date: 2000/05/12 17:30:00 JST
1130    Tag: blume-20000512-anchors
1131    Description:
1132    
1133    !!! NEW BOOT FILES !!!
1134    
1135    This change is in preparation of fading out support for "implicitly
1136    anchored path names".  I went through all sources and used the
1137    explicit (and relatively new) $-notation.  See system/README and the
1138    CM manual for more info on this.
1139    
1140    I also modified the anchoring scheme for some things such as "smlnj",
1141    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1142    anchors are more expressive: anchor name and first arc do not have to
1143    coincide.  This entails the following user-visible change:
1144    
1145    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1146    particular, when you fire up sml with a command-line argument, say,
1147    e.g.:
1148    
1149       sml '$smlnj/cmb.cm'
1150    
1151    At the ML toplevel prompt:
1152    
1153       CM.autoload "$smlnj/cmb.cm";
1154    
1155    There is also a new controller in CM.Control that can be used to turn
1156    off all remaining support for implicit anchors by saying:
1157    
1158        CM.autoload "$smlnj/
1159        #set CM.Control.implicit_anchors false;
1160    
1161    This causes CM to reject implicitly anchored paths.  This is (for the
1162    time being) less permissive than the "final" version where there will
1163    be no more such implicit anchors and relative paths will be just that:
1164    relative.
1165    
1166    The next step (version after next version?) will be to make the
1167    default for CM.Control.implicit_anchors false.  After the dust has
1168    settled, I can then produce the "final" version of this...
1169    
1170    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1171    
1172    ----------------------------------------------------------------------
1173    Name: Matthias Blume
1174    Date: 2000/05/11 16:30:00 JST
1175    Tag: blume-20000511-sources
1176    Description:
1177    
1178    The main change is that I added function CM.sources as a generalized
1179    version of the earlier CM.makedepend.  This entails the following
1180    additional changes:
1181    
1182      - CM.makedepend has been dropped.
1183    
1184      - CM manual has been updated.
1185    
1186      - TOOLS signature and API have been changed.
1187    
1188    ----------------------------------------------------------------------
1189    Name: Allen Leung
1190    Date: 2000/05/10 21:17 EDT
1191    Tag: leunga-20000510-moby-c--ssa
1192    Description:
1193    
1194      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1195    None of these affect SML/NJ.
1196    
1197    1. Register Allocation
1198    
1199        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1200           This module tries to remove local (i.e. basic block level) redundancies
1201           during spilling.
1202    
1203        b. A new framework for performing region based register allocation.
1204           Not yet entirely functional.
1205    
1206    2. X86
1207    
1208       a. DefUse for POP was missing the stack pointer [found by Lal]
1209       b. Reload for CALL was incorrect in X86Spill [found by John]
1210       c. Various fixes in X86Spill so that it can be used correctly for
1211          the new spilling module.
1212    
1213    3. SSA/IR
1214    
1215       a. New module ir/dj-dataflow.sml implements elimination based
1216          data flow analysis.
1217    
1218    4. MLRiscGen
1219    
1220       a. Fix for gc type annotation
1221    
1222    5. MDGen
1223    
1224       Various fixes for machine description -> ml code translation.  For ssa
1225       only.
1226    
1227    ----------------------------------------------------------------------
1228    Name: Allen Leung
1229    Date: 2000/05/08 22:17 EDT
1230    Tag: leunga-20000508-labexp
1231    Description:
1232    
1233      Fermin has found a few assembly problems with constant expressions
1234      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1235      which choke on dumb assemblers.  This is his fix.
1236    
1237    ----------------------------------------------------------------------
1238    Name: Dave MacQueen
1239    Date: 2000/04/09 14:00 EDT
1240    Tag: dbm-20000502-Version_110_28
1241    Description:
1242    
1243    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1244    
1245    2. Updated config/version to 110.28
1246    
1247    3. Updated config/srcarchiveurl
1248    
1249    3. New boot files!
1250       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1251    
1252    ----------------------------------------------------------------------
1253    Name: Matthias Blume
1254    Date: 2000/05/01 19:05:00 JST
1255    Tag: blume-20000501-noweb
1256    Description:
1257    
1258    A new noweb tool has been added.  The existing system is entirely
1259    unaffected by this, but some CM users have asked for renewed noweb
1260    support.  Everything is documented in the CM manual.
1261    
1262    New (plugin) libraries:
1263    
1264       noweb-tool.cm
1265       nw-ext.cm
1266    
1267    ----------------------------------------------------------------------
1268    Name: Dave MacQueen
1269    Date: 2000/04/30 12:40PM EDT
1270    Tag: dbm-20000430-bug_fixes
1271    Description:
1272    
1273    1. Fix for bug 1498
1274       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1275       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1276         added toRealArray function
1277       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1278         added check for tag Obj.RealArray to array printing case in ppObj
1279    
1280    2. Fix for bug 1510
1281       smlnj/src/compiler/Semant/types/typesutil.sml
1282         fixed definition of dummyargs (used by equalTycon) so that
1283         dummy args are distinct types
1284    
1285    ----------------------------------------------------------------------
1286    Name: Matthias Blume
1287    Date: 2000/04/30 01:00:00 JST
1288    Tag: blume-20000430-versions
1289    Description:
1290    
1291    1. CM version numbering added.  This is an implementation of Lal's
1292       proposal for adding version numbers and version checking to .cm
1293       files.  Lal said that his proposal was just that -- a proposal.
1294       For the time being I went ahead and implemented it so that people
1295       can comment on it.  Everything is completely backward-compatible
1296       (except for the stable library format, i.e., new bootfiles!).
1297    
1298       As usual, see the CM manual for details.
1299    
1300    2. An alternative syntax for anchored paths has been implemented.
1301       Dave has recently voiced the same concerns that I had when I did
1302       this, so there should be some support.  My take is that eventually
1303       I will let support for the current syntax (where anchors are
1304       "implicit") fade out in favor of the new, explicit syntax.
1305       In order to be backward-compatible, both old and new syntax are
1306       currently supported.
1307    
1308       Again, see the CM manual for details.
1309    
1310    3. Parallel make is trying to be slightly smarter:  When the master
1311       process finds a "bottleneck", i.e., when there is only one
1312       compilation unit that can be compiled and everybody else is
1313       waiting on it, then it will simply compile it directly instead
1314       of clumsily telling one of the slaves to do it.
1315    
1316    4. Support for "unsharing" added.  This is necessary in order to be
1317       able to have two different versions of the same library running
1318       at the same time (e.g., for trying out a new MLRISC while still
1319       having the old MLRISC linked into the current compiler, etc.)
1320       See the CM manual.
1321    
1322    5. Simple "makedepend" functionality added for generating Makefile
1323       dependency information.  (This is rather crude at the moment.
1324       Expect some changes here in the future.)
1325    
1326    6. ".fun" added as a recognized suffix for ML files. Also documented
1327       explicitly in the manual that the fallback behavior (unknown suffix
1328       -> ML file) is not an official feature!
1329    
1330    7. Small changes to the pickler for stable libraries.
1331    
1332    8. Several internal changes to CM (for cleanup/improvement).
1333    
1334    
1335    !!!! NEW BINFILES !!!!
1336    
1337    ----------------------------------------------------------------------
1338    Name: Matthias Blume
1339    Date: 2000/04/28 17:30:00 JST
1340    Tag: blume-20000428-pathconfig
1341    Description:
1342    
1343    1. I changed config/install.sh to remove duplicate entries from the
1344       lib/pathconfig file at the end.  Moreover, the final version of
1345       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1346       in src/system/installml.
1347    
1348    2. The config/install.sh script now consistently uses relative
1349       pathnames in lib/pathconfig whenever the anchor is in the lib
1350       directory.  (So far this was true for the libraries that come
1351       pre-compiled and bundled as part of the bootfiles but not for
1352       libraries that are compiled by the script itself.)
1353    
1354    ----------------------------------------------------------------------
1355    Name: Matthias Blume
1356    Date: 2000/04/26 13:10:00 JST
1357    Tag: blume-20000426-fun_suffix
1358    Description:
1359    
1360    Added ".fun" as a recognized file name suffix (for ML code).
1361    
1362    ----------------------------------------------------------------------
1363    Name: Allen Leung
1364    Date: 2000/04/25 17:00:00 EST
1365    Tag: leunga-20000425-alpha-ra
1366    Description:
1367    
1368    1. Alpha
1369    
1370        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1371    in C--.
1372    
1373    2. RA
1374    
1375       Added a flag "ra-dump-size" to print out the size of the flowgraph
1376       and the interference graph.
1377    
1378    ----------------------------------------------------------------------
1379    Name: Dave MacQueen
1380    Date: 2000/04/25/
1381    Tag: dbm-20000425-mlyacc_doc_examples
1382    Description:
1383      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1384      Updated all three examples in src/ml-yacc/examples to run
1385      under 110.* using CM.make.
1386    
1387    ----------------------------------------------------------------------
1388    Name: Allen Leung
1389    Date: 2000/04/20 23:04:00 EST
1390    Tag: leunga-20000420-ssa-c---stuff
1391    Description:
1392    
1393      This update synchronizes my repository with Yale's.  Most of these
1394    changes, however, do not affect SML/NJ at all (the RA is an exception).
1395    
1396    1. Register Allocator
1397    
1398       a. An improvement in the interference graph construction:
1399          Given a copy
1400    
1401                s <- t
1402    
1403          no interference edge between s and t is added for this definition of s.
1404    
1405       b. I've added two new spill heuristic modules that Fermin and I developed
1406          (in the new library RA.cm). These are unused in SML/NJ but maybe
1407          useful for others (Moby?)
1408    
1409    2. X86
1410    
1411       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1412    
1413    3. Alpha
1414    
1415       a. Added the BSR instruction and code generation that goes with it [C--]
1416       b. Other fixes too numerous to recount provided by Fermin [C--]
1417    
1418    4. Regmaps
1419    
1420       a. The regmaps are not initialized with the identity physical bindings
1421          at creation time.  This is unneeded.
1422    
1423    5. MLRISC Optimizations
1424    
1425       a. The DJ-Graph module can now compute the iterated dominance frontiers
1426          intersects with liveness incrementally in linear time! Woohoo!
1427          This is now used in my new SSA construction algorithm.
1428    
1429       b. THe branch reorganization module is now smarter about linear chains of
1430          basic blocks.
1431    
1432    
1433    ----------------------------------------------------------------------
1434    Name: Matthias Blume
1435    Date: 2000/04/12 13:52:00 JST
1436    Tag: blume_main_v110p27_1
1437    Description:
1438    
1439    Changed install.sh script to handle archive files without version number
1440    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1441    name of the boot file archive.
1442    
1443    ----------------------------------------------------------------------
1444    Name: Dave MacQueen
1445    Date: 2000/04/09 14:00 EDT
1446    Tag: dbm-20000410-Version_110_27
1447    Description:
1448    
1449    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1450    
1451    2. Updated src/config/version to 110.27
1452    
1453    3. New boot files!
1454    
1455    ----------------------------------------------------------------------
1456    Name: Allen Leung
1457    Date: 2000/04/09 19:09:00 EST
1458    Tag: leunga-20000409-misc
1459    Description:
1460    
1461    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1462    
1463    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1464    
1465    ----------------------------------------------------------------------
1466    Name: Stefan
1467    Date: 2000/04/07 10:00:00 EDT
1468    Tag: monnier-20000406-branch-handling
1469    Description:
1470    
1471    Improved handling of branches (mostly those generated from
1472    polymorphic equality), removed switchoff and changed the
1473    default optimization settings (more cpsopt and less flintopt).
1474    
1475    ----------------------------------------------------------------------
1476    Name: Allen Leung
1477    Date: 2000/04/06 01:30:00 EST
1478    Tag: leunga-20000406-peephole-x86-SSA-2
1479    Description:
1480    
1481       Forgot a few files.
1482    
1483    ----------------------------------------------------------------------
1484    Name: Allen Leung
1485    Date: 2000/04/06 00:36:00 EST
1486    Tag: leunga-20000406-peephole-x86-SSA
1487    Description:
1488    
1489    1.  New Peephole code
1490    
1491    2.  Minor improvement to X86 instruction selection
1492    
1493    3.  Various fixes to SSA and machine description -> code translator
1494    
1495    ----------------------------------------------------------------------
1496    Name: Matthias Blume
1497    Date: 2000/04/05 12:30:00 JST
1498    Tag: blume_main_v110p26p2_3
1499    Description:
1500    
1501    This update just merges three minor cosmetic updates to CM's sources
1502    to get ready for the 110.27 code freeze on Friday.  No functionality
1503    has changed.
1504    
1505    ----------------------------------------------------------------------
1506    Name: Allen Leung
1507    Date: 2000/04/04 19:39:00 EST
1508    Tag: leunga-20000404-x86-asm
1509    Description:
1510    
1511    1.  Fixed a problem in X86 assembly.
1512    
1513        Things like
1514    
1515           jmp %eax
1516           jmp (%eax)
1517    
1518        should be output as
1519    
1520           jmp *%eax
1521           jmp *(%eax)
1522    
1523    2.  Assembly output
1524    
1525          Added a new flag
1526    
1527              "asm-indent-copies" (default to false)
1528    
1529          When this flag is on, parallel copies will be indented an extra level.
1530    
1531    ----------------------------------------------------------------------
1532    Name: Allen Leung
1533    Date: 2000/04/04 03:18:00 EST
1534    Tag: leunga-20000404-C--Moby
1535    Description:
1536    
1537        All of these fixes are related to C--, Moby, and my own optimization
1538        stuff; so they shouldn't affect SML/NJ.
1539    
1540    1.  X86
1541    
1542        Various fixes related floating point, and extensions.
1543    
1544    2.  Alpha
1545    
1546        Some extra patterns related to loads with signed/zero extension
1547        provided by Fermin.
1548    
1549    3.  Assembly
1550    
1551        When generating assembly, resolve the value of client defined constants,
1552        instead of generating symbolic values.  This is controlled by the
1553        new flag "asm-resolve-constants", which is default to true.
1554    
1555    4.  Machine Descriptions
1556    
1557        a. The precedence parser was slightly broken when parsing infixr symbols.
1558        b. The type generalizing code had the bound variables reversed, resulting
1559           in a problem during arity raising.
1560        c. Various fixes in machine descriptions.
1561    
1562    ----------------------------------------------------------------------
1563    Name: Matthias Blume
1564    Date: 2000/04/03 16:05:00 JST
1565    Tag: blume_main_v110p26p2_2
1566    Description:
1567    
1568    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1569    now done via the ordinary static environment that is context to each
1570    compilation unit.
1571    
1572    To this end, I arranged that instead of "structure Core" as "structure
1573    _Core" is bound in the pervasive environment.  Core access is done via
1574    _Core (which can never be accidentally rebound because _Core is not a
1575    legal surface-syntax symbol).
1576    
1577    The current solution is much cleaner because the core environment is
1578    now simply part of the pervasive environment which is part of every
1579    compilation unit's context anyway.  In particular, this eliminates all
1580    special-case handling that was necessary until now in order to deal
1581    with dynamic and symbolic parts of the core environment.
1582    
1583    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1584    compilation manager's bootstrap compiler (actually: in the "init group"
1585    handling).  See the comments in src/system/smlnj/init/init.cmi for
1586    more details.
1587    
1588    I also tried to track down all mentions of "Core" (as string argument
1589    to Symbol.strSymbol) in the compiler and replaced them with a
1590    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1591    name appears in one place only.
1592    
1593    Binfile and bootfile format have not changed, but the switchover from
1594    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1595    bootfiles anyway.
1596    
1597    ----------------------------------------------------------------------
1598    Name: Allen Leung
1599    Date: 2000/04/02 21:17:00 EST
1600    Tag: leunga-20000402-mltree
1601    Description:
1602    
1603       1. Renamed the constructor CALL in MLTREE by popular demand.
1604       2. Added a bunch of files from my repository.  These are currently
1605          used by other non-SMLNJ backends.
1606    
1607    ----------------------------------------------------------------------
1608    Name: Allen Leung
1609    Date: 2000/03/31 21:15:00 EST
1610    Tag: leunga-20000331-aliasing
1611    Description:
1612    
1613    This update contains a rewritten (and hopefully more correct) module
1614    for extracting aliasing information from CPS.
1615    
1616       To turn on this feature:
1617    
1618            Compiler.Control.CG.memDisambiguate := true
1619    
1620       To pretty print the region information with assembly
1621    
1622           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1623    
1624       To control how many levels of aliasing information are printed, use:
1625    
1626           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1627    
1628       The default of n is 3.
1629    
1630    ----------------------------------------------------------------------
1631    Name: David MacQueen
1632    Date: 2000/03/31 11:15:00 EST
1633    Tag: dbm-20000331-runtime_fix
1634    Description:
1635    
1636    This update contains:
1637    
1638    1. runtime/c-lib/c-libraries.c
1639       includes added in revision 1.2 caused compilation errors on hppa-hpux
1640    
1641    2. fix for bug 1556
1642       system/Basis/Implementation/NJ/internal-signals.sml
1643    
1644    ----------------------------------------------------------------------
1645    Name: Matthias Blume
1646    Date: 2000/03/31 18:00:00 JST
1647    Tag: blume_main_v110p26p2_1
1648    Description:
1649    
1650    This update contains:
1651    
1652    1. A small change to CM's handling of stable libraries:
1653       CM now maintains one "global" modmap that is used for all stable
1654       libraries.  The use of such a global modmap maximizes sharing and
1655       minimizes the need for re-traversing parts of environments during
1656       modmap construction.  (However, this has minor impact since modmap
1657       construction seems to account for just one percent or less of total
1658       compile time.)
1659    
1660    2. I added a "genmap" phase to the statistics.  This is where I got the
1661       "one percent" number (see above).
1662    
1663    3. CM's new tool parameter mechanism just became _even_ better. :)
1664       - The parser understands named parameters and recursive options.
1665       - The "make" and "shell" tools use these new features.
1666         (This makes it a lot easier to cascade these tools.)
1667       - There is a small syntax change: named parameters use a
1668    
1669           <name> : ( <option> ... )            or
1670           <name> : <string>
1671    
1672         syntax.  Previously, named parameters were implemented in an
1673         ad-hoc fashion by each tool individually (by parsing strings)
1674         and had the form
1675    
1676           <name>=<string>
1677    
1678       See the CM manual for a full description of these issues.
1679    
1680    ----------------------------------------------------------------------
1681    Name: Matthias Blume
1682    Date: 2000/03/30 18:00:00 JST
1683    Tag: blume_main_v110p26p2_0
1684    Description:
1685    
1686    !!!!! WARNING !!!!!!
1687    !!  New binfiles  !!
1688    !!!!!!!!!!!!!!!!!!!!
1689    
1690    This update contains:
1691    
1692    1. Moderate changes to CM:
1693    
1694       - Changes to CM's tools mechanism.  In particular, it is now possible
1695       to have tools that accept additional "command line" parameters
1696       (specified in the .cm file at each instance where the tool's class is
1697       used).
1698    
1699       This was done to accommodate the new "make" and "shell" tools which
1700       facilitate fairly seamless hookup to portions of code managed using
1701       Makefiles or Shell scripts.
1702    
1703       There are no classes "shared" or "private" anymore.  Instead, the
1704       sharing annotation is now a parameter to the "sml" class.
1705    
1706       There is a bit of generic machinery for implementing one's own
1707       tools that accept command-line parameters.  However, I am not yet fully
1708       satisfied with that part, so expect changes here in the future.
1709    
1710       All existing tools are described in the CM manual.
1711    
1712       - Slightly better error handling.  (CM now suppresses many followup
1713       error messages that tended to be more annoying than helpful.)
1714    
1715    2. Major changes to the compiler's static environment data structures.
1716    
1717       - no CMStaticEnv anymore.
1718            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
1719              but it is called Environment), no conversions between different
1720              kinds of static environments
1721    
1722       - There is still a notion of a "modmap", but such modmaps are generated
1723         on demand at the time when they are needed.  This sounds slow, but I
1724         sped up the code that generates modmaps enough for this not to lead to
1725         a slowdown of the compiler (at least I didn't detect any).
1726    
1727       - To facilitate rapid modmap generation, static environments now
1728         contain an (optional) "modtree" structure.  Modtree annotations are
1729         constructed by the unpickler during unpickling.  (This means that
1730         the elaborator does not have to worry about modtrees at all.)
1731         Modtrees have the advantage that they are compositional in the same
1732         way as the environment data structure itself is compositional.
1733         As a result, modtrees never hang on to parts of an environment that
1734         has already been rendered "stale" by filtering or rebinding.
1735    
1736       - I went through many, many trials and errors before arriving at the
1737         current solution.  (The initial idea of "linkpaths" did not work.)
1738         But the result of all this is that I have touched a lot of files that
1739         depend on the "modules" and "types" data structures (most of the
1740         elaborator). There were a lot of changes during my "linkpath" trials
1741         that could have been reverted to their original state but weren't.
1742         Please, don't be too harsh on me for messing with this code a bit more
1743         than what was strictly necessary...  (I _did_ resist the tempation
1744         of doing any "global reformatting" to avoid an untimely death at
1745         Dave's hands. :)
1746    
1747       - One positive aspect of the previous point:  At least I made sure that
1748         all files that I touched now compile without warnings (other than
1749         "polyEqual").
1750    
1751       - compiler now tends to run "leaner" (i.e., ties up less memory in
1752         redundant modmaps)
1753    
1754    ----------------------------------------------------------------------
1755    Name: Allen Leung
1756    Date: 2000/03/29 18:00:00
1757    Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
1758    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
1759    Description:
1760    
1761       This update contains *MAJOR* changes to the way code is generated from CPS
1762    in the module mlriscGen, and in various backend modules.
1763    
1764    CHANGES
1765    =======
1766    
1767    1. MLRiscGen: forward propagation fix.
1768    
1769       There was a bug in forward propagation introduced at about the same time
1770       as the MLRISC x86 backend, which prohibits coalescing to be
1771       performed effectively in loops.
1772    
1773       Effect: speed up of loops in RISC architectures.
1774               By itself, this actually slowed down certain benchmarks on the x86.
1775    
1776    2. MLRiscGen:  forward propagating addresses from consing.
1777    
1778       I've changed the way consing code is generated.  Basically I separated
1779       out the initialization part:
1780    
1781            store tag,   offset(allocptr)
1782            store elem1, offset+4(allocptr)
1783            store elem2, offset+8(allocptr)
1784            ...
1785            store elemn, offset+4n(allocptr)
1786    
1787       and the address computation part:
1788    
1789            celladdr <- offset+4+alloctpr
1790    
1791       and move the address computation part
1792    
1793       Effect:  register pressure is generally lower as a result.  This
1794                makes compilation of certain expressions much faster, such as
1795                long lists with non-trivial elements.
1796    
1797                 [(0,0), (0,0), .... (0,0)]
1798    
1799    3. MLRiscGen: base pointer elimination.
1800    
1801        As part of the linkage mechanism, we generate the sequence:
1802    
1803         L:  ...  <- start of the code fragment
1804    
1805         L1:
1806             base pointer <- linkreg - L1 + L
1807    
1808         The base pointer was then used for computing relocatable addresses
1809       in the code fragment.  Frequently (such as in lots of continuations)
1810       this is not needed.  We now eliminate this sequence whenever possible.
1811    
1812         For compile time efficiency, I'm using a very stupid local heuristic.
1813       But in general, this should be done as a control flow analysis.
1814    
1815       Effect:  Smaller code size.  Speed up of most programs.
1816    
1817    4. Hppa back end
1818    
1819         Long jumps in span dependence resolution used to depend on the existence
1820      of the base pointer.
1821    
1822         A jump to a long label L was expanded into the following sequence:
1823    
1824          LDIL %hi(L-8192), %r29
1825          LDO  %lo(L-8192)(%r29), %r29
1826          ADD  %r29, baseptr, %r29
1827          BV,n %r0(%r29)
1828    
1829         In the presence of change (3) above, this will not work.  I've changed
1830       it so that the following sequence of instructions are generated, which
1831       doesn't mention the base pointer at all:
1832    
1833             BL,n  L', %r29           /* branch and link, L' + 4 -> %r29 */
1834        L':  ADDIL L-(L'+4), %r29     /* Compute address of L */
1835             BV,n  %r0(%r29)          /* Jump */
1836    
1837    5. Alpha back end
1838    
1839          New alpha instructions LDB/LDW have been added, as per Fermin's
1840       suggestions.   This is unrelated to all other changes.
1841    
1842    6. X86 back end
1843    
1844         I've changed andl to testl in the floating point test sequence
1845         whenever appropriate.  The Intel optimization guide states that
1846         testl is preferable to andl.
1847    
1848    7. RA (x86 only)
1849    
1850         I've improved the spill propagation algorithm, using an approximation
1851       of maximal weighted independent sets.   This seems to be necessary to
1852       alleviate the negative effect in light of the slow down in (1).
1853    
1854         I'll write down the algorithm one of these days.
1855    
1856    8. MLRiscGen: frequencies
1857    
1858         I've added an annotation that states that all call gc blocks have zero
1859       execution frequencies.  This improves register allocation on the x86.
1860    
1861    BENCHMARKS
1862    ==========
1863    
1864       I've only perform the comparison on 110.25.
1865    
1866       The platforms are:
1867    
1868        HPPA  A four processor HP machine (E9000) with 5G of memory.
1869        X86   A 300Hhz Pentium II with 128M of memory, and
1870        SPARC An Ultra sparc 2 with 512M of memory.
1871    
1872       I used the following parameters for the SML benchmarks:
1873    
1874                 @SMLalloc
1875         HPPA    256k
1876         SPARC   512k
1877         X86     256k
1878    
1879    COMPILATION TIME
1880    ----------------
1881       Here are the numbers comparing the compilation times of the compilers.
1882       I've only compared 110.25 compiling the new sources versus
1883       a fixpoint version of the new compiler compiling the same.
1884    
1885                     110.25                                  New
1886               Total  Time in RA  Spill+Reload   Total  Time In RA Spill+Reload
1887         HPPA   627s    116s        2684+3584     599s    95s       1003+1879
1888         SPARC  892s    173s        2891+3870     708s    116s      1004+1880
1889         X86    999s    315s       94006+130691   987s    296s    108877+141957
1890    
1891                   110.25         New
1892                Code Size      Code Size
1893         HPPA   8596736         8561421
1894         SPARC  8974299         8785143
1895         X86    9029180         8716783
1896    
1897       So in summary, things are at least as good as before.   Dramatic
1898       reduction in compilation is obtained on the Sparc; I can't explain it,
1899       but it is reproducible.  Perhaps someone should try to reproduce this
1900       on their own machines.
1901    
1902    SML BENCHMARKS
1903    --------------
1904    
1905        On the average, all benchmarks perform at least as well as before.
1906    
1907          HPPA         Compilation Time     Spill+Reload      Run Time
1908                     110.25  New            110.25    New   110.25  New
1909    
1910          barnesHut  3.158  3.015  4.75%    1+1       0+0   2.980  2.922   2.00%
1911              boyer  6.152  5.708  7.77%    0+0       0+0   0.218  0.213   2.34%
1912       count-graphs  1.168  1.120  4.32%    0+0       0+0  22.705 23.073  -1.60%
1913                fft  0.877  0.792 10.74%    1+3       1+3   0.602  0.587   2.56%
1914        knuthBendix  3.180  2.857 11.32%    0+0       0+0   0.675  0.662   2.02%
1915             lexgen  6.190  5.290 17.01%    0+0       0+0   0.913  0.788  15.86%
1916               life  0.803  0.703 14.22%   25+25      0+0   0.153  0.140   9.52%
1917              logic  2.048  2.007  2.08%    6+6       1+1   4.133  4.008   3.12%
1918         mandelbrot  0.077  0.080 -4.17%    0+0       0+0   0.765  0.712   7.49%
1919             mlyacc 22.932 20.937  9.53%  154+181    32+57  0.468  0.430   8.91%
1920            nucleic  5.183  5.060  2.44%    2+2       0+0   0.125  0.120   4.17%
1921      ratio-regions  3.357  3.142  6.84%    0+0       0+0  116.225 113.173 2.70%
1922                ray  1.283  1.290 -0.52%    0+0       0+0   2.887  2.855   1.11%
1923             simple  6.307  6.032  4.56%   28+30      5+7   3.705  3.658   1.28%
1924                tsp  0.888  0.862  3.09%    0+0       0+0   7.040  6.893   2.13%
1925               vliw 24.378 23.455  3.94%  106+127    25+45  2.758  2.707   1.91%
1926      --------------------------------------------------------------------------
1927       Average                     6.12%                                   4.09%
1928    
1929          SPARC        Compilation Time     Spill+Reload      Run Time
1930                     110.25  New            110.25    New   110.25  New
1931    
1932          barnesHut  3.778  3.592  5.20%    2+2       0+0   3.648  3.453    5.65%
1933              boyer  6.632  6.110  8.54%    0+0       0+0   0.258  0.242    6.90%
1934       count-graphs  1.435  1.325  8.30%    0+0       0+0  33.672 34.737   -3.07%
1935                fft  0.980  0.940  4.26%    3+9       2+6   0.838  0.827    1.41%
1936        knuthBendix  3.590  3.138 14.39%    0+0       0+0   0.962  0.967   -0.52%
1937             lexgen  6.593  6.072  8.59%    1+1       0+0   1.077  1.078   -0.15%
1938               life  0.972  0.868 11.90%   26+26      0+0   0.143  0.140    2.38%
1939              logic  2.525  2.387  5.80%    7+7       1+1   5.625  5.158    9.05%
1940         mandelbrot  0.090  0.093 -3.57%    0+0       0+0   0.855  0.728   17.39%
1941             mlyacc 26.732 23.827 12.19%  162+189    32+57  0.550  0.560   -1.79%
1942            nucleic  6.233  6.197  0.59%    3+3       0+0   0.163  0.173   -5.77%
1943      ratio-regions  3.780  3.507  7.79%    0+0       0+0 133.993 131.035   2.26%
1944                ray  1.595  1.550  2.90%    1+1       0+0   3.440  3.418    0.63%
1945             simple  6.972  6.487  7.48%   29+32      5+7   3.523  3.525   -0.05%
1946                tsp  1.115  1.063  4.86%    0+0       0+0   7.393  7.265    1.77%
1947               vliw 27.765 24.818 11.87%  110+135    25+45  2.265  2.135    6.09%
1948      ----------------------------------------------------------------------------
1949       Average                     6.94%                                    2.64%
1950    
1951          X86          Compilation Time     Spill+Reload      Run Time
1952                     110.25  New            110.25    New   110.25  New
1953    
1954          barnesHut  5.530  5.420  2.03%  593+893   597+915   3.532  3.440   2.66%
1955              boyer  8.768  7.747 13.19%  493+199   301+289   0.327  0.297  10.11%
1956       count-graphs  2.040  2.010  1.49%  298+394   315+457  26.578 28.660  -7.26%
1957                fft  1.327  1.302  1.92%  112+209   115+210   1.055  0.962   9.71%
1958        knuthBendix  5.218  5.475 -4.69%  451+598   510+650   0.928  0.932  -0.36%
1959             lexgen  9.970  9.623  3.60% 1014+841  1157+885   0.947  0.928   1.97%
1960               life  1.183  1.183  0.00%  162+182   145+148   0.127  0.103  22.58%
1961              logic  3.285  3.512 -6.45%  514+684   591+836   5.682  5.577   1.88%
1962         mandelbrot  0.147  0.143  2.33%   38+41     33+54    0.703  0.690   1.93%
1963             mlyacc 35.457 32.763  8.22% 3496+4564 3611+4860  0.552  0.550   0.30%
1964            nucleic  7.100  6.888  3.07%  239+168   201+158   0.175  0.173   0.96%
1965      ratio-regions  6.388  6.843 -6.65% 1182+257   981+300  120.142 120.345 -0.17%
1966                ray  2.332  2.338 -0.29%  346+398   402+494   3.593  3.540   1.51%
1967             simple  9.912  9.903  0.08% 1475+941  1579+1168  3.057  3.178  -3.83%
1968                tsp  1.623  1.532  5.98%  266+200   250+211   8.045  7.878   2.12%
1969               vliw 33.947 35.470 -4.29% 2629+2774 2877+3171  2.072  1.890   9.61%
1970      ----------------------------------------------------------------------------
1971       Average                     1.22%                                     3.36%
1972    
1973    ----------------------------------------------------------------------
1974    Name: Allen Leung
1975    Date: 2000/03/23 16:25:00
1976    Tag: leunga-20000323-fix_x86_alpha
1977    Description:
1978    
1979    1. X86 fixes/changes
1980    
1981       a.  The old code generated for SETcc was completely wrong.
1982           The Intel optimization guide is VERY misleading.
1983    
1984    2. ALPHA fixes/changes
1985    
1986       a.  Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
1987       b.  Added a new mode byteWordLoadStores to the functor parameter to Alpha()
1988       c.  Added reassociation code for address computation.
1989    
1990    ----------------------------------------------------------------------
1991    Name: Allen Leung
1992  Date: 2000/03/22 01:23:00  Date: 2000/03/22 01:23:00
1993  Tag: leunga-20000322-fix_x86_hppa_ra  Tag: leunga-20000322-fix_x86_hppa_ra
1994  Description:  Description:
# Line 49  Line 2025 
2025    
2026  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2027  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2028  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
2029  explanation.  explanation.
2030    
2031  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2032  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2033  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2034  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 96  Line 2072 
2072  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
2073  <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,
2074  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
2075  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2076  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2077  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2078    
# Line 146  Line 2122 
2122    it from that remote directory.    it from that remote directory.
2123    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2124    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2125    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2126    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2127    
2128    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.580  
changed lines
  Added in v.752

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