Estimated hours taken: 30+
arg_info:
Fix allocation to work properly for --args compact.
bytecode*:
Handle complex deconstruction unifications. Not really tested because
I can't find a test case.
bytecode_gen, call_gen, code_util:
Use the new method to handle builtin predicates/functions. We now
handle reverse mode arithmetic and unary plus/minus as builtins.
code_gen, code_init, follow_vars, hlds_pred:
Put back the initial follow_vars field of the proc_info, since this
may allow the code generator to emit better code at the starts of
of predicates.
inlining:
Don't inline recursive predicates.
goals_util:
Add a predicate to find out if a goal calls a particular predicate.
Used in inlining to find out if a predicate is recursive.
unused_args:
Remove code that used to set the mode of unused args to free->free.
Since this changes the arg from top_in to top_unused *without* code
in other modules being aware of the change, this screws up --args
compact.
llds, llds_out, garbage_out:
Prepare for the move to the new type_info structure by adding a new
"module" type for defining structures holding type_infos. Not
currently generated or output.
llds, opt_debug, opt_util, vn_type, vn_cost, vn_temploc:
Change the argument of temp to be a reg, not an int, allowing
floating point temporaries.
vn_type:
Add information about the number of floating point registers and
temporaries to the parameter structure (these are currently unused).
llds, dupelim, frameopt, livemap, middle_rec, value_number, vn_filter,
vn_verify:
Add an extra field to blocks giving the number of float temporaries.
options:
Add parameters to configure the number of floating point registers
and temporaries.
mercury_compile:
Add an extra excess assign phase at the start of the middle pass.
This should reduce the size of the code manipulated by the other
phases, and gives more accurate size information to inlining.
(The excess assign phase before code generation is I think still
needed since optimizations can introduce such assignments.)
value_number:
Optimize code sequences before and after assignments to curfr
separately, since such assignments change the meaning of framevars.
This fixes the bug that caused singleton variable warnings to contain
garbage.
vn_block, vn_flush, vn_order, vn_util:
Add special handling of assignments to curfr. This is probably
unnecessary after my change to value_number, and will be removed
again shortly :-(
vn_flush:
Improve the code generated by value numbering (1) by computing values
into the place that needs them in some special circumstances, and
(2) by fixing a bug that did not consider special registers to be
as fast as r1 etc.
vn_util:
Improve the code generated by value numbering by removing duplicates
from the list of uses of a value before trying to find out if there
is more than one use.
simplify:
Avoid overzealous optimization of main --> { ..., error(...) }.
handle_options:
Fix an error message.
code_aux:
Break an excessive long line.
Estimated hours taken: 5
peephole:
Fixed a bug that caused restores of succip to be put in the wrong
place, but only after predicate-wide value numbering.
opt_debug:
Added a couple of debugging predicates used in tracking down this bug.
value_number:
Fix a bug that left a livevals pseudo-op in the wrong place if a
single instruction sequence contained more than one such pseudo-op.
options:
Add --debug-opt. Rename --vndebug to --debug-vn.
Add --generate-bytecode.
optimize, vn_debug:
Use the new routines in opt_debug, and use the new/renamed options.
store_alloc:
Don't thread follow_vars through the module, since the follow_vars
information is not attached directly to branched structures. We
now also use the same slot to hold the store map computed by this
pass; this should allow the later deletion of the store map slot
from goal_infos.
follow_code:
Removed dead predicate.
livemap:
Added a comment.
Estimated hours taken: 20
det_analysis:
Make sure we don't change the goal being analyzed except possibly
for the introduction of `some's (which should not hurt anything).
Make sure we don't print any error messages except in the final
iteration, when all the inputs to the inference are stable.
If the --debug-detism options is set, print messages about the
progress of inference and checking.
Also moved some code around.
det_report:
Distinguish the handling of warning messages and error messages.
simplify:
Use the new ability of det_report to separate warnings and errors.
passes_aux:
Add a new generic pass form, for use by simplify.
option:
Add --debug-detism (as above), --aditi, which at the moment
only enables the disjunctive normal form transformation, and
--inlining/--no-inlining, which set the other three flags
involved in inlining depending on whether you want standard
inlining or none at all.
Follow_code used to be set twice and follow_vars not at all;
I fixed this.
Reenabled optimize_higher_order at -O3.
Moved value numbering to -O4 and pred_value_number to -O5.
This makes it easier to separate value numbering from the
other optimizations (which are likely to be more effective).
Divided options_help into sections to avoid excessive
compilation times.
store_alloc:
Base the store map on the follow_vars info attached to the
branched structure which I added recently, and not on the
follow_vars map being passed around, since it will be more accurate.
hlds_out:
Print information about follow_vars and store_maps when -D is given.
follow_code:
Undo an old hack that change to follow_vars has made counterproductive.
middle_rec:
Fix a bug uncovered by the change to follow_code. When looking for a
register to hold the counter, it is not enough to avoid picking a
register that appears in the recursive case; we must also avoid
registers that occur only in the base case.
livemap:
Mentioning the code address succip now causes the succip to be
considered live. This may or may not fix the bug with pred_value_number
miscompiling unused_args.m; the other changes have caused the input
to value numbering to change, and they no longer trigger the problem.
(Will try to test this later.)
mercury_compile:
Try to make sure that we print statistics only after passes that
were actually executed. Also, reduce the number of lookups of the
verbose option. Move some predicates so that the order of their
appearance matches the current order of invocation.
vn_table:
Loosen a sanity check to let xnuc2 pass through it.
code_exprn, switch_detection:
Minor changes.
Estimated hours taken: 1.5
Split llds into two parts. llds.m defines the data types, while llds_out.m
has the predicates for printing the code.
Removed the call_closure instruction. Instead, we use calls to the
system-defined addresses do_call_{det,semidet,nondet}_closure. This is
how call_closure was implemented already. The advantage of the new
implementation is that it allows jump optimization of what used to be
call_closures, without new code in jumpopt.
A Constraint Solver Interface For Mercury
<thunderous applause>
Estimated hours taken: 1 summer studentship
This is the implementation of a fairly general constraint solver interface. If
using a library grade *.cnstr, we emit C instructions to keep track of the
solver's implicit state. This is done by storing and restoring 'tickets' -
abstract handles on the solver's state.
We emit a store_ticket() macro:
-when entering the first disjunct of a disjunction
-when entering the condition of an if-then-else
We emit a restore_ticket() macro:
-when entering a disjunct other than the first of a disjunction
-when entering the else part of an if-then-else
We emit a discard_ticket() macro:
-after the restore_ticket() in the final disjunct of a disjunction
-at the start of the 'then' part of an if-then-else
The rules for emitting the macros is slightly more complicated than that shown
above for if-then-elses (determinism of the parts must be taken into account).
compiler/code_info.m:
Get an llds store_ticket/restore_ticket etc. instruction
compiler/disj_gen.m:
Emit ticket macros in the appropriate places in a disjunction.
compiler/dupelim.m:
Handle the new llds instruction.
compiler/frameopt.m:
Handle the new llds instruction.
compiler/handle_options.m:
If the grade is *.cnstr, set the constraints option on.
compiler/ite_gen.m:
Emit ticket macros in the appopriate places in an if-then-else.
compiler/livemap.m:
Handle the new llds instruction.
compiler/llds.m:
Output the ticket macros.
compiler/make_hlds.m:
An irrelevant tidy-up.
compiler/mercury_compile.pp:
If the grade is *.cnstr, pass -DCONSTRAINTS to mgnuc
compiler/middle_rec.m:
Handle the new llds instruction.
compiler/opt_*.m:
Handle the new llds instruction.
compiler/options.m:
Introduce a new boolean option 'constraints'.
compiler/shapes.m:
Output a new shape - 'ticket'.
compiler/unify_proc.m:
Handle the new llds instruction.
compiler/v*.m:
Handle the new llds instruction.
Estimated hours taken: 1.5
Undo dylan's changes in the names of some library entities,
by applying the following sed script
s/term_atom/term__atom/g
s/term_string/term__string/g
s/term_integer/term__integer/g
s/term_float/term__float/g
s/term_context/term__context/g
s/term_functor/term__functor/g
s/term_variable/term__variable/g
s/_term__/_term_/g
s/std_util__bool_/bool__/g
to all the `.m' and `.pp' files in the compiler and library directories.
The reason for undoing these changes was to minimize incompatibilities
with 0.4 (and besides, the changes were not a really good idea in the first
place).
I also moved `bool' to a separate module.
The main reason for that change is to ensure that the `__' prefix is
only used when it genuinely represents a module qualifier.
(That's what dylan's changes were trying to acheive, but `term__'
does genuinely represent a module qualifier.)
compiler/*.m:
Apply sed script above;
where appropriate, add `bool' to the list of imported modules.
The changes made allow declarations of the form:
:- pragma(c_code, predname(Varname1::mode1, Varname2::mode2, ...),
"Some C code to execute instead of a mercury clause;").
There are still a couple of minor problems to be fixed in the near future:
If there is a regular clause given as well as a pragma(c_code, ...) dec, it
is not handled well, and variables names '_' are not handled well.
prog_io.m:
parse the pragma(c_code, ...) dec.
hlds.m:
define a new hlds__goal_expr 'pragma_c_code'.
make_hlds.m:
insert the pragma(c_code, ...) dec. as a pragma_c_code into the hlds.
det_analysis.m:
infer that pragma_c_code goals are det.
modes.m:
convince the mode checker that the correct pragma variables are bound
etc.
quantification.m:
quantify the variables in the pragma(c_code, ...) dec.
code_gen.pp:
convert pragma_c_code into pragma_c (in the llds).
llds.m:
define a new instr, pragma_c. Output the pragma_c
hlds_out.m:
mercury_to_mercury.m:
mercury_to_goedel.m:
spit out pragma(c_code, ...) decs properly
*.m: handle the new pragma_c_code in the hlds or the new pragma_c in the llds
instructions, and the last argument from local labels. All these were
placeholders for info put in there by prof.m and used when emitting C
code.
The set of labels that serve as return points are now calculated in llds.m
just before each procedure has its C code generated. This set is passed to
output_instruction along with the label at the start of the procedure.
This set of changes includes most of the work necessary for
mode and determinism checking of higher-order predicates.
prog_io.m:
Change the syntax for lambda expressions: they need
to have a determinism declaration. Lambda
expressions must now look like this:
lambda([X::in, Y::out] is det, ...goal...).
^^^^^^
Note that both the modes and the determinism are mandatory,
not optional.
hlds.m:
Insert a determinism field in the lambda_goal structure.
hlds_out.m, inlining.m, make_hlds.m, modes.m, polymorphism.m, quantification.m,
switch_detection.m, typecheck.m:
Modified to use lambda_goal/4 rather than lambda_goal/3.
prog_io.m:
Add a new field to the `ground' inst, of type `maybe(pred_inst_info)'.
We use this to store the modes and determinism of higher-order
predicate terms.
code_info.m, inst_match.m, mercury_to_mercury.m, mode_util.m, modes.m,
polymorphism.m, shapes.m, undef_modes.m:
Modified to handle higher-order pred modes:
use ground/2 rather than ground/1.
(Note that modes.m still requires a bit more work on this.)
llds.m:
Add a new field to the call_closure/3 instruction to hold the
caller address for use with profiling, since the C macros
require a caller address.
dup_elim.m, frame_opt.m, garbage_out.m, live_map.m, middle_rec.m, opt_debug.m,
opt_util.m, value_number.m, vn_*.m:
Modified to use call_closure/4 rather than call_closure/3.
mercury_to_mercury.m:
Export mercury_output_det for use by hlds_out.m.
-------------------------------------------------------
Implement unique modes. We do not handle local aliasing yet, so this
is still not very useful, except for io__state. Destructive update is
not yet implemented. Also note that this really only implements
"mostly unique" variables that may be non-unique on backtracking - we
don't check that you don't backtrack over I/O, for example.
prog_io.m, mode_util.m, modes.m, inst_match.m:
Major changes to Handle unique modes.
mercury_to_mercury.m, polymorphism.m, prog_out.m, undef_modes.m:
Use `ground(Uniqueness)' rather than just `ground'.
compiler/*.m:
Fix compile errors now that unique modes are enforced: add a
few calls to copy/2, and comment out lots of unique mode
declarations that caused problems.
typecheck.m, mode_info.m:
Hack around the use of unique modes, which doesn't work
because we don't allow local aliasing yet: make the insts
`uniq_type_info' and `uniq_mode_info' not unique at all,
and add a call to copy/2 when extracting the io_state from
type_info or mode_info.
-------------------------------------------------------
Plus a couple of unrelated changes:
hlds.m:
Change the modes for the special predicates from `ground -> ground'
to `in', so that any error messages that show those modes
come out looking nicer.
Add a new shared_inst_table for shared versions of user-defined
insts.
mercury_to_goedel.m:
Use string__is_alnum_or_underscore.
code_gen.pp:
Put the comment about the contents of stack slots before the initial
label, since this way it will be preserved by optimizations.
cse_detection.m:
Extended the search to look for cses in if-then-elses and switches
as well as disjunctions. Removed InstmapDelta from preds in which it
was not being used.
det_analysis.m:
Make the diagnosis routines more robust. The changes here avoid the
Philip's problems with lexical.m.
jumpopt.m:
Minor formatting changes.
livemap.m:
Avoid duplicating livevals instructions when optimizations are
repeated, since this can confuse some optimizations.
llds.m:
Minor documentation change.
make_hlds.m:
Minor formatting change.
mercury_compile.pp:
Do not map arguments to registers if any semantic errors have been
found.
middle_rec.m and code_aux.m:
Apply middle recursion only if tail recursion is not possible,
since tail recursion yields more efficient code.
opt_util.m:
Added a predicate to recognize constant conditions in if_vals.
Modified a predicate to make it better suited for frameopt.
optimize.pp:
Changed the way optimizations were repeated to allow better control.
Repeat peephole once more after frameopt, since the new frameopt
can benefit from this.
options.m:
Removed the --compile-to-c option, which was obsolete. Added an
option for predicate-wide value numbering, which is off by default.
Changed some of the default values of optimization flags to reduce
compilation time while holding the loss of speed of generated code
to a minimum.
peephole.m:
Look for if_vals whose conditions are constants, and eliminate the
if_val or turn it into a goto depending on the value of the constant.
Generalized the condition for optimizing incr_sp/decr_sp pairs.
value_number.m:
Added a prepass to separate primary tag tests in if-then-elses from
the test of the secondary tag, which requires dereferencing the
pointer.
Added sanity check routines to test two aspects of the generated code.
First, whether it produces the same values for the live variables as
the original code, and second, whether it has moved any dereferences
of a pointer before a test of the tag of that pointer. If either test
fails, we use the old instruction sequence.
vn_debug.m:
New messages to announce the failure of the sanity checks. They are
enabled by default, but of course can only appear if value numbering
is turned on (it is still off by default).
vn_flush.m:
Threaded a list of forbidden lvals (lvals that may not be assigned to)
through the flushing routines. When saving the old value of an lval
that is being assigned to, we use this list to avoid modifying any of
the values used on the right hand side of the assignment, even if the
saving of an old value results in assignment that requires another
save, and so on recursively.
When the flushing of a node_lval referred to a shared vn, the uses of
the access vns of the node_lvals were not being adjusted properly.
Now they are.
vn_order.m:
The ctrl_vn phase of the ordering was designed to ensure that all
nodes that need not come before a control node come after it. However,
nodes were created after this phase operated, causing leakage of some
value nodes in front of control nodes. Some of these led to pointer
dereferences before tag tests, causing bus errors. The ctrl_vn phase
is now last to avoid this problem.
vn_table.m:
Added an extra interface predicate to support the sanity checks in
value_number.
vn_util.m:
The transformation of c1-e2 into (0-e2)+c1 during vnrval simplification
could lead to an infinite loop in the compiler if c1 was zero. A test
for this case now prevents the loop.
frameopt:
fix the problem with destroying stack frames and creating
them again later, accessing detstackvars that were earlier
nominally destroyed.
vn_livemap:
renamed it to livemap since frameopt now uses it also.
value_number, vn_*:
Fixed some bugs. Reorganized the handling of blocks: they are now
put in at the last minute before llds writes out the code.
Made a start towards exploiting info about cheaper copies of
values.
optimize, options:
Made value_numbering an iterated optimization. Added a new
option to control how many times it is iterated together
with other the jumpopt, peephole and labelopt.
llds, call_gen, code_gen, code_info, middle_rec, opt_debug:
changed type of the argument of livevals to plain set.
Warning: in more than a week I haven't been able to fully test this change,
dur to kryten's flakiness and bugs upstream of the optimizer.
