| Age | Commit message (Collapse) | Author |
|---|
|
|
|
hipe_range_split is a complex live range splitter, more sophisticated
thatn hipe_restore_reuse, but still targeted specifically at temporaries
forced onto stack by being live over call instructions.
hipe_range_split partitions the control flow graph at call instructions,
like hipe_regalloc_prepass. Splitting decisions are made on a per
partition and per temporary basis.
There are three different ways in which hipe_range_split may choose to
split a temporary in a program partition:
* Mode1: Spill the temp before calls, and restore it after them
* Mode2: Spill the temp after definitions, restore it after calls
* Mode3: Spill the temp after definitions, restore it before uses
To pick which of these should be used for each temp×partiton pair,
hipe_range_split uses a cost function. The cost is simply the sum of the
cost of all expected stack accesses, and the cost for an individual
stack access is based on the probability weight of the basic block that
it resides in. This biases the range splitter so that it attempts moving
stack accesses from a functions hot path to the cold path.
hipe_bb_weights is used to compute the probability weights.
mode3 is effectively the same as what hipe_restore_reuse does. Because
of this, hipe_restore_reuse reuses the analysis pass of
hipe_restore_reuse in order to compute the minimal needed set of spills
and restores. The reason mode3 was introduced to hipe_range_split rather
than simply composing it with hipe_restore_reuse (by running both) is
that such a composition resulted in poor register allocation results due
to insufficiently strong move coalescing in the register allocator.
The cost function heuristic has a couple of tuning knobs:
* {range_split_min_gain, Gain} (default: 1.1, range: [0.0, inf))
The minimum proportional improvement that the cost of all stack
accesses to a temp must display in order for that temp to be split.
* {range_split_mode1_fudge, Factor} (default: 1.1, range: [0.0, inf))
Costs for mode1 are multiplied by this factor in order to discourage
it when it provides marginal benefits. The justification is that
mode1 causes temps to be live for longest, thus leading to higher
register pressure.
* {range_split_weight_power, Factor} (default: 2, range: (0.0, inf))
Adjusts how much effect the basic block weights have on the cost of a
stack access. A stack access in a block with weight 1.0 has cost 1.0,
a stack access in a block with weight 0.01 has cost 1/Factor.
Additionally, the option range_split_weights chooses whether the basic
block weights are used at all.
In the case that the input is very big, hipe_range_split automatically
falls back to hipe_restore_reuse only in order to keep compile times
under control. Note that this is not only because of hipe_range_split
being slow, but also due to the resulting program being slow to register
allocate, and is not as partitionable by hipe_regalloc_prepass.
hipe_restore_reuse, on the other hand, does not affect the programs
partitionability.
The hipe_range_split pass is controlled by a new option ra_range_split.
ra_range_split is added to o2, and ra_restore_reuse is disabled in o2.
|
|
hipe_bb_weights computes basic block weights by using the branch
probability predictions as the coefficients in a linear equation system.
This linear equation system is then solved using Gauss-Jordan
Elimination.
The equation system representation is picked to be efficient with highly
sparse data. During triangelisation, the remaining equations are
dynamically reordered in order to prevent the equations from growing in
the common case, preserving the benefit of the sparse equation
representation.
In the case that the input is very big, hipe_bb_weights automatically
falls back to a rough approximation in order to keep compile times under
control.
|
|
hipe_restore_reuse is a simplistic range splitter that splits temps that
are forced onto the stack by being live over call instructions. In
particular, it attempts to avoid cases where there are several accesses
to such stack allocated temps in straight-line code, uninterrupted by
any calls. In order to achieve this it splits temps between just before
the first access(es) and just after the last access(es) in such
straight-line code groups.
The hipe_restore_reuse pass is controlled by a new option
ra_restore_reuse.
ra_restore_reuse is added to o1.
|
|
|
|
* rickard/time-unit/OTP-13831:
Replace usage of deprecated time units
|
|
These will not only be useful for hipe_regalloc_prepass, but also, after
the introduction of a mk_move/2 (or similar) callback, for the purpose
of range splitting.
Since the substitution needed to case over all the instructions, a new
module, hipe_ppc_subst, was introduced to the ppc backend.
|
|
These will not only be useful for hipe_regalloc_prepass, but also, after
the introduction of a mk_move/2 (or similar) callback, for the purpose
of range splitting.
Since the substitution needed to case over all the instructions, a new
module, hipe_sparc_subst, was introduced to the sparc backend.
|
|
These will not only be useful for hipe_regalloc_prepass, but also, after
the introduction of a mk_move/2 (or similar) callback, for the purpose
of range splitting.
Since the substitution needed to case over all the instructions, a new
module, hipe_arm_subst, was introduced to the arm backend.
|
|
These will not only be useful for hipe_regalloc_prepass, but also, after
the introduction of a mk_move/2 (or similar) callback, for the purpose
of range splitting.
Since the substitution needed to case over all the instructions, a new
module, hipe_x86_subst, was introduced to the x86 backend.
Due to differences in the 'jtab' field of a #jmp_switch{} between x86
and amd64, it regrettably needed to be duplicated to hipe_amd64_subst.
|
|
hipe_regalloc_prepass speeds up register allocation by spilling any temp
that is live over a call (which clobbers all register).
In order to detect these, a new function was added to the target
interface; defines_all_alloc/1, that takes an instruction and returns a
boolean.
|
|
There is little point offering LSRA for x86 if we're still going to call
hipe_graph_coloring_regalloc for the floats. In particular, all
allocators except LSRA allocates an N^2 interference matrix, making them
unusable for really large functions.
|
|
|
|
This speeds up parentsOfChild/2 from O(n) to O(lg n + k).
A new module misc/hipe_segment_trees.erl is introduced.
|
|
* Implemented removal of maps:is_key/2 calls of which the result is
known in a new pass during the typed phase, called
hipe_icode_call_elim.
* Added the option icode_call_elim that enables the
hipe_icode_call_elim pass, and made it default for o2.
|
|
Instead ask running VM for the value of THE_NON_VALUE,
which is different between opt and debug VM.
Same hipe compiler can now compile for both opt and debug VM.
|
|
|
|
Applications that use the new erl_anno module are depending on STDLIB 2.5.
Note that CosNotification, Megaco, SNMP, Xmerl, and Parsetools use the
erl_anno module via the Yecc parsers only (the header file in
lib/parsetools/include/yeccpre.hrl calls the erl_anno module).
HiPE does not call the erl_anno module, but uses an exported type.
We have chosen to make HiPE dependent on the erl_anno module.
|
|
|
|
* yiannist/hipe-llvm-backend:
Support the LLVM backend in HiPE
Implement the LLVM backend
Extend RTL API to support the LLVM backend
Add support for llvm unique symbols in hipe_gensym
Add a BIF that only returns the atom ok
Move some common code in hipe_pack_constants
Add better specs in hipe_pack_constants and cleanup
OTP-11801
|
|
Most dependencies introduced are exactly the dependencies to other
applications found by xref. That is, there might be real dependencies
missing. There might also be pure debug dependencies listed that
probably should be removed. Each application has to be manually
inspected in order to ensure that all real dependencies are listed.
All dependencies introduced are to application versions used in
OTP 17.0. This since the previously used version scheme wasn't
designed for this, and in order to minimize the work of introducing
the dependencies.
|
|
Add flags to enable and use the LLVM backend:
* to_llvm: use the LLVM pipeline for compilation (default optimization level
is O3),
* llvm_save_temps: save the intermediate files in current directory in order
to be able to debug or optimize the LLVM assembly,
* {to_llvm, O}: set the optimization level of LLVM opt and llc tools.
Add some debug support to the loader; no semantic change intented.
|
|
|
|
The hybrid heap emulator was last working in the non-SMP R11B
run-time system. When the constant pools were introduced in R12B,
the hybrid heap emulator was not updated to handle them.
At this point, the harm from reduced readability of the code is
greater than any potential usefulness of keeping the code.
|
|
|
|
hipe_ceach.erl has disappeared, but is still referred to from
hipe.app.src, causing reltool-generated releases crash when they
cannot find the module.
|
|
|
