aboutsummaryrefslogtreecommitdiffstats
path: root/lib/hipe
AgeCommit message (Collapse)Author
2016-09-05hipe: Reuse liveness between regalloc iterationsMagnus Lång
This is sound because the liveness data structure only stores liveness info at basic block boundaries, and the rewrites that happen in TargetSpecific:check_and_rewrite/2 preserves all existing definitions and uses, and all new liveness intervals, belonging to newly introduced temporaries, are always local to a basic block, and thus do not show up in the liveout or livein sets for the basic block.
2016-09-05hipe: Add ra_partitioned to o1 and upMagnus Lång
ra_partitioned significantly speeds up register allocation of larger functions without affecting allocation quality negatively. This is the final change needed to make o1 suitable for compiling really large functions without choking.
2016-09-05hipe_regalloc_prepass: Change splitting heuristicMagnus Lång
The division into an initial pass that may introduce temps, and following passes that must not forces us to make the same heuristic decision during each of these passes. Thus, the splitting heuristic can't be based on the number of temporaries -- at least without excluding temporaries above SpillLimit.
2016-09-02hipe: Make sure prepass temps are below SpillLimitMagnus Lång
If temps introduced by hipe_regalloc_prepass end up above SpillLimit, the register allocators will not spill them. This constraint is unnecessarily limiting the allocators and might theoretically lead to unallocatable programs (more temps above SpillLimit alive at a time than there are physical registers).
2016-09-02hipe_regalloc_prepass: Rename coloring collisionsMagnus Lång
2016-09-02hipe_ppc: Add code rewrite RA callbacksMagnus Lång
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.
2016-09-02hipe_sparc: Add code rewrite RA callbacksMagnus Lång
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.
2016-09-02hipe_arm: Add code rewrite RA callbacksMagnus Lång
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.
2016-09-02hipe_x86: Add code rewrite RA callbacksMagnus Lång
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.
2016-09-02hipe: Remove defun_to_cfg/1 RA callbackMagnus Lång
Now that all backends do register allocation on a CFG directly and define the defun_to_cfg/1 callback as the identity function, it can be removed.
2016-09-02Add new sanity assertion to hipe_regalloc_prepassMagnus Lång
As the just_as_good_as assertion was loosened with the `NowRegs >= CheckRegs` check, it no longer verified that hipe_regalloc_prepass had not incorrectly labeled a temp as unallocatable. We add that behaviour back.
2016-09-02Simplify hipe_x86_ra_finalise:conv_ra_maplet/3Magnus Lång
2016-09-02hipe_x86: Simplify ra_postconditions is_mem_opndMagnus Lång
This is due to the improvements in hipe_temp_map, removing the need for duplicated logic in the backends.
2016-09-02hipe_x86: Fix pseudo_tailcall prettyprintingMagnus Lång
2016-09-02hipe_x86: Extra sanity assertionsMagnus Lång
2016-09-02hipe: clean up unnecessary catchesMagnus Lång
2016-09-02hipe: Remove temp reuse from call_funMagnus Lång
2016-09-02hipe: Add IG partitioning to hipe_regalloc_prepassMagnus Lång
2016-09-02hipe: Add hipe_regalloc_prepassMagnus Lång
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.
2016-09-02Merge branch 'sverker/hipe-performance-o1/PR-1154/OTP-13862'Sverker Eriksson
* sverker/hipe-performance-o1/PR-1154: hipe_sparc: Minimise CFG<->linear conversions hipe_ppc: Minimise CFG<->linear conversions hipe_arm: Minimise CFG<->linear conversions hipe_x86: Use lea instead of move+add hipe_arm: Improve peephole optimiser hipe_arm: Be resilient to crappy RTL hipe_ppc: Be resilient to crappy RTL hipe_sparc: Be resilient to crappy RTL hipe: Reuse liveness info for spillmin hipe_x86: Minimise CFG<->linear conversions hipe: Fix o0 and o1 hipe: Add o0 and o1 to tests hipe_rtl_binary:get_word_integer/4: Handle imms hipe_x86: Be resilient to crappy RTL hipe_x86: LSRA for SSE2
2016-09-02Merge branch 'maint'Sverker Eriksson
2016-09-02Merge branch 'sverker/hipe-sparc-19/PR-1148/OTP-13861' into maintSverker Eriksson
* sverker/hipe-sparc-19/PR-1148: Eliminate catch-all clause from two functions Increase the time limit used by the test suite
2016-09-01Merge branch 'maint'Hans Bolinder
* maint: dialyzer: Increase time limit of suites dialyzer: Remove a check that always fails dialyzer: Optimize an opaque type case
2016-08-31dialyzer: Optimize an opaque type caseHans Bolinder
Fix a mistake in commit 85f6fe3b. Instead of using the declared opaque type, the form's type is used in a case where the opaque type is turned into a non-opaque type. The result is more general types (smaller Erlang terms) and faster analyses.
2016-08-30hipe_sparc: Minimise CFG<->linear conversionsMagnus Lång
Now, there will only ever be a single Linear->CFG conversion, just after lowering from RTL, and only ever a single CFG->Linear conversion, just before the finalise pass. Both of these now happen in hipe_sparc_main.
2016-08-30hipe_ppc: Minimise CFG<->linear conversionsMagnus Lång
Now, there will only ever be a single Linear->CFG conversion, just after lowering from RTL, and only ever a single CFG->Linear conversion, just before the finalise pass. Both of these now happen in hipe_ppc_main.
2016-08-30hipe_arm: Minimise CFG<->linear conversionsMagnus Lång
Now, there will only ever be a single Linear->CFG conversion, just after lowering from RTL, and only ever a single CFG->Linear conversion, just before the finalise pass. Both of these now happen in hipe_arm_main.
2016-08-30hipe_x86: Use lea instead of move+addMagnus Lång
This is primarily useful for heap allocations, as a two-address 'add' can't be used to both copy the heap pointer to another register, and add the tag.
2016-08-30hipe_arm: Improve peephole optimiserMagnus Lång
2016-08-30hipe_arm: Be resilient to crappy RTLMagnus Lång
The ARM backend crashes if certain RTL optimisations were omitted, preventing it from being usable at lower optimisation levels. One of the problems were caused by shift-by-immediate-zero, which wraps to immediate-32 with some shiftops. TODO: Someplace should be modified to crash when these are generated so debuging further instances of this gets easier in the future.
2016-08-30hipe_ppc: Be resilient to crappy RTLMagnus Lång
The PowerPC backend crashes if certain RTL optimisations were omitted, preventing it from being usable at lower optimisation levels.
2016-08-30hipe_sparc: Be resilient to crappy RTLMagnus Lång
The SPARC backend crashes if certain RTL optimisations were omitted, preventing it from being usable at lower optimisation levels.
2016-08-30hipe: Reuse liveness info for spillminMagnus Lång
For x86, additionally reuse liveness from float LSRA for the GP LSRA.
2016-08-30hipe_x86: Minimise CFG<->linear conversionsMagnus Lång
Most x86 passes were either linearise(pass(to_cfg(Code))) or trivially rewritable to process a CFG. This saves a great deal of time and memory churn when compiling large programs. Now, there will only ever be a single Linear->CFG conversion, just after lowering from RTL, and only ever a single CFG->Linear conversion, just before the finalise pass. Both of these now happen in hipe_x86_main.
2016-08-30hipe: Fix o0 and o1Magnus Lång
These options would not do anything, because they would not supress the 'o2' in ?COMPILE_DEFAULTS. Such behaviour is added to expand_options/2.
2016-08-30hipe: Add o0 and o1 to testsMagnus Lång
Now that x86 is no longer broken with these optimisation levels, we add them to the test suite to ensure they do not break again. Bump timeout to 6min since tests are run twice as many times. The option set of o1 was changed to all optimisations that run fast on both big and small programs, incurring only a slight compile time increase compared to the old set, but with a, presumably, significant improvement to speed of compiled code. Change o0 register allocator to linear_scan.
2016-08-30hipe_rtl_binary:get_word_integer/4: Handle immsMagnus Lång
Immediate arguments to get_word_integer/4 would lead to bad but unreachable RTL being generated. We omit its generation by testing for immediates and performing the logic at compile time.
2016-08-30hipe_x86: Be resilient to crappy RTLMagnus Lång
The x86 backend crashes if certain RTL optimisations were omitted, preventing it from being usable at lower optimisation levels.
2016-08-30hipe_x86: LSRA for SSE2Magnus Lång
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.
2016-08-26Merge branch 'maint'Sverker Eriksson
2016-08-26Eliminate catch-all clause from two functionsKostis Sagonas
A stronger version of Dialyzer complained that some case clauses in functions xaluop_is_shift/1 and xaluop_normalise/1 are unreachable. These clauses are now commented out. While at it, I thought that it would be better to eliminate the catch-all clauses in order to be certain we properly handle all RTL instructions that are used as inputs to these functions. Note: The code will now crash if there are unhandled cases.
2016-08-25Increase the time limit used by the test suiteKostis Sagonas
This is required in some really old SPARC machines running Solaris we still have access to.
2016-08-22hipe: Fix amd64 SSE2 encoding crashMagnus Lång
Register allocation could transform something like fmove u32, d99 to fmove $rdx, 0x20($rsp) which is an invalid instruction.
2016-08-22hipe: Fix tailcall stackarg clobber bugMagnus Lång
Since the link register/return address is restored before stack arguments are stored to the frame, we must not use it to store a stack argument. We do that by adding it to the registers clobbered by pseudo_tailcall_prepare.
2016-08-22hipe_arm: Fix translation of shift by 0Magnus Lång
The problem was caused by shift-by-immediate-zero, which wraps to immediate-32 with some shiftops. TODO: Someplace should be modified to crash when these are generated so debugging further instances of this gets easier in the future.
2016-08-22hipe_ppc: Fix PPC64 bug encoding large immediatesMagnus Lång
2016-08-22hipe_ppc: Fix incorrect encoding of shift by 0Magnus Lång
2016-08-22hipe_x86: Fix illegal inst from peephole optMagnus Lång
2016-07-11hipe_vectors: Change implementation to 'array'Magnus Lång
The 'array' module is highly optimised for the hipe_vectors use-case, and seems to perform slightly better than the gb_trees implementation. Also, we remove the completely unnecessary hipe_vectors.hrl header.
2016-07-11hipe: Faster unreachable basic block removalMagnus Lång