| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
| |
function so that it can live in Analysis instead of
VMCore.
llvm-svn: 121885
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
1. Fix pre-ra scheduler so it doesn't try to push instructions above calls to
"optimize for latency". Call instructions don't have the right latency and
this is more likely to use introduce spills.
2. Fix if-converter cost function. For ARM, it should use instruction latencies,
not # of micro-ops since multi-latency instructions is completely executed
even when the predicate is false. Also, some instruction will be "slower"
when they are predicated due to the register def becoming implicit input.
rdar://8598427
llvm-svn: 118135
|
|
|
|
|
|
| |
fallthroughs uses all registers, just gather the union of all successor liveins.
llvm-svn: 117506
|
|
|
|
|
|
| |
regression.
llvm-svn: 117329
|
|
|
|
|
|
|
|
|
|
|
| |
2) live-outs.
Previously the post-RA schedulers completely ignore these dependencies since
returns, branches, etc. are all scheduling barriers. This patch model the
latencies between instructions being scheduled and the barriers. It also
handle calls by marking their register uses.
llvm-svn: 117193
|
|
|
|
| |
llvm-svn: 116119
|
|
|
|
|
|
|
|
|
|
| |
implicit. e.g.
%D6<def>, %D7<def> = VLD1q16 %R2<kill>, 0, ..., %Q3<imp-def>
%Q1<def> = VMULv8i16 %Q1<kill>, %Q3<kill>, ...
The real definition indices are 0,1.
llvm-svn: 116080
|
|
|
|
| |
llvm-svn: 115802
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
allow target to correctly compute latency for cases where static scheduling
itineraries isn't sufficient. e.g. variable_ops instructions such as
ARM::ldm.
This also allows target without scheduling itineraries to compute operand
latencies. e.g. X86 can return (approximated) latencies for high latency
instructions such as division.
- Compute operand latencies for those defined by load multiple instructions,
e.g. ldm and those used by store multiple instructions, e.g. stm.
llvm-svn: 115755
|
|
|
|
|
|
| |
pipeline forwarding path.
llvm-svn: 115098
|
|
|
|
|
|
|
|
|
|
|
| |
take multiple cycles to decode.
For the current if-converter clients (actually only ARM), the instructions that
are predicated on false are not nops. They would still take machine cycles to
decode. Micro-coded instructions such as LDM / STM can potentially take multiple
cycles to decode. If-converter should take treat them as non-micro-coded
simple instructions.
llvm-svn: 113570
|
|
|
|
|
|
|
| |
instead of fixed size arrays, so that increasing FirstVirtualRegister to 16K
won't cause a compile time performance regression.
llvm-svn: 109330
|
|
|
|
| |
llvm-svn: 108452
|
|
|
|
|
|
| |
more than one dbg_value instruction. rdar://7759363
llvm-svn: 104174
|
|
|
|
|
|
| |
changes before doing phi lowering for switches.
llvm-svn: 102809
|
|
|
|
| |
llvm-svn: 101655
|
|
|
|
| |
llvm-svn: 99214
|
|
|
|
| |
llvm-svn: 99195
|
|
|
|
|
|
|
| |
No functional effect yet. This is still evolving and should
not be viewed as final.
llvm-svn: 98195
|
|
|
|
|
|
|
| |
and T->isPointerTy(). Convert most instances of the first form to the second form.
Requested by Chris.
llvm-svn: 96344
|
|
|
|
|
|
| |
dependencies were overly conservative for memory access that are known not to alias.
llvm-svn: 86580
|
|
|
|
|
|
| |
instructions.
llvm-svn: 86080
|
|
|
|
|
|
| |
information to avoid chaining loads/stores of spill slots with non-aliased memory ops.
llvm-svn: 85934
|
|
|
|
|
|
| |
(except for true dependency of Store followed by aliased Load... we estimate that case with a single cycle of latency assuming the hardware will bypass)
llvm-svn: 85807
|
|
|
|
|
|
|
|
|
|
| |
machineinstr whether the aliased register is dead, rather than the original
register is dead. This allows it to get the correct answer when examining
an instruction like this:
CALLpcrel32 <ga:foo>, %AL<imp-def>, %EAX<imp-def,dead>
where EAX is dead but a subregister of it is still live. This fixes PR5294.
llvm-svn: 85135
|
|
|
|
| |
llvm-svn: 84432
|
|
|
|
|
|
|
|
|
|
|
| |
stack slots and giving them different PseudoSourceValue's did not fix the
problem of post-alloc scheduling miscompiling llvm itself.
- Apply Dan's conservative workaround by assuming any non fixed stack slots can
alias other memory locations. This means a load from spill slot #1 cannot
move above a store of spill slot #2.
- Enable post-alloc scheduling for x86 at optimization leverl Default and above.
llvm-svn: 84424
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
is trivially rematerializable and integrate it into
TargetInstrInfo::isTriviallyReMaterializable. This way, all places that
need to know whether an instruction is rematerializable will get the
same answer.
This enables the useful parts of the aggressive-remat option by
default -- using AliasAnalysis to determine whether a memory location
is invariant, and removes the questionable parts -- rematting operations
with virtual register inputs that may not be live everywhere.
llvm-svn: 83687
|
|
|
|
|
|
|
|
|
| |
implementations with a new MachineInstr::isInvariantLoad, which uses
MachineMemOperands and is target-independent. This brings MachineLICM
and other functionality to targets which previously lacked an
isInvariantLoad implementation.
llvm-svn: 83475
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Allocate MachineMemOperands and MachineMemOperand lists in MachineFunctions.
This eliminates MachineInstr's std::list member and allows the data to be
created by isel and live for the remainder of codegen, avoiding a lot of
copying and unnecessary translation. This also shrinks MemSDNode.
- Delete MemOperandSDNode. Introduce MachineSDNode which has dedicated
fields for MachineMemOperands.
- Change MemSDNode to have a MachineMemOperand member instead of its own
fields with the same information. This introduces some redundancy, but
it's more consistent with what MachineInstr will eventually want.
- Ignore alignment when searching for redundant loads for CSE, but remember
the greatest alignment.
Target-specific code which previously used MemOperandSDNodes with generic
SDNodes now use MemIntrinsicSDNodes, with opcodes in a designated range
so that the SelectionDAG framework knows that MachineMemOperand information
is available.
llvm-svn: 82794
|
|
|
|
|
|
|
|
| |
sdisel will use to properly complete phi nodes.
Not functionality change yet.
llvm-svn: 82273
|
|
|
|
|
|
| |
dependence edge latency for post-RA scheduling.
llvm-svn: 79425
|
|
|
|
| |
llvm-svn: 78910
|
|
|
|
|
|
| |
target schedule information to accurately model the pipeline. Update the scheduler to correctly handle multi-issue targets.
llvm-svn: 78563
|
|
|
|
| |
llvm-svn: 78362
|
|
|
|
| |
llvm-svn: 76236
|
|
|
|
|
|
|
|
| |
and related functions out of LoopBase and into Loop, since they
are specific to BasicBlock-based loops. This also allows the code
to be moved out-of-line.
llvm-svn: 75523
|
|
|
|
|
|
|
|
|
|
|
| |
instruction index across each part. Instruction indices are used
to make live range queries, and live ranges can extend beyond
scheduling region boundaries.
Refactor the ScheduleDAGSDNodes class some more so that it
doesn't have to worry about this additional information.
llvm-svn: 64288
|
|
|
|
|
|
|
|
|
|
|
| |
scheduling, and generalize is so that preserves state across
scheduling regions. This fixes incorrect live-range information around
terminators and labels, which are effective region boundaries.
In place of looking for terminators to anchor inter-block dependencies,
introduce special entry and exit scheduling units for this purpose.
llvm-svn: 64254
|
|
|
|
|
|
|
|
| |
that used this header to select a scheduling policy should
use SchedulerRegistry.h instead (llvm-gcc and clang were
updated a while ago).
llvm-svn: 63934
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If a MachineInstr doesn't have a memoperand but has an opcode that
is known to load or store, assume its memory reference may alias
*anything*, including stack slots which the compiler completely
controls.
To partially compensate for this, teach the ScheduleDAG building
code to do basic getUnderlyingValue analysis. This greatly
reduces the number of instructions that require restrictive
dependencies. This code will need to be revisited when we start
doing real alias analysis, but it should suffice for now.
llvm-svn: 63370
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
and every other instruction in their blocks to keep the terminator
instructions at the end, teach the post-RA scheduler how to operate
on ranges of instructions, and exclude terminators from the range
of instructions that get scheduled.
Also, exclude mid-block labels, such as EH_LABEL instructions, and
schedule code before them separately from code after them. This
fixes problems with the post-RA scheduler moving code past
EH_LABELs.
llvm-svn: 62366
|
|
|
|
|
|
|
|
|
|
|
| |
and into the ScheduleDAGInstrs class, so that they don't get
destructed and re-constructed for each block. This fixes a
compile-time hot spot in the post-pass scheduler.
To help facilitate this, tidy and do some minor reorganization
in the scheduler constructor functions.
llvm-svn: 62275
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This removes all the _8, _16, _32, and _64 opcodes and replaces each
group with an unsuffixed opcode. The MemoryVT field of the AtomicSDNode
is now used to carry the size information. In tablegen, the size-specific
opcodes are replaced by size-independent opcodes that utilize the
ability to compose them with predicates.
This shrinks the per-opcode tables and makes the code that handles
atomics much more concise.
llvm-svn: 61389
|
|
|
|
|
|
| |
code in ScheduleDAGSDNodes' BuildSchedGraph into separate functions.
llvm-svn: 61376
|
|
|
|
|
|
|
|
| |
several places. isTerminator() returns true for a superset
of cases, and includes things like FP_REG_KILL, which are
nither return or branch but aren't safe to move/remat/etc.
llvm-svn: 61373
|
|
|
|
|
|
|
| |
especially in the case of addresses computed from loop induction
variables.
llvm-svn: 61075
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
computation code. Also, avoid adding output-depenency edges when both
defs are dead, which frequently happens with EFLAGS defs.
Compute Depth and Height lazily, and always in terms of edge latency
values. For the schedulers that don't care about latency, edge latencies
are set to 1.
Eliminate Cycle and CycleBound, and LatencyPriorityQueue's Latencies array.
These are all subsumed by the Depth and Height fields.
llvm-svn: 61073
|
|
|
|
|
|
| |
instruction itinerary data to back-schedule loads.
llvm-svn: 61070
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Cost field is removed. It was only being used in a very limited way,
to indicate when the scheduler should attempt to protect a live register,
and it isn't really needed to do that. If we ever want the scheduler to
start inserting copies in non-prohibitive situations, we'll have to
rethink some things anyway.
A Latency field is added. Instead of giving each node a single
fixed latency, each edge can have its own latency. This will eventually
be used to model various micro-architecture properties more accurately.
The PointerIntPair class and an internal union are now used, which
reduce the overall size.
llvm-svn: 60806
|