code-size win, and not when it's only likely to be code-size neutral,
such as when only a single instruction would be eliminated and a new
branch would be required.
This fixes rdar://7392894.
llvm-svn: 88692
instead of typedefs for std::pair. This simplifies the type of
SameTails, which previously was std::vector<std::pair<std::vector<std::pair<unsigned, MachineBasicBlock *> >::iterator, MachineBasicBlock::iterator>
llvm-svn: 86885
tail merging support to handle more cases.
- Recognize several cases where tail merging is beneficial even when
the tail size is smaller than the generic threshold.
- Make use of MachineInstrDesc::isBarrier to help detect
non-fallthrough blocks.
- Check for and avoid disrupting fall-through edges in more cases.
llvm-svn: 86871
an unconditional branch (possibly from tail merging), this code is
trying to redirect all of its predecessors to go directly to the branch
target, but that isn't feasible for indirect branches. The other
predecessors (that don't end with indirect branches) could theoretically
still be handled, but that is not easily done right now.
The AnalyzeBranch interface doesn't currently let us distinguish jump table
branches from indirect branches, and this code is currently handling
jump tables. To avoid punting on address-taken blocks, we would have to give
up handling jump tables. That seems like a bad tradeoff.
llvm-svn: 85975
use it to control tail merging when there is a tradeoff between performance
and code size. When there is only 1 instruction in the common tail, we have
been merging. That can be good for code size but is a definite loss for
performance. Now we will avoid tail merging in that case when the
optimization level is "Aggressive", i.e., "-O3". Radar 7338114.
Since the IfConversion pass invokes BranchFolding, it too needs to know
the optimization level. Note that I removed the RegisterPass instantiation
for IfConversion because it required a default constructor. If someone
wants to keep that for some reason, we can add a default constructor with
a hard-wired optimization level.
llvm-svn: 85346
to break up CFG diamonds by banishing one of the blocks to the end of
the function, which is bad for code density and branch size.
This does pessimize MultiSource/Benchmarks/Ptrdist/yacr2, the
benchmark cited as the reason for the change, however I've examined
the code and it looks more like a case of gaming a particular
branch than of being generally applicable.
llvm-svn: 84803
remove RemoveDuplicateSuccessor, as it is no longer necessary, and because
it breaks assumptions made in
MachineBasicBlock::isOnlyReachableByFallthrough.
Convert test/CodeGen/X86/omit-label.ll to FileCheck and add a testcase
for PR4732.
test/CodeGen/Thumb2/thumb2-ifcvt2.ll sees a diff with this commit due to
it being bugpoint-reduced to the point where it doesn't matter what the
condition for the branch is.
Add some more interesting code to
test/CodeGen/X86/2009-08-06-branchfolder-crash.ll, which is the testcase
that originally motivated the RemoveDuplicateSuccessor code, to help
verify that the original problem isn't being re-broken.
llvm-svn: 79338
This adds location info for all llvm_unreachable calls (which is a macro now) in
!NDEBUG builds.
In NDEBUG builds location info and the message is off (it only prints
"UREACHABLE executed").
llvm-svn: 75640
Make llvm_unreachable take an optional string, thus moving the cerr<< out of
line.
LLVM_UNREACHABLE is now a simple wrapper that makes the message go away for
NDEBUG builds.
llvm-svn: 75379
suprise to some callers, e.g. register coalescer. For now, add an parameter
that tells AnalyzeBranch whether it's safe to modify the mbb. A better
solution is out there, but I don't have time to deal with it right now.
llvm-svn: 64124
is set but mayLoad is not set. Fix all the problems this turned up.
Change code to not use isSimpleLoad instead of mayLoad unless it
really wants isSimpleLoad.
llvm-svn: 60459
isImmediate(), isRegister(), and friends, to avoid confusion
about having two different names with the same meaning. I'm
not attached to the longer names, and would be ok with
changing to the shorter names if others prefer it.
llvm-svn: 56189
MachineMemOperands. The pools are owned by MachineFunctions.
This drastically reduces the number of calls to malloc/free made
during the "Emit" phase of scheduling, as well as later phases
in CodeGen. Combined with other changes, this speeds up the
"instruction selection" phase of CodeGen by 10% in some cases.
llvm-svn: 53212
the need for a flavor operand, and add a new SDNode subclass,
LabelSDNode, for use with them to eliminate the need for a label id
operand.
Change instruction selection to let these label nodes through
unmodified instead of creating copies of them. Teach the MachineInstr
emitter how to emit a MachineInstr directly from an ISD label node.
This avoids the need for allocating SDNodes for the label id and
flavor value, as well as SDNodes for each of the post-isel label,
label id, and label flavor.
llvm-svn: 52943
that it is cheap and efficient to get.
Move a variety of predicates from TargetInstrInfo into
TargetInstrDescriptor, which makes it much easier to query a predicate
when you don't have TII around. Now you can use MI->getDesc()->isBranch()
instead of going through TII, and this is much more efficient anyway. Not
all of the predicates have been moved over yet.
Update old code that used MI->getInstrDescriptor()->Flags to use the
new predicates in many places.
llvm-svn: 45674
e.g. MO.isMBB() instead of MO.isMachineBasicBlock(). I don't plan on
switching everything over, so new clients should just start using the
shorter names.
Remove old long accessors, switching everything over to use the short
accessor: getMachineBasicBlock() -> getMBB(),
getConstantPoolIndex() -> getIndex(), setMachineBasicBlock -> setMBB(), etc.
llvm-svn: 45464
See test/CodeGen/X86/test-pic-jtbl.ll for a case where it works well;
shaves another 10K off our favorite benchmark. I was hesitant about
this because of compile speed, but seems to do OK on a bootstrap.
llvm-svn: 37392
When considering blocks with more than 2 predecessors, merge the block with
the largest number of matching insns, rather than the first block found.
Considering that 1 matching insn is enough to show a win for candidates that
already end with a branch.
llvm-svn: 37315
if (cond) goto BB2
BB1:
...
return;
BB2:
...
Move BB1 to the end of the function so that the code falls through in the
non-return case. This has the effect of moving assert (and other no-return
call) bodies and return blocks out of loops.
llvm-svn: 31855
LBB1_57: #bb207.i
movl 72(%esp), %ecx
movb (%ecx,%eax), %cl
movl 80(%esp), %edx
movb %cl, 1(%edx,%eax)
incl %eax
cmpl $143, %eax
jne LBB1_57 #bb207.i
jmp LBB1_64 #cond_next255.i
intead of:
LBB1_57: #bb207.i
movl 72(%esp), %ecx
movb (%ecx,%eax), %cl
movl 80(%esp), %edx
movb %cl, 1(%edx,%eax)
incl %eax
cmpl $143, %eax
je LBB1_64 #cond_next255.i
jmp LBB1_57 #bb207.i
This eliminates a branch per iteration of the loop. This hurted PPC
particularly, because the extra branch meant another dispatch group for each
iteration of the loop.
llvm-svn: 31530
Turn on -Wunused and -Wno-unused-parameter. Clean up most of the resulting
fall out by removing unused variables. Remaining warnings have to do with
unused functions (I didn't want to delete code without review) and unused
variables in generated code. Maintainers should clean up the remaining
issues when they see them. All changes pass DejaGnu tests and Olden.
llvm-svn: 31380
a place more useful. In particular, if we can put them in a place where code
will be able to fall into it, do so. Otherwise, put it in a place it can fall
through into a successor. Otherwise, if preventing a fallthrough, move to the
end of the function, out of the way.
This deletes several hundred unconditional branches from spass.
llvm-svn: 31149
1. Remove a bunch of ifdef'd code.
2. When a block just contains an uncond branch, change all blocks branching
to it to jump to the destination instead.
3. If branch analysis tells us some edges in the machinecfg are not actually
possible, remove them.
#2 triggers a suprisingly large number of times.
llvm-svn: 31094
branch folding can now compile stuff like this:
void foo(int W, int X, int Y, int Z) {
if (W & 1) {
for (; X;--X) bar();
} else if (W & 2) {
for (; Y;--Y) bar();
} else if (W & 4) {
for (; Z;--Z) bar();
} else if (W & 8) {
for (; W;--W) bar();
}
if (W) {
bar();
}
}
contrived testcase where loops exits all end up merging together. To have
the loop merges be:
...
cmplw cr0, r30, r27
bne cr0, LBB1_14 ;bb38
LBB1_16: ;cond_next48.loopexit
mr r27, r29
LBB1_20: ;cond_next48
cmplwi cr0, r27, 0
beq cr0, LBB1_22 ;UnifiedReturnBlock
...
instead of:
...
cmplw cr0, r30, r27
bne cr0, LBB1_14 ;bb38
LBB1_16: ;cond_next48.loopexit
mr r27, r29
b LBB1_20 ;cond_next48
LBB1_17: ;cond_next48.loopexit1
b LBB1_20 ;cond_next48
LBB1_18: ;cond_next48.loopexit2
b LBB1_20 ;cond_next48
LBB1_19: ;cond_next48.loopexit3
LBB1_20: ;cond_next48
cmplwi cr0, r27, 0
beq cr0, LBB1_22 ;UnifiedReturnBlock
...
This is CodeGen/PowerPC/branch-opt.ll
llvm-svn: 31006
Move include/Config and include/Support into include/llvm/Config,
include/llvm/ADT and include/llvm/Support. From here on out, all LLVM
public header files must be under include/llvm/.
llvm-svn: 16137
Bob Wilson