6975602024
If a copy isn't coalesced, but its src is defined by trivial computation. Re-materialize the src to replace the copy.
...
llvm-svn: 55467
2008-08-28 07:53:51 +00:00
4f6bf04616
Convert uses of std::vector in TargetInstrInfo to SmallVector. This change had to be propoagated down into all the targets and up into all clients of this API.
...
llvm-svn: 54802
2008-08-14 22:49:33 +00:00
8de6d22392
Use empty() instead of begin() == end().
...
llvm-svn: 54780
2008-08-14 18:13:49 +00:00
51f689a652
Make the allocation of LiveIntervals explicit, rather than holding them in the r2iMap_ by value. This will prevent references to them from being invalidated
...
if the map is changed.
llvm-svn: 54763
2008-08-13 21:49:13 +00:00
a4d6d884d6
Remove #if 0.
...
llvm-svn: 54347
2008-08-05 07:20:57 +00:00
0ca10c9572
Fix PR2568: Fix bug that cause redudant kill marker after its live interval has been extended due to coalescing.
...
llvm-svn: 54346
2008-08-05 07:10:38 +00:00
7c800ad977
Fix a compile-time regression introduced by my heuristic-changing patch. I forgot
...
to multiply the instruction count by a constant factor in a few places, which
caused the register allocator to require many more iterations.
llvm-svn: 53959
2008-07-23 19:47:27 +00:00
029182f3a3
Change the heuristics used in the coalescer, register allocator, and within
...
live intervals itself to use an instruction count approximation that is
not affected by inserting empty indices.
llvm-svn: 53937
2008-07-22 22:46:49 +00:00
cefd6e62fa
Subreg live interval valno may not have a corresponding def machineinstr since it's less precise.
...
llvm-svn: 53734
2008-07-17 19:48:53 +00:00
79d2fa52fa
Use getMBBEndIdx rather than assuming that the end is right after the last instruction in the block.
...
llvm-svn: 52649
2008-06-23 22:12:23 +00:00
f593a65497
Undo spill weight tweak. Need to investigate the performance regressions.
...
llvm-svn: 52572
2008-06-21 06:45:54 +00:00
0c8ef553f5
Coalesce copy from one register class to a sub register class. e.g. X86::MOV16to16_.
...
llvm-svn: 52480
2008-06-19 01:39:21 +00:00
1eb69314fa
When extending a liveinterval by commuting, don't throw away the live ranges that are not affected.
...
llvm-svn: 52430
2008-06-17 20:11:16 +00:00
c5e21e4f38
The coalescer doesn't need LiveVariables now that we have register use iterators.
...
llvm-svn: 51790
2008-05-30 22:37:27 +00:00
68079268f5
Fix PR2289: vr defined by multiple implicit_def as result of coalescing.
...
llvm-svn: 51648
2008-05-28 17:40:10 +00:00
a5d27ae586
Fix PR2343. An *interesting* coalescer bug.
...
BB1:
vr1025 = copy vr1024
..
BB2:
vr1024 = op
= op vr1025
<loop eventually branch back to BB1>
Even though vr1025 is copied from vr1024, it's not safe to coalesced them since live range of vr1025 intersects the def of vr1024. This happens when vr1025 is assigned the value of the previous iteration of vr1024 in the loop.
llvm-svn: 51394
2008-05-21 22:34:12 +00:00
0479aa5c0b
Change class' public PassInfo variables to by initialized with the
...
address of the PassInfo directly instead of calling getPassInfo.
This eliminates a bunch of dynamic initializations of static data.
Also, fold RegisterPassBase into PassInfo, make a bunch of its
data members const, and rearrange some code to initialize data
members in constructors instead of using setter member functions.
llvm-svn: 51022
2008-05-13 02:05:11 +00:00
d78c400b5b
Clean up the use of static and anonymous namespaces. This turned up
...
several things that were neither in an anonymous namespace nor static
but not intended to be global.
llvm-svn: 51017
2008-05-13 00:00:25 +00:00
11b98b6612
Another extract_subreg coalescing bug.
...
e.g.
vr1024<2> extract_subreg vr1025, 2
If vr1024 do not have the same register class as vr1025, it's not safe to coalesce this away. For example, vr1024 might be a GPR32 while vr1025 might be a GPR64.
llvm-svn: 50385
2008-04-29 01:41:44 +00:00
3980a7911a
- Check if a register is livein before removing it. It may have already been removed.
...
- Do not iterate over SmallPtrSet, the order of iteration is not deterministic.
llvm-svn: 50209
2008-04-24 09:06:33 +00:00
d556115e7e
Correct comment.
...
llvm-svn: 49913
2008-04-18 19:25:26 +00:00
495a516390
Not safe to "kill" a register if its live range extends pass the end of block branch.
...
llvm-svn: 49911
2008-04-18 19:22:23 +00:00
7e4a55bc58
Be more careful with insert_subreg and extract_subreg where either source or destination operand has already been coalesced with another register that's defined by a insert_subreg or extract_subreg.
...
llvm-svn: 49843
2008-04-17 07:58:04 +00:00
288ef83b8a
Use correct name for method in comment.
...
llvm-svn: 49841
2008-04-17 05:20:39 +00:00
c8c3a899c0
Fix a sub-register indice propagation bug.
...
llvm-svn: 49832
2008-04-17 00:06:42 +00:00
59aa126e48
After reading memory that's already freed.
...
llvm-svn: 49810
2008-04-16 20:24:25 +00:00
23f12757ed
Fix PR2226. Avoid using uninitialized variables.
...
llvm-svn: 49807
2008-04-16 18:48:43 +00:00
e29e9774a4
Avoid read after free.
...
llvm-svn: 49760
2008-04-16 01:22:28 +00:00
c6864b6652
Remove implicit_def instructions that become dead as result of coalescing.
...
llvm-svn: 49513
2008-04-10 23:48:35 +00:00
16ea87d6ee
A copy instruction may use a register multiple times on some targets. Change them all.
...
llvm-svn: 49491
2008-04-10 18:38:47 +00:00
c8eeb752a3
- More aggressively coalescing away copies whose source is defined by an implicit_def.
...
- Added insert_subreg coalescing support.
llvm-svn: 49448
2008-04-09 20:57:25 +00:00
aa3b55f842
Missed a hasInterval check.
...
llvm-svn: 49415
2008-04-09 01:30:15 +00:00
58936a48ee
- Turn copies of implicit_def into implicit_def instructions.
...
- Be smarter about coalescing copies from implicit_def.
llvm-svn: 49168
2008-04-03 16:41:54 +00:00
db390694ff
One more coalescer fix wrt deadness propagation.
...
llvm-svn: 48837
2008-03-26 20:15:49 +00:00
289ba4f335
Avoid commuting a def MI in order to coalesce a copy instruction away if any use of the same val# is a copy instruction that has already been coalesced.
...
llvm-svn: 48833
2008-03-26 19:03:01 +00:00
7d564c3b4a
lastRegisterUse() should ignore identity copies. Those will be erased.
...
llvm-svn: 48759
2008-03-25 02:02:19 +00:00
69a3f9c417
If the coalescer commuted a def MI to allow coalescing, it can changed a previously coalesced copy into an non-identity copy.
...
llvm-svn: 48752
2008-03-24 23:31:21 +00:00
8c19af1b7e
A couple of kill marker maintainence bug.
...
llvm-svn: 48653
2008-03-21 19:09:30 +00:00
56e9e57d28
Fixed a coalescer bug caused by a typo.
...
llvm-svn: 48526
2008-03-19 02:26:36 +00:00
d096ec0a86
Rewrite code that propagate isDead information after a dead copy is coalesced. This remove some ugly spaghetti code and fixed a number of subtle bugs.
...
llvm-svn: 48490
2008-03-18 08:26:47 +00:00
4f610c0de1
Remove unused options.
...
llvm-svn: 48319
2008-03-13 02:41:34 +00:00
4a3c5eab34
- Fix a subtle bug in RemoveCopyByCommutingDef. ALR is the live range where the source is defined; BLR is the live range which is defined by the copy.
...
If ALR and BLR overlaps and end of BLR extends beyond end of ALR, e.g.
A = or A, B
...
B = A
...
C = A<kill>
...
= B
then do not add kills of A to the newly created B interval.
- Also fix some kill info update bug.
llvm-svn: 48141
2008-03-10 08:11:32 +00:00
29b502e0e0
Fix a coalescer bug wrt how dead copy interval is shortened.
...
llvm-svn: 47966
2008-03-05 22:09:42 +00:00
6325446666
Refactor code. Remove duplicated functions that basically do the same thing as
...
findRegisterUseOperandIdx, findRegisterDefOperandIndx. Fix some naming inconsistencies.
llvm-svn: 47927
2008-03-05 00:59:57 +00:00
ca7c61e79a
No need for coalescer to update kills. Only copies are coalesced and those instructions will be deleted. Doh.
...
llvm-svn: 47749
2008-02-29 02:50:03 +00:00
d7a258d325
Rename PrintableName to Name.
...
llvm-svn: 47629
2008-02-26 21:47:57 +00:00
c24ea4fb41
Change "Name" to "AsmName" in the target register info. Gee, a refactoring tool
...
would have been a Godsend here!
llvm-svn: 47625
2008-02-26 21:11:01 +00:00
fa6b366892
Enable -coalescer-commute-instrs by default.
...
llvm-svn: 47623
2008-02-26 20:40:22 +00:00
2ff0b0e681
This is possible:
...
vr1 = extract_subreg vr2, 3
...
vr3 = extract_subreg vr1, 2
The end result is vr3 is equal to vr2 with subidx 2.
llvm-svn: 47592
2008-02-26 08:03:41 +00:00
271aef2b03
Fix compiler warning.
...
llvm-svn: 47468
2008-02-22 01:48:00 +00:00
31160f5b98
Help testing.
...
llvm-svn: 47448
2008-02-21 19:20:21 +00:00
b2e4b7adde
- Remove the previous check which broke coalescer-commute3.ll
...
- For now, conservatively ignore copy MI whose source is a physical register. Commuting its def MI can cause a physical register live interval to be live through a loop (since we know it's live coming into the def MI).
llvm-svn: 47281
2008-02-18 18:56:31 +00:00
8f90724a53
For now, avoid commuting def MI for copy MI's whose source is not killed. That simply trade a live interval for another and because only the non-two-address operands can be folded into loads, may end up pessimising code.
...
llvm-svn: 47262
2008-02-18 08:40:53 +00:00
652e4618e2
Refactor some code; check if commuteInstruction is able to commute the instruction.
...
llvm-svn: 47208
2008-02-16 02:32:17 +00:00
803bb6d699
The copy instruction being coalesced will be removed, it is not a kill.
...
llvm-svn: 47179
2008-02-15 21:36:51 +00:00
2ff2da89ab
- Removing the infamous r2rMap_ and rep() method. Now the coalescer will update
...
register defs and uses after each successful coalescing.
- Also removed a number of hacks and fixed some subtle kill information bugs.
llvm-svn: 47167
2008-02-15 18:24:29 +00:00
587c66ed96
Some code clean up.
...
llvm-svn: 47060
2008-02-13 09:56:03 +00:00
8cc58728a8
* Cannot safely commute an instruction there are other defs which can reach its uses.
...
* Ignore copy instructions which have already been coalesced.
llvm-svn: 47056
2008-02-13 08:41:08 +00:00
1446726f3e
Initial support for copy elimination by commuting its definition MI.
...
PR1877.
A3 = op A2 B0<kill>
...
B1 = A3 <- this copy
...
= op A3 <- more uses
==>
B2 = op B0 A2<kill>
...
B1 = B2 <- now an identify copy
...
= op B2 <- more uses
This speeds up FreeBench/neural by 29%, Olden/bh by 12%, oopack_v1p8 by 53%.
llvm-svn: 47046
2008-02-13 03:01:43 +00:00
3a4be0fdef
Rename MRegisterInfo to TargetRegisterInfo.
...
llvm-svn: 46930
2008-02-10 18:45:23 +00:00
6aabf837fe
Remove unused hidden option.
...
llvm-svn: 46903
2008-02-09 08:36:28 +00:00
0c209430b4
Don't recalculate the loop info and loop dominators analyses if they're
...
preserved.
llvm-svn: 45596
2008-01-04 20:54:55 +00:00
118ae4cd61
80-column violations.
...
llvm-svn: 45574
2008-01-04 08:59:18 +00:00
21ec2b4769
update a couple of references to SSARegMap.
...
llvm-svn: 45468
2007-12-31 04:16:08 +00:00
a10fff51d9
Rename SSARegMap -> MachineRegisterInfo in keeping with the idea
...
that "machine" classes are used to represent the current state of
the code being compiled. Given this expanded name, we can start
moving other stuff into it. For now, move the UsedPhysRegs and
LiveIn/LoveOuts vectors from MachineFunction into it.
Update all the clients to match.
This also reduces some needless #includes, such as MachineModuleInfo
from MachineFunction.
llvm-svn: 45467
2007-12-31 04:13:23 +00:00
6005589faf
More cleanups for MachineOperand:
...
- Eliminate the static "print" method for operands, moving it
into MachineOperand::print.
- Change various set* methods for register flags to take a bool
for the value to set it to. Remove unset* methods.
- Group methods more logically by operand flavor in MachineOperand.h
llvm-svn: 45461
2007-12-30 21:56:09 +00:00
f3ebc3f3d2
Remove attribution from file headers, per discussion on llvmdev.
...
llvm-svn: 45418
2007-12-29 20:36:04 +00:00
a509537e25
The physical register + virtual register joining requirement was much too strict.
...
llvm-svn: 45253
2007-12-20 02:23:25 +00:00
303417d242
Switch over to MachineLoopInfo.
...
llvm-svn: 44838
2007-12-11 02:09:15 +00:00
85cdba29b0
Add an option to control this heuristic tweak so I can test it.
...
llvm-svn: 44671
2007-12-07 00:28:32 +00:00
7fc1d98353
Fix for PR1831: if all defs of an interval are re-materializable, then it's a preferred spill candiate.
...
llvm-svn: 44644
2007-12-06 00:01:56 +00:00
f85c063ec0
Replace the odd kill# hack with something less fragile.
...
llvm-svn: 44434
2007-11-29 09:49:23 +00:00
8e22379303
Live interval splitting:
...
When a live interval is being spilled, rather than creating short, non-spillable
intervals for every def / use, split the interval at BB boundaries. That is, for
every BB where the live interval is defined or used, create a new interval that
covers all the defs and uses in the BB.
This is designed to eliminate one common problem: multiple reloads of the same
value in a single basic block. Note, it does *not* decrease the number of spills
since no copies are inserted so the split intervals are *connected* through
spill and reloads (or rematerialization). The newly created intervals can be
spilled again, in that case, since it does not span multiple basic blocks, it's
spilled in the usual manner. However, it can reuse the same stack slot as the
previously split interval.
This is currently controlled by -split-intervals-at-bb.
llvm-svn: 44198
2007-11-17 00:40:40 +00:00
7f02cfa599
Clean up sub-register implementation by moving subReg information back to
...
MachineOperand auxInfo. Previous clunky implementation uses an external map
to track sub-register uses. That works because register allocator uses
a new virtual register for each spilled use. With interval splitting (coming
soon), we may have multiple uses of the same register some of which are
of using different sub-registers from others. It's too fragile to constantly
update the information.
llvm-svn: 44104
2007-11-14 07:59:08 +00:00
be51f28e2b
Refactor some code.
...
llvm-svn: 44010
2007-11-12 06:35:08 +00:00
d5d59ad634
First step towards moving the coalescer to priority_queue based machinery.
...
llvm-svn: 43764
2007-11-06 08:52:21 +00:00
8bb30184a8
Move SimpleRegisterCoalescing.h to lib/CodeGen since there is now a common
...
register coalescer interface: RegisterCoalescing.
llvm-svn: 43714
2007-11-05 17:41:38 +00:00
17b0e3e1ae
Skip over deleted val#'s.
...
llvm-svn: 43700
2007-11-05 06:46:45 +00:00
fe1ac52836
- Coalesce extract_subreg when both intervals are relatively small.
...
- Some code clean up.
llvm-svn: 43606
2007-11-01 06:22:48 +00:00
e6a41c066a
Really fix PR1734. Carefully track which register uses are sub-register uses by
...
traversing inverse register coalescing map.
llvm-svn: 43118
2007-10-18 07:49:59 +00:00
c8b5397000
One more extract_subreg coalescing bug fix.
...
llvm-svn: 43065
2007-10-17 05:29:37 +00:00
fab7ca89d5
Fix PR1734.
...
llvm-svn: 43035
2007-10-16 19:29:47 +00:00
ecf62cb763
Code clean up.
...
llvm-svn: 43026
2007-10-16 08:04:24 +00:00
a5abba65b6
Fix PR1729: watch out for val# with no def.
...
llvm-svn: 42996
2007-10-15 18:33:50 +00:00
8d6da9142c
When coalescing an EXTRACT_SUBREG and the dst register is a physical register,
...
the source register will be coalesced to the super register of the LHS. Properly
merge in the live ranges of the resulting coalesced interval that were part of
the original source interval to the live interval of the super-register.
llvm-svn: 42961
2007-10-14 10:08:34 +00:00
11330f7526
Restrict EXTRACT_SUBREG coalescing to avoid negative performance impact.
...
llvm-svn: 42903
2007-10-12 09:15:53 +00:00
aa2d6ef81d
EXTRACT_SUBREG coalescing support. The coalescer now treats EXTRACT_SUBREG like
...
(almost) a register copy. However, it always coalesced to the register of the
RHS (the super-register). All uses of the result of a EXTRACT_SUBREG are sub-
register uses which adds subtle complications to load folding, spiller rewrite,
etc.
llvm-svn: 42899
2007-10-12 08:50:34 +00:00
a9830a04eb
Bad choice of variable name.
...
llvm-svn: 42821
2007-10-10 00:11:40 +00:00
ad55a6079a
Fix an extremely stupid bug that prevented first round of coalescing (physical registers only) from happening.
...
llvm-svn: 42820
2007-10-09 23:36:27 +00:00
9da02f5ee2
Remove isReg, isImm, and isMBB, and change all their users to use
...
isRegister, isImmediate, and isMachineBasicBlock, which are equivalent,
and more popular.
llvm-svn: 41958
2007-09-14 20:33:02 +00:00
65d2a697dc
Pluggable coalescers inplementation.
...
llvm-svn: 41743
2007-09-06 16:18:45 +00:00
db53aef53e
Use pool allocator for all the VNInfo's to improve memory access locality. This reduces coalescing time on siod Mac OS X PPC by 35%. Also remove the back ptr from VNInfo to LiveInterval and other tweaks.
...
llvm-svn: 41729
2007-09-05 21:46:51 +00:00
2089a21360
More tweaks to improve compile time.
...
llvm-svn: 41669
2007-09-01 02:03:17 +00:00
7bef79a1c3
std::map -> DenseMap for slight compile time benefit.
...
llvm-svn: 41650
2007-08-31 21:23:06 +00:00
262596e76a
Use std::map instead of a (potentially very sparse) array to track val# defined by copy from the other live range. Minor compile time win when number of val# is large.
...
llvm-svn: 41640
2007-08-31 08:04:17 +00:00
1ad4a6117b
Change LiveRange so it keeps a pointer to the VNInfo rather than an index.
...
Changes related modules so VNInfo's are not copied. This decrease
copy coalescing time by 45% and overall compilation time by 10% on siod.
llvm-svn: 41579
2007-08-29 20:45:00 +00:00
a5b10b334f
Recover most of the compile time regression due to recent live interval changes.
...
1. Eliminate the costly live interval "swapping".
2. Change ValueNumberInfo container from SmallVector to std::vector. The former
performs slowly when the vector size is very large.
llvm-svn: 41536
2007-08-28 08:28:51 +00:00
70c2de7bf1
Fix some kill info update bugs; add hidden option -disable-rematerialization to turn off remat for debugging.
...
llvm-svn: 41118
2007-08-16 07:24:22 +00:00
94168a4ed0
Fix for PR1596: AdjustCopiesBackFrom() should conservatively check if any of its sub-registers may overlap with the interval of the copy that's being coalesced.
...
llvm-svn: 41084
2007-08-14 23:19:28 +00:00
74c69f7588
Kill info update bugs.
...
llvm-svn: 41064
2007-08-14 01:56:58 +00:00
Evan Cheng