Commit Graph

1134 Commits

Author SHA1 Message Date
Alkis Evlogimenos
f978a1db51 Make OptimizeBlock take a MachineFunction::iterator instead of a
MachineBasicBlock* as a parameter so that nxext() and prior() helper
functions can work naturally on it.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15376 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-31 19:24:41 +00:00
Chris Lattner
7f74340733 Next on a pointer increments the pointer, not an iterator
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15375 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-31 18:40:36 +00:00
Alkis Evlogimenos
9fd332392c Use next() helper to make code more readable. Use
MachineFunction::iterator instead of MachineBasicBlock* to avoid
dereferencing end iterators.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15373 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-31 15:14:29 +00:00
Alkis Evlogimenos
dd04583781 Use MachineFunction::iterator instead of a MachineBasicBlock* because
FallThrough maybe == to MF.end().


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15372 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-31 15:03:52 +00:00
Chris Lattner
21ab22e475 Implement a simple target-independent CFG cleanup pass
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15368 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-31 10:01:27 +00:00
Tanya Lattner
0a88d2d8b9 Updated ModuloScheduling. It makes it all the wya through register allocation on the new code!!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15351 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-30 23:36:10 +00:00
Brian Gaeke
365f54c7bc Convert a few assertions with side-effects into regular old runtime checks.
These side-effects seem to make a difference when using llc -march=sparcv9
in Release mode (i.e., with -DNDEBUG); when they are left out, lots of
instructions just get dropped on the floor, because they never end up
in the schedule.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15339 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-29 21:31:20 +00:00
Misha Brukman
47b14a4a6a Fix #includes of i*.h => Instructions.h as per PR403.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15334 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-29 17:30:56 +00:00
Chris Lattner
879313ae3a Fix #includes of i*.h => Instructions.h as per PR403:
http://llvm.cs.uiuc.edu/PR403 .


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15333 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-29 17:23:00 +00:00
Chris Lattner
9670eecf33 Fix #includes of i*.h => Instructions.h as per PR403:
http://llvm.cs.uiuc.edu/PR403 .


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15331 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-29 17:11:37 +00:00
Brian Gaeke
b2f30a3792 TargetInstrInfo::hasOperandInterlock() is always true, because it is
never overridden by any target.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15308 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-28 19:24:48 +00:00
Alkis Evlogimenos
5a0a1c2cba Add some comments to the backtracking code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15200 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-25 08:10:33 +00:00
Chris Lattner
c25b55a5b2 Fix the sense of joinable
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15196 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-25 07:47:25 +00:00
Chris Lattner
f542649f1b This patch makes use of the infrastructure implemented before to safely and
aggressively coallesce live ranges even if they overlap.  Consider this LLVM
code for example:

int %test(int %X) {
        %Y = mul int %X, 1      ;; Codegens to Y = X
        %Z = add int %X, %Y
        ret int %Z
}

The mul is just there to get a copy into the code stream.  This produces
this machine code:

 (0x869e5a8, LLVM BB @0x869b9a0):
        %reg1024 = mov <fi#-2>, 1, %NOREG, 0    ;; "X"
        %reg1025 = mov %reg1024                 ;; "Y"  (subsumed by X)
        %reg1026 = add %reg1024, %reg1025
        %EAX = mov %reg1026
        ret

Note that the life times of reg1024 and reg1025 overlap, even though they
contain the same value.  This results in this machine code:

test:
        mov %EAX, DWORD PTR [%ESP + 4]
        mov %ECX, %EAX
        add %EAX, %ECX
        ret

Another, worse case involves loops and PHI nodes.  Consider this trivial loop:
testcase:

int %test2(int %X) {
entry:
        br label %Loop
Loop:
        %Y = phi int [%X, %entry], [%Z, %Loop]
        %Z = add int %Y, 1
        %cond = seteq int %Z, 100
        br bool %cond, label %Out, label %Loop
Out:
        ret int %Z
}

Because of interactions between the PHI elimination pass and the register
allocator, this got compiled to this code:

test2:
        mov %ECX, DWORD PTR [%ESP + 4]
.LBBtest2_1:
***     mov %EAX, %ECX
        inc %EAX
        cmp %EAX, 100
***     mov %ECX, %EAX
        jne .LBBtest2_1

        ret

Or on powerpc, this code:

_test2:
        mflr r0
        stw r0, 8(r1)
        stwu r1, -60(r1)
.LBB_test2_1:
        addi r2, r3, 1
        cmpwi cr0, r2, 100
***     or r3, r2, r2
        bne cr0, .LBB_test2_1

***     or r3, r2, r2
        lwz r0, 68(r1)
        mtlr r0
        addi r1, r1, 60
        blr 0



With this improvement in place, we now generate this code for these two
testcases, which is what we want:


test:
        mov %EAX, DWORD PTR [%ESP + 4]
        add %EAX, %EAX
        ret

test2:
        mov %EAX, DWORD PTR [%ESP + 4]
.LBBtest2_1:
        inc %EAX
        cmp %EAX, 100
        jne .LBBtest2_1 # Loop
        ret

Or on PPC:

_test2:
        mflr r0
        stw r0, 8(r1)
        stwu r1, -60(r1)
.LBB_test2_1:
        addi r3, r3, 1
        cmpwi cr0, r3, 100
        bne cr0, .LBB_test2_1

        lwz r0, 68(r1)
        mtlr r0
        addi r1, r1, 60
        blr 0


Static numbers for spill code loads/stores/reg-reg copies (smaller is better):

em3d:       before: 47/25/26         after: 44/22/24
164.gzip:   before: 433/245/310      after: 403/231/278
175.vpr:    before: 3721/2189/1581   after: 4144/2081/1423
176.gcc:    before: 26195/8866/9235  after: 25942/8082/8275
186.crafty: before: 4295/2587/3079   after: 4119/2519/2916
252.eon:    before: 12754/7585/5803  after: 12508/7425/5643
256.bzip2:  before: 463/226/315      after: 482:241/309


Runtime perf number samples on X86:

gzip: before: 41.09 after: 39.86
bzip2: runtime: before: 56.71s after: 57.07s
gcc: before: 6.16 after: 6.12
eon: before: 2.03s after: 2.00s


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15194 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-25 07:11:19 +00:00
Chris Lattner
d3a205eab5 Make a method const, no functionality changes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15193 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-25 06:23:01 +00:00
Chris Lattner
62d4e16a05 Fix a bug where we incorrectly value numbered the first PHI definition the
same as the PHI use.  This is not correct as the PHI use value is different
depending on which branch is taken.  This fixes espresso with aggressive
coallescing, and perhaps others.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15189 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-25 05:45:18 +00:00
Chris Lattner
6925a9f9cc Fix a bug in the range remover
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15188 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-25 05:43:53 +00:00
Chris Lattner
c83e40d1b6 Add debugging output for joining assignments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15187 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-25 03:24:11 +00:00
Alkis Evlogimenos
3f86193cd2 Remove implementation of operator= and make it private so that it is
not used accidentally.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15172 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-24 18:55:15 +00:00
Alkis Evlogimenos
a1613db62f Change std::map<unsigned, LiveInterval*> into a std::map<unsigned,
LiveInterval>. This saves some space and removes the pointer
indirection caused by following the pointer.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15167 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-24 11:44:15 +00:00
Chris Lattner
060913cce4 whoops, didn't mean to remove this
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15157 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-24 04:32:22 +00:00
Chris Lattner
deb9971061 In the joiner, merge the small interval into the large interval. This restores
us back to taking about 10.5s on gcc, instead of taking 15.6s!  The net result
is that my big patches have hand no significant effect on compile time or code
quality.  heh.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15156 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-24 03:41:50 +00:00
Chris Lattner
4df98e546d Completely eliminate the intervals_ list. instead, the r2iMap_ maintains
ownership of the intervals.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15155 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-24 03:32:06 +00:00
Chris Lattner
7ac2d3146a Big change to compute logical value numbers for each LiveRange added to an
Interval.  This generalizes the isDefinedOnce mechanism that we used before
to help us coallesce ranges that overlap.  As part of this, every logical
range with a different value is assigned a different number in the interval.
For example, for code that looks like this:

0  X = ...
4  X += ...
  ...
N    = X

We now generate a live interval that contains two ranges: [2,6:0),[6,?:1)
reflecting the fact that there are two different values in the range at
different positions in the code.

Currently we are not using this information at all, so this just slows down
liveintervals.  In the future, this will change.

Note that this change also substantially refactors the joinIntervalsInMachineBB
method to merge the cases for virt-virt and phys-virt joining into a single
case, adds comments, and makes the code a bit easier to follow.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15154 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-24 02:59:07 +00:00
Chris Lattner
0f4c076140 Add a new differingRegisterClasses method
make overlapsAliases take pointers instead of references
fix indentation


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15153 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-24 02:53:43 +00:00
Chris Lattner
abf295fc6c Little stuff:
* Fix comment typeo
* add dump() methods
* add a few new methods like getLiveRangeContaining, removeRange & joinable
  (which is currently the same as overlaps)
* Remove the unused operator==

Bigger change:

* In LiveInterval, instead of using a boolean isDefinedOnce to keep track of
  if there are > 1 definitions in a particular interval, keep a counter,
  NumValues to keep track of exactly how many there are.
* In LiveRange, add a new ValId element to indicate which of the numbered
  values each LiveRange belongs to.   We now no longer merge LiveRanges if
  they are of differing value ID's even if they are neighbors.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15152 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-24 02:52:23 +00:00
Chris Lattner
f35fef7060 More minor changes:
* Inline some functions
 * Eliminate some comparisons from the release build

This is good for another .3 on gcc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15144 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 21:24:19 +00:00
Chris Lattner
b26c215c05 Change addRange and join to be a little bit smarter. In particular, we don't
want to insert a new range into the middle of the vector, then delete ranges
one at a time next to the inserted one as they are merged.

Instead, if the inserted interval overlaps, just start merging.  The only time
we insert into the middle of the vector is when we don't overlap at all.  Also
delete blocks of live ranges if we overlap with many of them.

This patch speeds up joining by .7 seconds on a large testcase, but more
importantly gets all of the range adding code into addRangeFrom.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15141 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 19:38:44 +00:00
Chris Lattner
aa14147cd6 Search by the start point, not by the whole interval. This saves some
comparisons, reducing linscan by another .1 seconds :)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15139 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 18:40:00 +00:00
Chris Lattner
e6ad392802 New helper method
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15138 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 18:39:12 +00:00
Chris Lattner
4dc54ae0d9 Speedup debug builds a bit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15137 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 18:38:52 +00:00
Chris Lattner
ebd7e6c54d Instead of searching for a live interval pair, search for a location. This gives
a very modest speedup of .3 seconds compiling 176.gcc (out of 20s).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15136 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 18:13:24 +00:00
Chris Lattner
a3b8b5c0e0 Rename LiveIntervals.(cpp|h) -> LiveIntervalAnalysis.(cpp|h)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15135 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 17:56:30 +00:00
Chris Lattner
fb449b9ea5 Pull the LiveRange and LiveInterval classes out of LiveIntervals.h (which
will soon be renamed) into their own file.  The new file should not emit
DEBUG output or have other side effects.  The LiveInterval class also now
doesn't know whether its working on registers or some other thing.

In the future we will want to use the LiveInterval class and friends to do
stack packing.  In addition to a code simplification, this will allow us to
do it more easily.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15134 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 17:49:16 +00:00
Chris Lattner
ec2bc64505 Improve comments a bit
Use an explicit LiveRange class to represent ranges instead of an std::pair.
This is a minor cleanup, but is really intended to make a future patch simpler
and less invasive.

Alkis, could you please take a look at LiveInterval::liveAt?  I suspect that
you can add an operator<(unsigned) to LiveRange, allowing us to speed up the
upper_bound call by quite a bit (this would also apply to other callers of
upper/lower_bound).  I would do it myself, but I still don't understand that
crazy liveAt function, despite the comment. :)

Basically I would like to see this:
    LiveRange dummy(index, index+1);
    Ranges::const_iterator r = std::upper_bound(ranges.begin(),
                                                ranges.end(),
                                                dummy);

Turn into:
    Ranges::const_iterator r = std::upper_bound(ranges.begin(),
                                                ranges.end(),
                                                index);


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15130 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 08:24:23 +00:00
Chris Lattner
94881e8c45 Update live intervals more accurately for PHI elim. This slightly reduces
the live intervals for some registers.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15125 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 05:27:43 +00:00
Chris Lattner
fe1630b43e Force coallescing of live ranges that have a single definition, even if they
interfere.  Because these intervals have a single definition, and one of them
is a copy instruction, they are always safe to merge even if their lifetimes
interfere.  This slightly reduces the amount of spill code, for example on
252.eon, from:

 12837 spiller               - Number of loads added
  7604 spiller               - Number of stores added
  5842 spiller               - Number of register spills
 18155 liveintervals         - Number of identity moves eliminated after coalescing

to:

  12754 spiller               - Number of loads added
   7585 spiller               - Number of stores added
   5803 spiller               - Number of register spills
  18262 liveintervals         - Number of identity moves eliminated after coalescing

The much much bigger win would be to merge intervals with multiple definitions
(aka phi nodes) but this is not that day.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15124 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-23 05:26:05 +00:00
Chris Lattner
a7bfbba856 costmetic changes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15118 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 23:05:12 +00:00
Chris Lattner
e8850f476e Fix broken -debug printing
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15115 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 21:54:22 +00:00
Chris Lattner
4951d48a16 The default has not been 'simple' for AGES!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15114 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 21:46:02 +00:00
Chris Lattner
0bafa9818b Make linear scan the default
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15111 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 18:42:00 +00:00
Alkis Evlogimenos
77c4c4da24 Put variable name to a separate line.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15108 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 15:30:33 +00:00
Misha Brukman
75fa4e4f99 Fix indentation and wrap code at 80 cols
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15107 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 15:26:23 +00:00
Alkis Evlogimenos
d3014edc51 Sorting is now handled by both linearscan and iterative scan so live
intervals need not be sorted anymore. Removing this redundant step
improves LiveIntervals running time by 5% on 176.gcc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15106 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 15:18:10 +00:00
Alkis Evlogimenos
82c24fba8a Fit to 80 columns.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15105 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 14:29:31 +00:00
Alkis Evlogimenos
53eb37340c Some compile time improvements resulting in a 1sec speedup in the 5sec
compilation of gcc:

* Use vectors instead of lists for the intervals sets
* Use a heap for the unhandled set to keep intervals always sorted and
  makes insertions back to the heap very fast (compared to scanning a
  list)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15103 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 08:14:44 +00:00
Chris Lattner
199edde707 Remove extraneous punctuation
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15098 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 05:51:56 +00:00
Alkis Evlogimenos
10e169b038 Use reverse iterators when updating the vector, since scanning from
the end will reduce erase() runtimes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15093 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 02:16:53 +00:00
Chris Lattner
6beef3e1a0 That funny 2-address lowering pass can also cause multiple definitions,
fortunately, they are easy to handle if we know about them.  This patch fixes
some serious pessimization of code produced by the linscan register allocator.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15092 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-22 00:04:14 +00:00
Chris Lattner
1e31363a9e Minor cleanups
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@15091 91177308-0d34-0410-b5e6-96231b3b80d8
2004-07-21 23:17:57 +00:00