Commit Graph

1245 Commits

Author SHA1 Message Date
Chris Lattner
b2f12a2c38 Initial support for edge profiling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12225 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-08 17:54:34 +00:00
Chris Lattner
467dd2ec61 Split utility functions out of BlockProfiling.cpp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12224 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-08 17:06:13 +00:00
Chris Lattner
c51733c7b5 finegrainify namespacification
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12221 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-08 16:45:53 +00:00
Chris Lattner
9440db8866 Implement ArgumentPromotion/aggregate-promote.ll
This allows pointers to aggregate objects, whose elements are only read, to
be promoted and passed in by element instead of by reference.  This can
enable a LOT of subsequent optimizations in the caller function.

It's worth pointing out that this stuff happens a LOT of C++ programs, because
objects in templates are generally passed around by reference.  When these
templates are instantiated on small aggregate or scalar types, however, it is
more efficient to pass them in by value than by reference.

This transformation triggers most on C++ codes (e.g. 334 times on eon), but
does happen on C codes as well.  For example, on mesa it triggers 72 times,
and on gcc it triggers 35 times.  this is amazingly good considering that
we are using 'basicaa' so far.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12202 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-08 01:04:36 +00:00
Chris Lattner
86a734bd40 Implement: ArgumentPromotion/chained.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12200 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-07 22:52:53 +00:00
Chris Lattner
7db5a6df78 Fix another minor bug, exposed by perlbmk
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12198 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-07 22:43:27 +00:00
Chris Lattner
c76d80342e Since 'load null' is undefined, we can make it do whatever we want. Returning
a zero value is the most likely way to cause further simplification, so we do it.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12197 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-07 22:16:24 +00:00
Chris Lattner
7e6f5090af Fix a minor bug and turn debug output into, well, debug output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12195 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-07 21:54:50 +00:00
Chris Lattner
ed570a7dca New LLVM pass: argument promotion. This version only handles simple scalar
variables.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12193 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-07 21:29:54 +00:00
Chris Lattner
0a94348fb9 Don't emit things like malloc(16*1). Allocation instructions are fixed arity now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12086 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-03 01:40:53 +00:00
Misha Brukman
0256e4bbf9 Implement ExtractCodeRegion()
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12070 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-02 00:20:57 +00:00
Misha Brukman
38b8fd1078 Make a note that this is usually used via bugpoint.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12068 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-02 00:19:09 +00:00
Misha Brukman
b97fce5252 * Add implementation of ExtractBasicBlock()
* Add comments to ExtractLoop()


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12053 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-01 18:28:34 +00:00
Chris Lattner
99cca7d3aa Disable tail duplication in a case that breaks on Olden/tsp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12021 91177308-0d34-0410-b5e6-96231b3b80d8
2004-03-01 01:12:13 +00:00
Misha Brukman
99cc88bb64 * Remove function to find "main" in a Module, there's a method for that
* Removing extraneous empty space and empty comment lines


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12014 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-29 23:09:10 +00:00
Chris Lattner
06887c9a2a Fix bug: test/Regression/Transforms/LowerInvoke/2004-02-29-PHICrash.llx
... which tickled the lowerinvoke pass because it used the BCE routines.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12012 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-29 22:24:41 +00:00
Chris Lattner
4279f3c984 Fix PR255: [tailduplication] Single basic block loops are very rare
Note that this is a band-aid put over a band-aid.  This just undisables
tail duplication in on very specific case that it seems to work in.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11989 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-29 06:41:20 +00:00
Chris Lattner
542f149f00 Implement switch->br and br->switch folding by ripping out the switch->switch
and br->br code and generalizing it.  This allows us to compile code like this:

int test(Instruction *I) {
  if (isa<CastInst>(I))
    return foo(7);
  else if (isa<BranchInst>(I))
    return foo(123);
  else if (isa<UnwindInst>(I))
    return foo(1241);
  else if (isa<SetCondInst>(I))
    return foo(1);
  else if (isa<VAArgInst>(I))
    return foo(42);
  return foo(-1);
}

into:

int %_Z4testPN4llvm11InstructionE("struct.llvm::Instruction"* %I) {
entry:
        %tmp.1.i.i.i.i.i.i.i = getelementptr "struct.llvm::Instruction"* %I, long 0, ubyte 4            ; <uint*> [#uses=1]
        %tmp.2.i.i.i.i.i.i.i = load uint* %tmp.1.i.i.i.i.i.i.i          ; <uint> [#uses=2]
        %tmp.2.i.i.i.i.i.i = seteq uint %tmp.2.i.i.i.i.i.i.i, 27                ; <bool> [#uses=0]
        switch uint %tmp.2.i.i.i.i.i.i.i, label %endif.0 [
                 uint 27, label %then.0
                 uint 2, label %then.1
                 uint 5, label %then.2
                 uint 14, label %then.3
                 uint 15, label %then.3
                 uint 16, label %then.3
                 uint 17, label %then.3
                 uint 18, label %then.3
                 uint 19, label %then.3
                 uint 32, label %then.4
        ]
...

As well as handling the cases in 176.gcc and many other programs more effectively.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11964 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-28 21:28:10 +00:00
Chris Lattner
b99df4f8c2 if there is already a prototype for malloc/free, use it, even if it's incorrect.
Do not just inject a new prototype.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11951 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-28 18:51:45 +00:00
Chris Lattner
7bcc0e7fff Rename AddUsesToWorkList -> AddUsersToWorkList, as that is what it does.
Create a new AddUsesToWorkList method
optimize memmove/set/cpy of zero bytes to a noop.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11941 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-28 05:22:00 +00:00
Chris Lattner
6160e85201 Turn 'free null' into nothing
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11940 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-28 04:57:37 +00:00
Misha Brukman
9e4a642c03 Right, it's really Extractor, not Extraction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11939 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-28 03:37:58 +00:00
Misha Brukman
9401deb320 A pass that uses the generic CodeExtractor to rip out *every* loop in every
function, as long as the loop isn't the only one in that function. This should
help debugging passes easier with BugPoint.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11936 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-28 03:33:01 +00:00
Misha Brukman
e6336031b8 A generic code extractor: given a list of BasicBlocks, it will rip them out into
a new function, taking care of inputs and outputs.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11935 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-28 03:26:20 +00:00
Chris Lattner
7d90a2738e setcond instructions don't have aliasing implications.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11919 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-27 18:09:25 +00:00
Chris Lattner
a0ebcebc06 Implement test/Regression/Transforms/InstCombine/canonicalize_branch.ll
This is a really minor thing, but might help out the 'switch statement induction'
code in simplifycfg.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11900 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-27 06:27:46 +00:00
Chris Lattner
4f77caaa3d Since LLVM uses structure type equivalence, it isn't useful to keep around
multiple type names for the same structural type.  Make DTE eliminate all
but one of the type names


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11879 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-26 20:02:23 +00:00
Chris Lattner
f2dbf50efa turn things like:
if (X == 0 || X == 2)

...where the comparisons and branches are in different blocks... into a switch
instruction.  This comes up a lot in various programs, and works well with
the switch/switch merging code I checked earlier.  For example, this testcase:

int switchtest(int C) {
  return C == 0 ? f(123) :
         C == 1 ? f(3123) :
         C == 4 ? f(312) :
         C == 5 ? f(1234): f(444);
}

is converted into this:
        switch int %C, label %cond_false.3 [
                 int 0, label %cond_true.0
                 int 1, label %cond_true.1
                 int 4, label %cond_true.2
                 int 5, label %cond_true.3
        ]

instead of a whole bunch of conditional branches.

Admittedly the code is ugly, and incomplete.  To be complete, we need to add
br -> switch merging and switch -> br merging.  For example, this testcase:

struct foo { int Q, R, Z; };
#define A (X->Q+X->R * 123)
int test(struct foo *X) {
  return A  == 123 ? X1() :
        A == 12321 ? X2():
        (A == 111 || A == 222) ? X3() :
        A == 875 ? X4() : X5();
}

Gets compiled to this:
        switch int %tmp.7, label %cond_false.2 [
                 int 123, label %cond_true.0
                 int 12321, label %cond_true.1
                 int 111, label %cond_true.2
                 int 222, label %cond_true.2
        ]
...
cond_false.2:           ; preds = %entry
        %tmp.52 = seteq int %tmp.7, 875         ; <bool> [#uses=1]
        br bool %tmp.52, label %cond_true.3, label %cond_false.3

where the branch could be folded into the switch.

This kind of thing occurs *ALL OF THE TIME*, especially in programs like
176.gcc, which is a horrible mess of code.  It contains stuff like *shudder*:

#define SWITCH_TAKES_ARG(CHAR) \
  (   (CHAR) == 'D' \
   || (CHAR) == 'U' \
   || (CHAR) == 'o' \
   || (CHAR) == 'e' \
   || (CHAR) == 'u' \
   || (CHAR) == 'I' \
   || (CHAR) == 'm' \
   || (CHAR) == 'L' \
   || (CHAR) == 'A' \
   || (CHAR) == 'h' \
   || (CHAR) == 'z')

and

#define CONST_OK_FOR_LETTER_P(VALUE, C)                 \
  ((C) == 'I' ? SMALL_INTVAL (VALUE)                    \
   : (C) == 'J' ? SMALL_INTVAL (-(VALUE))               \
   : (C) == 'K' ? (unsigned)(VALUE) < 32                \
   : (C) == 'L' ? ((VALUE) & 0xffff) == 0               \
   : (C) == 'M' ? integer_ok_for_set (VALUE)            \
   : (C) == 'N' ? (VALUE) < 0                           \
   : (C) == 'O' ? (VALUE) == 0                          \
   : (C) == 'P' ? (VALUE) >= 0                          \
   : 0)

and

#define LEGITIMIZE_ADDRESS(X,OLDX,MODE,WIN)                     \
{                                                               \
  if (GET_CODE (X) == PLUS && CONSTANT_ADDRESS_P (XEXP (X, 1))) \
    (X) = gen_rtx (PLUS, SImode, XEXP (X, 0),                   \
                   copy_to_mode_reg (SImode, XEXP (X, 1)));     \
  if (GET_CODE (X) == PLUS && CONSTANT_ADDRESS_P (XEXP (X, 0))) \
    (X) = gen_rtx (PLUS, SImode, XEXP (X, 1),                   \
                   copy_to_mode_reg (SImode, XEXP (X, 0)));     \
  if (GET_CODE (X) == PLUS && GET_CODE (XEXP (X, 0)) == MULT)   \
    (X) = gen_rtx (PLUS, SImode, XEXP (X, 1),                   \
                   force_operand (XEXP (X, 0), 0));             \
  if (GET_CODE (X) == PLUS && GET_CODE (XEXP (X, 1)) == MULT)   \
    (X) = gen_rtx (PLUS, SImode, XEXP (X, 0),                   \
                   force_operand (XEXP (X, 1), 0));             \
  if (GET_CODE (X) == PLUS && GET_CODE (XEXP (X, 0)) == PLUS)   \
    (X) = gen_rtx (PLUS, Pmode, force_operand (XEXP (X, 0), NULL_RTX),\
                   XEXP (X, 1));                                \
  if (GET_CODE (X) == PLUS && GET_CODE (XEXP (X, 1)) == PLUS)   \
    (X) = gen_rtx (PLUS, Pmode, XEXP (X, 0),                    \
                   force_operand (XEXP (X, 1), NULL_RTX));      \
  if (GET_CODE (X) == SYMBOL_REF || GET_CODE (X) == CONST       \
           || GET_CODE (X) == LABEL_REF)                        \
    (X) = legitimize_address (flag_pic, X, 0, 0);               \
  if (memory_address_p (MODE, X))                               \
    goto WIN; }

and others.  These macros get used multiple times of course.  These are such
lovely candidates for macros, aren't they?  :)

This code also nicely handles LLVM constructs that look like this:

  if (isa<CastInst>(I))
   ...
  else if (isa<BranchInst>(I))
   ...
  else if (isa<SetCondInst>(I))
   ...
  else if (isa<UnwindInst>(I))
   ...
  else if (isa<VAArgInst>(I))
   ...

where the isa can obviously be a dyn_cast as well.  Switch instructions are a
good thing.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11870 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-26 07:13:46 +00:00
Chris Lattner
079236d1c9 My faith in programmers has been found to be totally misplaced. One would
assume that if they don't intend to write to a global variable, that they
would mark it as constant.  However, there are people that don't understand
that the compiler can do nice things for them if they give it the information
it needs.

This pass looks for blatently obvious globals that are only ever read from.
Though it uses a trivially simple "alias analysis" of sorts, it is still able
to do amazing things to important benchmarks.  253.perlbmk, for example,
contains several ***GIANT*** function pointer tables that are not marked
constant and should be.  Marking them constant allows the optimizer to turn
a whole bunch of indirect calls into direct calls.  Note that only a link-time
optimizer can do this transformation, but perlbmk does have several strings
and other minor globals that can be marked constant by this pass when run
from GCCAS.

176.gcc has a ton of strings and large tables that are marked constant, both
at compile time (38 of them) and at link time (48 more).  Other benchmarks
give similar results, though it seems like big ones have disproportionally
more than small ones.

This pass is extremely quick and does good things.  I'm going to enable it
in gccas & gccld.  Not bad for 50 SLOC.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11836 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-25 21:34:36 +00:00
Chris Lattner
940ff563f7 Fix incorrect debug code
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11821 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-25 15:15:04 +00:00
Chris Lattner
d38f208cd9 Fix a faulty optimization on FP values
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11801 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-24 18:10:14 +00:00
Chris Lattner
8e509dd5e8 If a block is made dead, make sure to promptly remove it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11799 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-24 16:09:21 +00:00
Chris Lattner
d52c261cf8 Implement SimplifyCFG/switch_switch_fold.ll
This case occurs many times in various benchmarks, especially when combined
with the previous patch.  This allows it to get stuff like:
  if (X == 4 || X == 3)
    if (X == 5 || X == 8)

and

switch (X) {
case 4: case 5: case 6:
  if (X == 4 || X == 5)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11797 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-24 07:23:58 +00:00
Chris Lattner
e14ea0804a Rearrange code a bit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11793 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-24 05:54:22 +00:00
Chris Lattner
0d56008f53 Implement: test/Regression/Transforms/SimplifyCFG/switch_create.ll
This turns code like this:
  if (X == 4 | X == 7)
and
  if (X != 4 & X != 7)
into switch instructions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11792 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-24 05:38:11 +00:00
Chris Lattner
d8864ce766 Generate much more efficient code in programs like pifft
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11775 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-23 21:46:58 +00:00
Chris Lattner
077a373791 Fix a small typeo in my checkin last night that broke vortex and other programs :(
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11774 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-23 21:46:42 +00:00
Chris Lattner
8adac75298 Fix InstCombine/2004-02-23-ShiftShiftOverflow.ll
Also, turn 'shr int %X, 1234' into 'shr int %X, 31'


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11768 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-23 20:30:06 +00:00
Chris Lattner
c5943fb186 Implement cast.ll::test14/15
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11742 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-23 07:16:20 +00:00
Chris Lattner
4cb170cc6e Refactor some code. In the mul - setcc folding case, we really care about
whether this is the sign bit or not, so check unsigned comparisons as well.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11740 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-23 06:38:22 +00:00
Chris Lattner
fed58fd72e Implement mul.ll:test11
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11737 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-23 06:00:11 +00:00
Chris Lattner
45aaafef49 Implement "strength reduction" of X <= C and X >= C
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11735 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-23 05:47:48 +00:00
Chris Lattner
fb54b2b744 Implement InstCombine/mul.ll:test10, which is a case that occurs when dealing
with "predication"


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11734 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-23 05:39:21 +00:00
Chris Lattner
c6bd195336 Implement Transforms/InstCombine/cast.ll:test13, a case which occurs in a
hot 164.gzip loop.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11702 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-22 05:25:17 +00:00
Chris Lattner
57cb9883aa Fix PR245: Linking weak and strong global variables is dependent on link order
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11565 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-17 21:56:04 +00:00
Chris Lattner
19831ec853 Implement test/Regression/Transforms/SimplifyCFG/UncondBranchToReturn.ll,
see the testcase for the reasoning.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11496 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-16 06:35:48 +00:00
Chris Lattner
7059f2e76b Fold PHI nodes of constants which are only used by a single cast. This implements
phi.ll:test4


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11494 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-16 05:07:08 +00:00
Chris Lattner
26ca7e145e Teach LLVM to unravel the "swap idiom". This implements:
Regression/Transforms/InstCombine/xor.ll:test20


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11492 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-16 03:54:20 +00:00
Chris Lattner
c317d39887 Implement Transforms/InstCombine/xor.ll:test19
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11490 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-16 01:20:27 +00:00
Chris Lattner
860a16143c Instead of producing calls to setjmp/longjmp, produce uses of the
llvm.setjmp/llvm.longjmp intrinsics.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@11482 91177308-0d34-0410-b5e6-96231b3b80d8
2004-02-15 22:24:27 +00:00