Commit Graph

1361 Commits

Author SHA1 Message Date
Chris Lattner
3280f7bd55 Move isLoopInvariant to the Loop class
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13051 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 22:46:08 +00:00
Chris Lattner
59fb87d469 Correct rewriting of exit blocks after my last patch
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13048 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 22:27:10 +00:00
Chris Lattner
5fa802fe27 Loop exit sets are no longer explicitly held, they are dynamically computed on demand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13046 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 22:15:13 +00:00
Chris Lattner
f1ab4b4eac Change the ExitBlocks list from being explicitly contained in the Loop
structure to being dynamically computed on demand.  This makes updating
loop information MUCH easier.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13045 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 22:14:10 +00:00
Chris Lattner
e3a55865bb Reduce the unrolling limit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13040 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 18:06:14 +00:00
Chris Lattner
cc43909bb6 If the preheader of the loop was the entry block of the function, make sure
that the exit block of the loop becomes the new entry block of the function.

This was causing a verifier assertion on 252.eon.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13039 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 17:38:42 +00:00
Chris Lattner
998f44f859 Be much more careful about how we update instructions outside of the loop
using instructions inside of the loop.  This should fix the MishaTest failure
from last night.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13038 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 17:32:39 +00:00
Chris Lattner
50ca0a195a After unrolling our single basic block loop, fold it into the preheader and exit
block.  The primary motivation for doing this is that we can now unroll nested loops.

This makes a pretty big difference in some cases.  For example, in 183.equake,
we are now beating the native compiler with the CBE, and we are a lot closer
with LLC.

I'm now going to play around a bit with the unroll factor and see what effect
it really has.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13034 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 06:27:43 +00:00
Chris Lattner
9c2cc4626b Fix a bug: this does not preserve the CFG!
While we're at it, add support for updating loop information correctly.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13033 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 05:38:37 +00:00
Chris Lattner
83bf2880c3 Initial checkin of a simple loop unroller. This pass is extremely basic and
limited.  Even in it's extremely simple state (it can only *fully* unroll single
basic block loops that execute a constant number of times), it already helps improve
performance a LOT on some benchmarks, particularly with the native code generators.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13028 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 05:20:17 +00:00
Chris Lattner
3c85eef81a Make the tail duplication threshold accessible from the command line instead of hardcoded
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13025 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-18 00:52:43 +00:00
Chris Lattner
20aa098ba6 If the loop executes a constant number of times, try a bit harder to replace
exit values.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13018 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-17 18:44:09 +00:00
Chris Lattner
4f1134e51a Fix a HUGE pessimization on X86. The indvars pass was taking this
(familiar) function:

int _strlen(const char *str) {
    int len = 0;
    while (*str++) len++;
    return len;
}

And transforming it to use a ulong induction variable, because the type of
the pointer index was left as a constant long.  This is obviously very bad.

The fix is to shrink long constants in getelementptr instructions to intptr_t,
making the indvars pass insert a uint induction variable, which is much more
efficient.

Here's the before code for this function:

int %_strlen(sbyte* %str) {
entry:
        %tmp.13 = load sbyte* %str              ; <sbyte> [#uses=1]
        %tmp.24 = seteq sbyte %tmp.13, 0                ; <bool> [#uses=1]
        br bool %tmp.24, label %loopexit, label %no_exit

no_exit:                ; preds = %entry, %no_exit
***     %indvar = phi uint [ %indvar.next, %no_exit ], [ 0, %entry ]            ; <uint> [#uses=2]
***     %indvar = phi ulong [ %indvar.next, %no_exit ], [ 0, %entry ]           ; <ulong> [#uses=2]
        %indvar1 = cast ulong %indvar to uint           ; <uint> [#uses=1]
        %inc.02.sum = add uint %indvar1, 1              ; <uint> [#uses=1]
        %inc.0.0 = getelementptr sbyte* %str, uint %inc.02.sum          ; <sbyte*> [#uses=1]
        %tmp.1 = load sbyte* %inc.0.0           ; <sbyte> [#uses=1]
        %tmp.2 = seteq sbyte %tmp.1, 0          ; <bool> [#uses=1]
        %indvar.next = add ulong %indvar, 1             ; <ulong> [#uses=1]
        %indvar.next = add uint %indvar, 1              ; <uint> [#uses=1]
        br bool %tmp.2, label %loopexit.loopexit, label %no_exit

loopexit.loopexit:              ; preds = %no_exit
        %indvar = cast uint %indvar to int              ; <int> [#uses=1]
        %inc.1 = add int %indvar, 1             ; <int> [#uses=1]
        ret int %inc.1

loopexit:               ; preds = %entry
        ret int 0
}


Here's the after code:

int %_strlen(sbyte* %str) {
entry:
        %inc.02 = getelementptr sbyte* %str, uint 1             ; <sbyte*> [#uses=1]
        %tmp.13 = load sbyte* %str              ; <sbyte> [#uses=1]
        %tmp.24 = seteq sbyte %tmp.13, 0                ; <bool> [#uses=1]
        br bool %tmp.24, label %loopexit, label %no_exit

no_exit:                ; preds = %entry, %no_exit
***     %indvar = phi uint [ %indvar.next, %no_exit ], [ 0, %entry ]            ; <uint> [#uses=3]
        %indvar = cast uint %indvar to int              ; <int> [#uses=1]
        %inc.0.0 = getelementptr sbyte* %inc.02, uint %indvar           ; <sbyte*> [#uses=1]
        %inc.1 = add int %indvar, 1             ; <int> [#uses=1]
        %tmp.1 = load sbyte* %inc.0.0           ; <sbyte> [#uses=1]
        %tmp.2 = seteq sbyte %tmp.1, 0          ; <bool> [#uses=1]
        %indvar.next = add uint %indvar, 1              ; <uint> [#uses=1]
        br bool %tmp.2, label %loopexit, label %no_exit

loopexit:               ; preds = %entry, %no_exit
        %len.0.1 = phi int [ 0, %entry ], [ %inc.1, %no_exit ]          ; <int> [#uses=1]
        ret int %len.0.1
}


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13016 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-17 18:16:10 +00:00
Chris Lattner
f50af088f1 Even if there are not any induction variables in the loop, if we can compute
the trip count for the loop, insert one so that we can canonicalize the exit
condition.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13015 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-17 18:08:33 +00:00
Chris Lattner
f5b9eb37a9 Add support for evaluation of exp/log/log10/pow
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13011 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-16 22:35:33 +00:00
Chris Lattner
657253119e Fix some really nasty dominance bugs that were exposed by my patch to
make the verifier more strict.  This fixes building zlib


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13002 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-16 18:08:07 +00:00
Brian Gaeke
c598517c24 Include <cmath> for compatibility with gcc 3.0.x (the system compiler on
Debian.)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12986 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-16 15:57:32 +00:00
Chris Lattner
006118fe8c Fix some of the strange CBE-only failures that happened last night.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12980 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-16 06:03:17 +00:00
Chris Lattner
cd01ae5c70 Fix Inline/2004-04-15-InlineDeletesCall.ll
Basically we were using SimplifyCFG as a huge sledgehammer for a simple
optimization.  Because simplifycfg does so many things, we can't use it
for this purpose.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12977 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-16 05:17:59 +00:00
Chris Lattner
d244057a48 Fix a bug in the previous checkin: if the exit block is not the same as
the back-edge block, we must check the preincremented value.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12968 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-15 20:26:22 +00:00
Chris Lattner
59fdaeeae8 Change the canonical induction variable that we insert.
Instead of producing code like this:

Loop:
  X = phi 0, X2
  ...

  X2 = X + 1
  if (X != N-1) goto Loop

We now generate code that looks like this:

Loop:
  X = phi 0, X2
  ...

  X2 = X + 1
  if (X2 != N) goto Loop

This has two big advantages:
  1. The trip count of the loop is now explicit in the code, allowing
     the direct implementation of Loop::getTripCount()
  2. This reduces register pressure in the loop, and allows X and X2 to be
     put into the same register.

As a consequence of the second point, the code we generate for loops went
from:

.LBB2:  # no_exit.1
	...
        mov %EDI, %ESI
        inc %EDI
        cmp %ESI, 2
        mov %ESI, %EDI
        jne .LBB2 # PC rel: no_exit.1

To:

.LBB2:  # no_exit.1
	...
        inc %ESI
        cmp %ESI, 3
        jne .LBB2 # PC rel: no_exit.1

... which has two fewer moves, and uses one less register.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12961 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-15 15:21:43 +00:00
Chris Lattner
d9955aaa61 ADd a trivial instcombine: load null -> null
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12940 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-14 03:28:36 +00:00
Chris Lattner
58b7b08ad7 Add SCCP support for constant folding calls, implementing:
test/Regression/Transforms/SCCP/calltest.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12921 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-13 19:43:54 +00:00
Chris Lattner
25b8390bf2 Add a simple call constant propagation interface.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12919 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-13 19:28:52 +00:00
Chris Lattner
4647b7cd69 Constant propagation should remove the dead instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12917 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-13 19:28:20 +00:00
Chris Lattner
4b66242c54 Fix LoopSimplify/2004-04-13-LoopSimplifyUpdateDomFrontier.ll
LoopSimplify was not updating dominator frontiers correctly in some cases.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12890 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-13 16:23:25 +00:00
Chris Lattner
1f62f82b05 Refactor code a bit to make it simpler and eliminate the goto
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12888 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-13 15:21:18 +00:00
Chris Lattner
529b28da45 This patch addresses PR35: Loop simplify should reconstruct nested loops.
This is fairly straight-forward, but was a real nightmare to get just
perfect.  aarg.  :)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12884 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-13 05:05:33 +00:00
Chris Lattner
ce1a3a292d Actually update the call graph as the inliner changes it. This allows us to
execute other CallGraphSCCPasses after the inliner without crashing.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12861 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-12 05:37:29 +00:00
Chris Lattner
1708d12916 Add support for removing invoke instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12858 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-12 05:15:13 +00:00
Chris Lattner
85eb157d96 Stop printing Function*
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12857 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-12 04:06:56 +00:00
Chris Lattner
d08ddd3300 Simplify code a bit, and be sure to mark the external node as potentially throwing
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12856 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-12 04:06:38 +00:00
Chris Lattner
fbede52af4 Fix a bug in my select transformation
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12826 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-11 01:39:19 +00:00
Chris Lattner
adb7c0de39 Update the value numbering interface.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12824 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 22:33:34 +00:00
Chris Lattner
d76956d444 Implement InstCombine/select.ll:test13*
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12821 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 22:21:27 +00:00
Chris Lattner
66331a4e33 Implement InstCombine/add.ll:test20
Canonicalize add of sign bit constant into a xor


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12819 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 22:01:55 +00:00
Chris Lattner
38e66bd9d1 Rewrite the GCSE pass to be *substantially* simpler, a bit more efficient,
and a bit more powerful


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12817 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 21:11:11 +00:00
Chris Lattner
326c0f3013 Fix spurious warning in release mode
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12816 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 19:15:56 +00:00
Chris Lattner
a5f4103931 Simplify code a bit, and fix a bug that was breaking perlbmk
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12814 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 18:06:21 +00:00
Chris Lattner
562cb66b84 Fix a bug in my checkin last night that was breaking programs using invoke.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12813 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 16:53:29 +00:00
Chris Lattner
27c694bacb Fix previous patch
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12811 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 07:27:48 +00:00
Chris Lattner
17177601fb Correctly update counters
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12810 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 07:02:02 +00:00
Chris Lattner
ede6ac60a5 Simplify code a bit, and use alias analysis to allow us to delete unused
call and invoke instructions that are known to not write to memory.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12807 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-10 06:53:09 +00:00
Chris Lattner
e576b91948 Implement select.ll:test12*
This transforms code like this:

   %C = or %A, %B
   %D = select %cond, %C, %A
into:
   %C = select %cond, %B, 0
   %D = or %A, %C

Since B is often a constant, the select can often be eliminated.  In any case,
this reduces the usage count of A, allowing subsequent optimizations to happen.

This xform applies when the operator is any of:
  add, sub, mul, or, xor, and, shl, shr


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12800 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-09 23:46:01 +00:00
Chris Lattner
570751c2a7 Fold code like:
if (C)
    V1 |= V2;

into:
  Vx = V1 | V2;
  V1 = select C, V1, Vx

when the expression can be evaluated unconditionally and is *cheap* to
execute.  This limited form of if conversion is quite handy in lots of cases.
For example, it turns this testcase into straight-line code:

int in0 ; int in1 ; int in2 ; int in3 ;
int in4 ; int in5 ; int in6 ; int in7 ;
int in8 ; int in9 ; int in10; int in11;
int in12; int in13; int in14; int in15;
long output;

void mux(void) {
  output =
      (in0   ?  0x00000001 : 0) | (in1   ?  0x00000002 : 0) |
      (in2   ?  0x00000004 : 0) | (in3   ?  0x00000008 : 0) |
      (in4   ?  0x00000010 : 0) | (in5   ?  0x00000020 : 0) |
      (in6   ?  0x00000040 : 0) | (in7   ?  0x00000080 : 0) |
      (in8   ?  0x00000100 : 0) | (in9   ?  0x00000200 : 0) |
      (in10  ?  0x00000400 : 0) | (in11  ?  0x00000800 : 0) |
      (in12  ?  0x00001000 : 0) | (in13  ?  0x00002000 : 0) |
      (in14  ?  0x00004000 : 0) | (in15  ?  0x00008000 : 0) ;
}


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12798 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-09 22:50:22 +00:00
Chris Lattner
2eefe518de Fold binary operators with a constant operand into select instructions
that have a constant operand.  This implements
add.ll:test19, shift.ll:test15*, and others that are not tested


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12794 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-09 19:05:30 +00:00
Chris Lattner
82e14fe168 Implement select.ll:test11
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12793 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-09 18:19:44 +00:00
Chris Lattner
f499eaca36 Implement InstCombine/cast-propagate.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12784 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-08 20:39:49 +00:00
Chris Lattner
8ba725b456 Implement ScalarRepl/select_promote.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12779 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-08 19:59:34 +00:00
Chris Lattner
775cbdd51a Remove the "really gross hacks" that are there to deal with recursive functions.
Now we collect all of the call sites we are interested in inlining, then inline
them.  This entirely avoids issues with trying to inline a call site we got by
inlining another call site.  This also eliminates iterator invalidation issues.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@12770 91177308-0d34-0410-b5e6-96231b3b80d8
2004-04-08 06:34:31 +00:00