Commit Graph

1933 Commits

Author SHA1 Message Date
Chris Lattner
641f02f10f Fix Reassociate/shifttest.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21839 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-10 03:39:25 +00:00
Chris Lattner
7f78f218fd If a function contains no allocas, all of the calls in it are trivially
suitable for tail calls.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21836 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-09 23:51:13 +00:00
Chris Lattner
6e4c6495eb implement and.ll:test33
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21809 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-09 04:58:36 +00:00
Chris Lattner
f201dbc1a4 Preserve calling conventions when doing IPO
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21798 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-09 01:05:50 +00:00
Chris Lattner
c154cef9a1 wrap long lines, preserve calling conventions when cloning functions and
turning calls into invokes


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21797 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-09 01:04:34 +00:00
Chris Lattner
fb217adf92 Convert non-address taken functions with C calling conventions to fastcc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21791 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 22:18:06 +00:00
Chris Lattner
44b8c7d5d3 Implement Reassociate/mul-neg-add.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21788 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 21:41:35 +00:00
Chris Lattner
6f156856ca Bail out earlier
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21786 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 21:33:47 +00:00
Chris Lattner
f33151aff0 Teach reassociate that 0-X === X*-1
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21785 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 21:28:52 +00:00
Chris Lattner
9c72319938 Fix PR557 and basictest[34].ll.
This makes reassociate realize that loads should be treated as unmovable, and
gives distinct ranks to distinct values defined in the same basic block, allowing
reassociate to do its thing.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21783 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 20:57:04 +00:00
Chris Lattner
c9fd097a01 Add debugging information
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21781 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 20:09:57 +00:00
Chris Lattner
989f6229bc eliminate gotos
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21780 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 19:48:43 +00:00
Chris Lattner
109d34d6ff Improve reassociation handling of inverses, implementing inverses.ll.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21778 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 18:59:37 +00:00
Chris Lattner
928128281f clean up and modernize this pass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21776 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 18:45:26 +00:00
Chris Lattner
120347e8d1 Strength reduce SAR into SHR if there is no way sign bits could be shifted
in.  This tends to get cases like this:

  X = cast ubyte to int
  Y = shr int X, ...

Tested by: shift.ll:test24


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21775 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 17:34:56 +00:00
Chris Lattner
4690010006 Refactor some code
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21772 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 00:19:31 +00:00
Chris Lattner
cc8a2b98f2 Handle some simple cases where we can see that values get annihilated.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21771 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-08 00:08:33 +00:00
Chris Lattner
f4d4c87985 Fix a miscompilation of crafty by clobbering the "A" variable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21770 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-07 23:49:08 +00:00
Chris Lattner
c0649ac931 Rewrite the guts of the reassociate pass to be more efficient and logical. Instead
of trying to do local reassociation tweaks at each level, only process an expression
tree once (at its root).  This does not improve the reassociation pass in any real way.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21768 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-07 21:59:39 +00:00
Reid Spencer
789082af03 * Add two strlen optimizations:
strlen(x) != 0 -> *x != 0
    strlen(x) == 0 -> *x == 0
* Change nested statistics to use style of other LLVM statistics so that
  only the name of the optimization (simplify-libcalls) is used as the
  statistic name, and the description indicates which specific all is
  optimized. Cuts down on some redundancy and saves a few bytes of space.
* Make note of stpcpy optimization that could be done.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21766 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-07 20:15:59 +00:00
Reid Spencer
673c1a9170 Don't increment the counter unless the debug flag is set.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21762 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-07 04:59:45 +00:00
Chris Lattner
0975ed5f4e Convert shifts to muls to assist reassociation. This implements
Reassociate/shifttest.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21761 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-07 04:24:13 +00:00
Chris Lattner
08b43921e1 Simplify the code and rearrange it. No major functionality changes here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21759 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-07 04:08:02 +00:00
Chris Lattner
3799ed83b4 BAD typeo which caused many testsuite failures last night. Note to self, do
not change code after testing it without retesting!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21741 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 17:13:16 +00:00
Chris Lattner
a9e9211bbb Preserve tail marker
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21737 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 06:48:21 +00:00
Chris Lattner
1b49141821 Implement Transforms/Inline/inline-tail.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21736 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 06:47:52 +00:00
Chris Lattner
1430ef134d preserve the tail marker
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21734 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 06:46:58 +00:00
Chris Lattner
19bdc03adc Wrap long lines
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21720 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 05:34:40 +00:00
Chris Lattner
ec710c5b12 DCE intrinsic instructions without side effects.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21719 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 05:27:34 +00:00
Chris Lattner
eabfa4787a Teach instcombine propagate zeroness through shl instructions, implementing
and.ll:test31


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21717 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 04:53:20 +00:00
Chris Lattner
d7115b01a0 Implement shift.ll:test23. If we are shifting right then immediately truncating
the result, turn signed shift rights into unsigned shift rights if possible.

This leads to later simplification and happens *often* in 176.gcc.  For example,
this testcase:

struct xxx { unsigned int code : 8; };
enum codes { A, B, C, D, E, F };
int foo(struct xxx *P) {
  if ((enum codes)P->code == A)
     bar();
}

used to be compiled to:

int %foo(%struct.xxx* %P) {
        %tmp.1 = getelementptr %struct.xxx* %P, int 0, uint 0           ; <uint*> [#uses=1]
        %tmp.2 = load uint* %tmp.1              ; <uint> [#uses=1]
        %tmp.3 = cast uint %tmp.2 to int                ; <int> [#uses=1]
        %tmp.4 = shl int %tmp.3, ubyte 24               ; <int> [#uses=1]
        %tmp.5 = shr int %tmp.4, ubyte 24               ; <int> [#uses=1]
        %tmp.6 = cast int %tmp.5 to sbyte               ; <sbyte> [#uses=1]
        %tmp.8 = seteq sbyte %tmp.6, 0          ; <bool> [#uses=1]
        br bool %tmp.8, label %then, label %UnifiedReturnBlock

Now it is compiled to:

        %tmp.1 = getelementptr %struct.xxx* %P, int 0, uint 0           ; <uint*> [#uses=1]
        %tmp.2 = load uint* %tmp.1              ; <uint> [#uses=1]
        %tmp.2 = cast uint %tmp.2 to sbyte              ; <sbyte> [#uses=1]
        %tmp.8 = seteq sbyte %tmp.2, 0          ; <bool> [#uses=1]
        br bool %tmp.8, label %then, label %UnifiedReturnBlock

which is the difference between this:

foo:
        subl $4, %esp
        movl 8(%esp), %eax
        movl (%eax), %eax
        shll $24, %eax
        sarl $24, %eax
        testb %al, %al
        jne .LBBfoo_2

and this:

foo:
        subl $4, %esp
        movl 8(%esp), %eax
        movl (%eax), %eax
        testb %al, %al
        jne .LBBfoo_2

This occurs 3243 times total in the External tests, 215x in povray,
6x in each f2c'd program, 1451x in 176.gcc, 7x in crafty, 20x in perl,
25x in gap, 3x in m88ksim, 25x in ijpeg.

Maybe this will cause a little jump on gcc tommorow :)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21715 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 04:18:52 +00:00
Chris Lattner
7aed7ac5e1 Implement xor.ll:test22
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21713 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 02:07:39 +00:00
Chris Lattner
d152380ba8 implement and.ll:test30 and set.ll:test21
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21712 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 01:53:19 +00:00
Chris Lattner
828eeddb38 implement or.ll:test20
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21709 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-06 00:58:50 +00:00
Chris Lattner
3987abdfe3 Fix a bug compimling Ruby, fixing this testcase:
LowerSetJmp/2005-05-05-OldUses.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21696 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-05 15:47:43 +00:00
Chris Lattner
693787a5ec Instcombine: cast (X != 0) to int, cast (X == 1) to int -> X iff X has only the low bit set.
This implements set.ll:test20.

This triggers 2x on povray, 9x on mesa, 11x on gcc, 2x on crafty, 1x on eon,
6x on perlbmk and 11x on m88ksim.

It allows us to compile these two functions into the same code:

struct s { unsigned int bit : 1; };
unsigned foo(struct s *p) {
  if (p->bit)
    return 1;
  else
    return 0;
}
unsigned bar(struct s *p) { return p->bit; }


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21690 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-04 19:10:26 +00:00
Reid Spencer
cea6559f60 Implement the IsDigitOptimization for simplifying calls to the isdigit
library function:
  isdigit(chr) -> 0 or 1 if chr is constant
  isdigit(chr) -> chr - '0' <= 9 otherwise

Although there are many calls to isdigit in llvm-test, most of them are
compiled away by macros leaving only this:

2 MultiSource/Applications/hexxagon


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21688 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-04 18:58:28 +00:00
Reid Spencer
58b563ce43 * Correct the function prototypes for some of the functions to match the
actual spec (int -> uint)
* Add the ability to get/cache the strlen function prototype.
* Make sure generated values are appropriately named for debugging purposes
* Add the SPrintFOptimiation for 4 casts of sprintf optimization:
    sprintf(str,cstr) -> llvm.memcpy(str,cstr) (if cstr has no %)
    sprintf(str,"")   -> store sbyte 0, str
    sprintf(str,"%s",src) -> llvm.memcpy(str,src) (if src is constant)
    sprintf(str,"%c",chr) -> store chr, str   ; store sbyte 0, str+1

The sprintf optimization didn't fire as much as I had hoped:

  2 MultiSource/Applications/SPASS
  5 MultiSource/Benchmarks/McCat/18-imp
 22 MultiSource/Benchmarks/Prolangs-C/TimberWolfMC
  1 MultiSource/Benchmarks/Prolangs-C/assembler
  6 MultiSource/Benchmarks/Prolangs-C/unix-smail
  2 MultiSource/Benchmarks/mediabench/mpeg2/mpeg2dec


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21679 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-04 03:20:21 +00:00
Reid Spencer
21506ff821 Implement optimizations for the strchr and llvm.memset library calls.
Neither of these activated as many times as was hoped:

strchr:
9 MultiSource/Applications/siod
1 MultiSource/Applications/d
2 MultiSource/Prolangs-C/archie-client
1 External/SPEC/CINT2000/176.gcc/176.gcc

llvm.memset:
no hits


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21669 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-03 07:23:44 +00:00
Reid Spencer
9974ddab2b Avoid garbage output in the statistics display by ensuring that the
strings passed to Statistic's constructor are not destructable. The stats
are printed during static destruction and the SimplifyLibCalls module was
getting destructed before the statistics.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21661 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-03 02:54:54 +00:00
Reid Spencer
e6ec8cc8c6 Add the StrNCmpOptimization which is similar to strcmp.
Unfortunately, this optimization didn't trigger on any llvm-test tests.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21660 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-03 01:43:45 +00:00
Reid Spencer
a1b4390938 Implement the fprintf optimization which converts calls like this:
fprintf(F,"hello") -> fwrite("hello",strlen("hello"),1,F)
  fprintf(F,"%s","hello") -> fwrite("hello",strlen("hello"),1,F)
  fprintf(F,"%c",'x') -> fputc('c',F)

This optimization fires severals times in llvm-test:

313 MultiSource/Applications/Burg
302 MultiSource/Benchmarks/Prolangs-C/TimberWolfMC
189 MultiSource/Benchmarks/Prolangs-C/mybison
175 MultiSource/Benchmarks/Prolangs-C/football
130 MultiSource/Benchmarks/Prolangs-C/unix-tbl


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21657 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-02 23:59:26 +00:00
John Criswell
fe3706a12b Fixed a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21653 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-02 14:47:42 +00:00
Chris Lattner
9fb25db11e Implement getelementptr.ll:test11
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21647 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-01 04:42:15 +00:00
Chris Lattner
37366c1ba8 Check for volatile loads only once.
Implement load.ll:test7


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21645 91177308-0d34-0410-b5e6-96231b3b80d8
2005-05-01 04:24:53 +00:00
Reid Spencer
63a7513ac1 Fix a comment that stated the wrong thing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21638 91177308-0d34-0410-b5e6-96231b3b80d8
2005-04-30 06:45:47 +00:00
Reid Spencer
9f56b1feeb * Don't depend on "guessing" what a FILE* is, just require that the actual
type be obtained from a CallInst we're optimizing.
* Make it possible for getConstantStringLength to return the ConstantArray
  that it extracts in case the content is needed by an Optimization.
* Implement the strcmp optimization
* Implement the toascii optimization

This pass is now firing several to many times in the following MultiSource
tests:

Applications/Burg      -   7 (strcat,strcpy)
Applications/siod      -  13 (strcat,strcpy,strlen)
Applications/spiff     - 120 (exit,fputs,strcat,strcpy,strlen)
Applications/treecc    -  66 (exit,fputs,strcat,strcpy)
Applications/kimwitu++ -  34 (strcmp,strcpy,strlen)
Applications/SPASS     - 588 (exit,fputs,strcat,strcpy,strlen)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21626 91177308-0d34-0410-b5e6-96231b3b80d8
2005-04-30 03:17:54 +00:00
Reid Spencer
ff5525daf3 Implement the optimizations for "pow" and "fputs" library calls.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21618 91177308-0d34-0410-b5e6-96231b3b80d8
2005-04-29 09:39:47 +00:00
Reid Spencer
df4945bcb5 Remove optimizations that don't require both operands to be constant. These
are moved to simplify-libcalls pass.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21614 91177308-0d34-0410-b5e6-96231b3b80d8
2005-04-29 05:55:35 +00:00
Jeff Cohen
5882b92360 Consistently use 'class' to silence VC++
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@21612 91177308-0d34-0410-b5e6-96231b3b80d8
2005-04-29 03:05:44 +00:00