Commit Graph

34006 Commits

Author SHA1 Message Date
Devang Patel
bacf5193cf ExitCondition and Induction variable are loop constraints
not split condition constraints.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40977 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-10 00:33:50 +00:00
Chris Lattner
b31e91cb16 unbreak the build
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40976 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 23:55:17 +00:00
Evan Cheng
a86e347f02 Bug fix. ~1U marks the val# dead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40975 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 23:14:39 +00:00
Dale Johannesen
ebbc95da49 Patch 10 for long double. Doing constants right needs expanding ConstantFP
to handle values bigger than double.  If we assume host==target and host
long double works correctly, this is not too bad, but we don't want to
have that limitation longterm.  I could implement accepting double
constants as long double or something like that, which would lead to
incorrect codegen with no errors; the more I think about that the worse
it seems.  Rather than do such a hack that would be backed out later,
I'm settling for giving reasonable error messages, for now.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40974 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 22:51:36 +00:00
Evan Cheng
a3231ba237 Temporarily backing out this change until we know why some dejagnu tests are failing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40973 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 22:25:35 +00:00
Evan Cheng
f7ef26e701 divb / mulb outputs to ah. Under x86-64 it's not legal to read ah if the instruction requires a rex prefix (i.e. outputs to r8b, etc.). So issue shift right by 8 on AX and then truncate it to 8 bits instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40972 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 21:59:35 +00:00
Chris Lattner
f2369f2042 when we see a unaligned load from an insufficiently aligned global or
alloca, increase the alignment of the load, turning it into an aligned load.

This allows us to compile:

#include <xmmintrin.h>
__m128i foo(__m128i x){
 static const unsigned int c_0[4] = { 0, 0, 0, 0 };
	  __m128i v_Zero = _mm_loadu_si128((__m128i*)c_0);
  x  = _mm_unpacklo_epi8(x,  v_Zero);
  return x;
}

into:

_foo:
	punpcklbw	_c_0.5944, %xmm0
	ret
	.data
	.lcomm	_c_0.5944,16,4		# c_0.5944

instead of:

_foo:
	movdqu	_c_0.5944, %xmm1
	punpcklbw	%xmm1, %xmm0
	ret
	.data
	.lcomm	_c_0.5944,16,2		# c_0.5944



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40971 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 19:05:49 +00:00
Evan Cheng
a333b41af9 GR16_ sub-register class should be GR8_, not GR8. That is, it should only be 8-bit registers in 32-bit mode. Ditto for GR32_.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40970 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 18:05:17 +00:00
Dale Johannesen
ab081c7195 Update per review comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40965 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 17:27:48 +00:00
Owen Anderson
9528f11481 Make NonLocal and None const in the right way. :-)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40961 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 04:42:44 +00:00
Devang Patel
c9d123dca9 Traverse loop blocks' terminators to find split candidates.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40960 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 01:39:01 +00:00
Dale Johannesen
5411a3937f long double 9 of N. This finishes up the X86-32 bits
(constants are still not handled).  Adds ConvertActions
to control fp-to-fp conversions (these are currently
defaulted for all other targets, so no changes there).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40958 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 01:04:01 +00:00
Dale Johannesen
48bd15ed72 Fix arguments for some Altivec instructions. From SWB.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40957 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-09 00:49:19 +00:00
Scott Michel
c1513d24ea If a target really needs to custom lower constants, it should be allowed
to do so.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40955 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 23:23:31 +00:00
Owen Anderson
642a9e3436 Add more comments to memdep.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40953 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 22:26:03 +00:00
Devang Patel
9704fcf505 Add cost analysis.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40952 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 22:25:28 +00:00
Owen Anderson
80b1f09693 Make memdep fit in 80 cols.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40950 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 22:01:54 +00:00
Owen Anderson
49609a5ddf Add one more comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40949 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 21:54:33 +00:00
Owen Anderson
179463c0d4 Cleanup and comment-ize the memdep header.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40948 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 21:53:20 +00:00
Devang Patel
787a713bb2 Preserve dom info while processing one iteration loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40947 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 21:39:47 +00:00
Owen Anderson
8cad423d94 Change the None and NonLocal markers in memdep to be const.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40946 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 21:39:39 +00:00
Reid Spencer
d2887e4dcd Separate program name from error message with a :
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40945 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 21:19:01 +00:00
Devang Patel
9021c7001e Clear split info.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40944 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 21:18:27 +00:00
Reid Spencer
c52ca08c56 Terminate an error message with a newline.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40943 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 21:17:59 +00:00
Devang Patel
71554b8e3c Handle multiple split conditions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40941 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 21:02:17 +00:00
Reid Spencer
eb9422b19d Make it clear that getSTDIN returns null when stdin is empty.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40940 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 20:02:20 +00:00
Reid Spencer
2372ccc111 Make getSTDIN return null if the standard input is empty, as the header file
documentation implies and as its uses depend.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40939 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 20:01:58 +00:00
Reid Spencer
53424ad141 Allow the filename "-" to be a place holder for stdin. This allows directing
stdin through llvm-ld and llvm-link.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40938 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 19:52:29 +00:00
Owen Anderson
838014ee5b Global values also don't undead-ify pointers in our dead alloca's set.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40936 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 19:12:31 +00:00
Owen Anderson
362bb52108 Make handleEndBlock significantly faster with one trivial improvement,
and one hack to avoid hitting a bad case when the alias analysis is imprecise.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40935 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 18:38:28 +00:00
Owen Anderson
df359c264e Small improvement: if a function doesn't access memory, we don't need to scan
it for potentially undeading pointers.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40933 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 17:58:56 +00:00
Owen Anderson
bb3abf41a8 Add some comments, remove a dead argument, and simplify some control flow.
No functionality change.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40932 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 17:50:09 +00:00
Chris Lattner
79e038a16d eliminate redundant conditions from the signless types conversion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40927 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 16:19:57 +00:00
Chris Lattner
a099b6c7bb Handle functions with no name better.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40926 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 16:07:23 +00:00
Evan Cheng
24c2e5cf7e Adding kill info to val#.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40925 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 07:03:29 +00:00
Chris Lattner
c5f6a1f9d6 significantly speed up constant folding of calls (and thus all clients that use
ConstantFoldInstruction on calls) by avoiding Value::getName().  getName() constructs
and returns an std::string, which does heap allocation stuff.  This slightly speeds up
instcombine.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40924 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 06:55:43 +00:00
Chris Lattner
de0e42d3c0 Speed up updateDFSNumbers with two observations:
1. domtree is a tree, not a graph.  There is no need to avoid revisiting nodes with a set.
2. the worklist can contain the child iterator pointers so we don't get N^2 rescanning of children.

This speeds up updateDFSNumbers significantly, making it basically free.  On the testcase in PR1432,
this speeds up loopsimplify by another 3x, dropping it from the 12th most expensive pass to the to
the 30th. :)  It used to be #1.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40923 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 06:24:20 +00:00
Owen Anderson
c182f17aed A few more small cleanups.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40922 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 06:06:02 +00:00
Evan Cheng
a141cfe5b2 Clean up and bug fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40921 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 05:56:18 +00:00
Chris Lattner
3e089ae0b8 reimplement dfs number computation to be significantly faster. This speeds up
natural loop canonicalization (which does many cfg xforms) by 4.3x, for 
example.  This also fixes a bug in postdom dfnumber computation.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40920 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 05:51:24 +00:00
Owen Anderson
6ca4cb3755 First round of cleanups from Chris' feedback.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40919 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 04:52:29 +00:00
Evan Cheng
8df786012d - Each val# can have multiple kills.
- Fix some minor bugs related to special markers on val# def. ~0U means
  undefined, ~1U means dead val#.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40916 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 03:00:28 +00:00
Devang Patel
6a2bfdaab6 Embrace patch review feedback.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40915 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 01:51:27 +00:00
Evan Cheng
51cbf3c980 Remove a dead assertion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40914 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-08 01:00:21 +00:00
Evan Cheng
a8d94f1315 - LiveInterval value#'s now have 3 components: def instruction #,
kill instruction #, and source register number (iff the value# is defined by a
copy).
- Now def instruction # is set for every value#, not just for copy defined ones.
- Update some outdated code related inactive live ranges.
- Kill info not yet set. That's next patch.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40913 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-07 23:49:57 +00:00
Devang Patel
002fe25dd7 Fix new compare instruction's signness. Caught by Chris during review.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40912 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-07 23:17:52 +00:00
Devang Patel
a9e9286630 Fix comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40911 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-07 23:16:03 +00:00
Owen Anderson
a37226af81 Don't insert nearly as many redundant phi nodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40909 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-07 23:12:31 +00:00
Dale Johannesen
84109cd015 Fix spelling of mtvscr and mfvscr.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40908 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-07 23:08:00 +00:00
Dale Johannesen
cdbe4d3ebe Long double patch 8 of N: make it partially work in
SSE mode (all but conversions <-> other FP types, I think):
>>Do not mark all-80-bit operations as "Requires[FPStack]"
(which really means "not SSE").
>>Refactor load-and-extend to facilitate this.
>>Update comments.
>>Handle long double in SSE when computing FP_REG_KILL.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40906 91177308-0d34-0410-b5e6-96231b3b80d8
2007-08-07 20:29:26 +00:00