Commit Graph

30758 Commits

Author SHA1 Message Date
Evan Cheng
04677a3b49 Recognize target instruction flag 'isReMaterializable'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35159 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 06:20:37 +00:00
Evan Cheng
5d5c93f659 Add a TargetInstrDescriptor flag to mark an instruction as "re-materializable".
It means the instruction can be easily re-materialized at any point. e.g.
constant generation, load from constantpool.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35158 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 06:19:16 +00:00
Zhou Sheng
fdcd5a7740 Add isStrictPositive() to APInt to determine if this APInt Value > 0.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35156 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 05:22:18 +00:00
Evan Cheng
79b098445b Minor bug fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35153 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 04:22:35 +00:00
Chris Lattner
d435dbcbf9 fix a warning
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35152 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 00:39:32 +00:00
Chris Lattner
12779797f7 add PR#
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35151 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 00:17:19 +00:00
Chris Lattner
8bf9911932 implement the next chunk of SROA with memset/memcpy's of aggregates. This
implements Transforms/ScalarRepl/memset-aggregate-byte-leader.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35150 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 00:16:43 +00:00
Chris Lattner
fb73d9cf0c add pr#
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35149 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 00:15:43 +00:00
Chris Lattner
7863a63bb9 new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35148 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 00:11:30 +00:00
Chris Lattner
7829828c4b testcase for SROA with memset etc
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35147 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-19 00:09:00 +00:00
Nick Lewycky
3f64b1ab1d Clean up this code and fix subtract miscompile.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35146 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-18 22:58:46 +00:00
Chris Lattner
318bf7963a Implement InstCombine/and-xor-merge.ll:test[12].
Rearrange some code to simplify it now that shifts are binops


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35145 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-18 22:51:34 +00:00
Chris Lattner
296d7a778a new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35144 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-18 22:50:57 +00:00
Chris Lattner
11a3a9d27d minor updates
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35143 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-18 22:41:33 +00:00
Nick Lewycky
7f4ba44f37 This is implemented. We now generate:
entry:
        icmp ugt i32 %x, 4              ; <i1>:0 [#uses=1]
        br i1 %0, label %cond_true, label %cond_false

cond_true:              ; preds = %entry
        %tmp1 = tail call i32 (...)* @bar( i32 12 )             ; <i32> [#uses=0]
        ret void

cond_false:             ; preds = %entry
        switch i32 %x, label %cond_true15 [
                 i32 4, label %cond_true3
                 i32 3, label %cond_true7
                 i32 2, label %cond_true11
                 i32 0, label %cond_false17
        ]

...


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35142 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-18 14:37:20 +00:00
Evan Cheng
cf596c54d4 - Merge UsedBlocks info after two virtual registers are coalesced.
- Use distance to closest use to determine whether to abort coalescing.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35141 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-18 09:05:55 +00:00
Evan Cheng
e951cd1647 Keep UsedBlocks info accurate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35140 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-18 09:02:31 +00:00
Evan Cheng
7e79e396a4 Fix comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35139 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-18 03:26:04 +00:00
Nick Lewycky
1eda0f60d7 Propagate ValueRanges across equality.
Add some more micro-optimizations: x * 0 = 0, a - x = a --> x = 0.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35138 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-18 01:09:32 +00:00
Anton Korobeynikov
1cc6452181 Silence warning
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35137 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-17 14:48:06 +00:00
Evan Cheng
c6a2410d58 Track the BB's where each virtual register is used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35135 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-17 09:29:54 +00:00
Evan Cheng
ba1a3df608 Joining a live interval of a physical register with a virtual one can turn out
to be really bad. Once they are joined they are not broken apart. Also, physical
intervals cannot be spilled!

Added a heuristic as a workaround for this. Be careful coalescing with a
physical register if the virtual register uses are "far". Check if there are
uses in the same loop as the source (copy instruction). Check if it is in the
loop preheader, etc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35134 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-17 09:27:35 +00:00
Evan Cheng
db264ae07f Use SmallSet instead of std::set.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35133 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-17 08:53:30 +00:00
Evan Cheng
2edd563825 If sdisel has decided to sink GEP index expression into any BB. Replace all uses
in that BB.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35132 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-17 08:22:49 +00:00
Evan Cheng
ade0162fdf GEP index sink test case.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35131 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-17 03:18:32 +00:00
Devang Patel
0e7039a5a8 Test case for X86 inline asm constraint 'I'
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35130 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-17 00:14:52 +00:00
Devang Patel
84f7fd2483 Support 'I' inline asm constraint.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35129 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-17 00:13:28 +00:00
Lauro Ramos Venancio
368f20fda4 Only ARMv6 has BSWAP.
Fix MultiSource/Applications/aha test.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35128 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-16 22:54:16 +00:00
Evan Cheng
9e2d86add8 Turn on GEP index sinking by default.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35127 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-16 18:32:30 +00:00
Evan Cheng
9f5ead9601 Stupid bug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35126 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-16 17:50:20 +00:00
Bill Wendling
1b7a81d3ae And now support for MMX logical operations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35125 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-16 09:44:46 +00:00
Evan Cheng
3cd4e5095b Sink a binary expression into its use blocks if it is a loop invariant
computation used as GEP indexes and if the expression can be folded into
target addressing mode of GEP load / store use types.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35123 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-16 08:46:27 +00:00
Evan Cheng
2770747216 Added isLegalAddressExpression(). Only allows X +/- C for now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35122 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-16 08:43:56 +00:00
Evan Cheng
14245a9d62 Added isLegalAddressExpression hook to test if the given expression can be
folded into target addressing mode for the given type.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35121 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-16 08:42:32 +00:00
Evan Cheng
2864eee7a3 These forward declarations are not needed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35120 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-16 08:41:06 +00:00
Nick Lewycky
4c70875774 Add more comments and update to new asm syntax.
Add new micro-optimizations.

Add icmp predicate snuggling. Given %x ULT 4, "icmp ugt %x, 2" becomes
"icmp eq %x, 3". This doesn't apply in any non-trivial cases yet due to missing
support for NE values in ValueRanges.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35119 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-16 02:37:39 +00:00
Bill Wendling
74027e98f1 Multiplication support for MMX.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35118 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-15 21:24:36 +00:00
Evan Cheng
56fdd7af88 Debugging output stuff.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35117 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-15 21:19:28 +00:00
Reid Spencer
5eb77c7674 Regenerate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35116 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-15 03:26:42 +00:00
Reid Spencer
44f87ee746 Revert last changes as they introduced other problems.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35115 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-15 03:25:34 +00:00
Evan Cheng
8f7d26bce7 ARM isel should match ldr x +/- x * (2^n) to ldr [x, +/- x, lsl #log2(n)].
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35114 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 23:26:40 +00:00
Reid Spencer
3fae7ba5b5 Regenerate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35113 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 23:13:06 +00:00
Reid Spencer
be5c50cb33 The sign information was not propagating into the rename map so only the
last entry stored in the map could be retrieved for a given integer type.
Propagating the sign information required an invasive change to ensure that
all ValueRef (ValID) instances get the right sign information as well. Also,
put in some assertions to ensure the RenameMap always gives us out the type
that is expected.

This fixes PR1256 and
test/Assembler/2007-03-14-UgpradeLocalSignless.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35112 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 23:11:45 +00:00
Reid Spencer
c1882215a0 For PR1256:
Carry sign with ValID and make TypeInfo sortable (useful in a map).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35111 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 23:08:04 +00:00
Reid Spencer
8ee16fbc73 Test case for PR1256.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35110 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 23:07:24 +00:00
Evan Cheng
c6deb3d447 Estimate a cost using the possible number of scratch registers required and use
it as a late BURR scheduling tie-breaker.
Intuitively, it's good to push down instructions whose results are liveout so
their long live ranges won't conflict with other values which are needed inside
the BB. Further prioritize liveout instructions by the number of operands which
are calculated within the BB.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35109 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 22:43:40 +00:00
Evan Cheng
ba693005e9 Under X86-64 large code model, do not emit 32-bit pc relative calls.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35108 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 22:11:11 +00:00
Evan Cheng
4485d3897b Notes about codegen issues.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35107 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 21:03:53 +00:00
Duncan Sands
edd174bcca Test that the size of a view converted object is determined by the target
type, not the source type.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35106 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 20:24:53 +00:00
Evan Cheng
e70ef98043 Clean up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35105 91177308-0d34-0410-b5e6-96231b3b80d8
2007-03-14 20:20:19 +00:00