Chris Lattner
bbf1c72d51
implement TII::insertNoop
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26562 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 23:49:55 +00:00
Chris Lattner
6cc3f0a774
Implement G5HazardRecognizer as a trivial thing that wants 5 cycles between
...
copyfromreg nodes. Clearly useful!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26559 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 23:13:56 +00:00
Chris Lattner
ad0f78ace2
Add basic hazard recognizer support. noop insertion isn't complete yet though.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26558 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 22:45:01 +00:00
Jeff Cohen
3d68e15c20
Fix VC++ compilation error.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26554 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 21:43:37 +00:00
Chris Lattner
a5de484bc7
Split the list scheduler into top-down and bottom-up pieces. The priority
...
function of the top-down scheduler are completely bogus currently, and
having (future) PPC specific in this file is also wrong, but this is a
small incremental step.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26552 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 21:10:33 +00:00
Chris Lattner
41f5ea0665
Move the available queue to being inside the ListSchedule method, since it
...
bounds its lifetime.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26550 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 20:21:55 +00:00
Chris Lattner
5b0fe7d91d
add a note
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26549 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 20:00:08 +00:00
Chris Lattner
cac7059d0f
Reinstate this now that the offending opposite xform has been removed.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26548 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 19:53:55 +00:00
Chris Lattner
c1d9f1de41
Do not fold (add (shl x, c1), (shl c2, c1)) -> (shl (add x, c2), c1),
...
we want to canonicalize the other way.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26547 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 19:52:57 +00:00
Chris Lattner
5267651057
print arbitrary constant pool entries
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26545 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 09:38:03 +00:00
Evan Cheng
9176b0169c
Back out fold (shl (add x, c1), c2) -> (add (shl x, c2), c1<<c2) for now.
...
It's causing an infinite loop compiling ldecod on x86 / Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26544 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 07:30:16 +00:00
Chris Lattner
12d830346b
Add some simple copysign folds
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26543 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 05:30:57 +00:00
Chris Lattner
a09f848c11
Codegen copysign[f] into a FCOPYSIGN node
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26542 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 05:09:38 +00:00
Chris Lattner
9601a86a64
Copysign needs to be expanded everywhere. Note that Alpha and IA64 should
...
implement copysign as a native op if they have it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26541 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 05:08:37 +00:00
Chris Lattner
a4929df2da
add a note for something evan noticed
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26539 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 01:15:18 +00:00
Chris Lattner
220b0cf3e4
Make vector narrowing more effective, implementing
...
Transforms/InstCombine/vec_narrow.ll. This add support for narrowing
extract_element(insertelement) also.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26538 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-05 00:22:33 +00:00
Chris Lattner
9f17be690e
Implemented.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26536 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 23:33:44 +00:00
Chris Lattner
a1deca3cd6
fold (mul (add x, c1), c2) -> (add (mul x, c2), c1*c2)
...
fold (shl (add x, c1), c2) -> (add (shl x, c2), c1<<c2)
This allows us to compile CodeGen/PowerPC/addi-reassoc.ll into:
_test1:
slwi r2, r4, 4
add r2, r2, r3
lwz r3, 36(r2)
blr
_test2:
mulli r2, r4, 5
add r2, r2, r3
lbz r2, 11(r2)
extsb r3, r2
blr
instead of:
_test1:
addi r2, r4, 2
slwi r2, r2, 4
add r2, r3, r2
lwz r3, 4(r2)
blr
_test2:
addi r2, r4, 2
mulli r2, r2, 5
add r2, r3, r2
lbz r2, 1(r2)
extsb r3, r2
blr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26535 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 23:33:26 +00:00
Chris Lattner
2cfdd2854c
Fix a crash compiling Obsequi
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26529 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 21:48:01 +00:00
Chris Lattner
e5022fe4cd
Add factoring of multiplications, e.g. turning A*A+A*B into A*(A+B).
...
Testcase here: Transforms/Reassociate/mulfactor.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26524 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 09:31:13 +00:00
Chris Lattner
ad01993194
Add a note
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26523 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 08:44:51 +00:00
Chris Lattner
c2c6038913
Regenerate
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26522 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 07:53:41 +00:00
Chris Lattner
7d5c1e1ed1
Don't use invalidated iterators!
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26521 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 07:53:16 +00:00
Evan Cheng
f42f516984
Add an entry
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26520 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 07:49:50 +00:00
Chris Lattner
ab51f3fa96
Canonicalize (X+C1)*C2 -> X*C2+C1*C2
...
This implements Transforms/InstCombine/add.ll:test31
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26519 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 06:04:02 +00:00
Evan Cheng
62bec2ca4c
MEMSET / MEMCPY lowering bugs: we can't issue a single WORD / DWORD version of
...
rep/stos and rep/mov if the count is not a constant. We could do
rep/stosl; and $count, 3; rep/stosb
For now, I will lower them to memset / memcpy calls. We will revisit this after
a little bit experiment.
Also need to take care of the trailing bytes even if the count is a constant.
Since the max. number of trailing bytes are 3, we will simply issue loads /
stores.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26517 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 02:48:56 +00:00
Chris Lattner
5a3cf8de5d
Be more conservative with our symbolic alias analysis. In particular,
...
don't assume that A[1][0] and A[0][i] can't alias. "i" might be out of
range, or even negative. This fixes a miscompilation of 188.ammp (which
does bad pointer tricks) with the new CFE.
Testcase here: Analysis/BasicAA/2006-03-03-BadArraySubscript.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26515 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 02:06:34 +00:00
Chris Lattner
b27b69f283
add a note
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26513 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 01:19:34 +00:00
Evan Cheng
8df346b4e8
Typo
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26512 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-04 01:12:00 +00:00
Jim Laskey
b80af6f572
Added support for dwarf block data entries.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26509 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 21:00:14 +00:00
Chris Lattner
329b951b69
Unbreak autouprade of llvm.sqrt, simplify some code.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26506 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 16:31:22 +00:00
Jim Laskey
f01e547700
Adding basic structure support.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26505 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 15:06:57 +00:00
Evan Cheng
3e1ce5a44d
Add more vector NodeTypes: VSDIV, VUDIV, VAND, VOR, and VXOR.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26504 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 07:01:07 +00:00
Evan Cheng
33143dce15
Number of NodeTypes now exceeds 128.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26503 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 06:58:59 +00:00
Evan Cheng
bfa284f697
SDOperand::isOperand should not be a forwarding. It must check *this against N's operands.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26502 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 06:42:32 +00:00
Evan Cheng
80d8eaae05
Added isOperand(N): true if this is an operand of N
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26501 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 06:24:54 +00:00
Evan Cheng
fb43331226
A bit more tweaking
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26500 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 06:23:43 +00:00
Jeff Cohen
d84525480a
Fix VC++ compilation errors.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26498 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 03:25:07 +00:00
Evan Cheng
932f022b82
Move #include "llvm/CodeGen/AsmPrinter.h" to top since it's the interface.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26493 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 02:04:29 +00:00
Chris Lattner
a8309ae1cb
Split the valuetypes out of Target.td into ValueTypes.td
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26490 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 01:55:26 +00:00
Chris Lattner
21959390c1
Change this to work with renamed intrinsics.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26484 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 01:34:17 +00:00
Chris Lattner
aecb0627fa
Make this work with renamed intrinsics.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26482 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 01:30:23 +00:00
Chris Lattner
01ac91e2fd
updates for recent changes
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26481 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 01:21:36 +00:00
Chris Lattner
41edaa0529
remove the read/write port/io intrinsics.
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26479 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 00:19:58 +00:00
Chris Lattner
03dd465215
Split memcpy/memset/memmove intrinsics into i32/i64 versions, resolving
...
PR709, and paving the way for future progress.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26476 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-03 00:00:25 +00:00
Chris Lattner
bad08000f1
Simplify the autoupgrade interface
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26475 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-02 23:59:12 +00:00
Chris Lattner
ffa987d3ee
Split memcpy/memset/memmove intrinsics into i32/i64 versions, resolving
...
PR709, and paving the way for future progress.
Significantly refactor autoupgrading code, to handle the more complex case
(where we upgrade one argument in a function), and fix some bugs in it.
Testcase here: llvm/test/Regression/Bytecode/memcpy.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26474 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-02 23:58:40 +00:00
Chris Lattner
89188a1096
add a note
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26472 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-02 22:34:38 +00:00
Evan Cheng
86ec7d1d07
- Fixed some priority calculation bugs that were causing bug 478. Among them:
...
a predecessor appearing more than once in the operand list was counted as
multiple predecessor; priority1 should be updated during scheduling;
CycleBound was updated after the node is inserted into priority queue; one
of the tie breaking condition was flipped.
- Take into consideration of two address opcodes. If a predecessor is a def&use
operand, it should have a higher priority.
- Scheduler should also favor floaters, i.e. nodes that do not have real
predecessors such as MOV32ri.
- The scheduling fixes / tweaks fixed bug 478:
.text
.align 4
.globl _f
_f:
movl 4(%esp), %eax
movl 8(%esp), %ecx
movl %eax, %edx
imull %ecx, %edx
imull %eax, %eax
imull %ecx, %ecx
addl %eax, %ecx
leal (%ecx,%edx,2), %eax
ret
It is also a slight performance win (1% - 3%) for most tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26470 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-02 21:38:29 +00:00
Chris Lattner
db3f873bd8
Generalize the REM folding code to handle another case Nick Lewycky
...
pointed out: realize the AND can provide factors and look through Casts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26469 91177308-0d34-0410-b5e6-96231b3b80d8
2006-03-02 06:50:58 +00:00