Scott Michel [Tue, 2 Dec 2008 19:53:53 +0000 (19:53 +0000)]
CellSPU:
- Incorporate Tilmann Scheller's ISD::TRUNCATE custom lowering patch
- Update SPU calling convention info, even if it's not used yet (but can be
at some point or another)
- Ensure that any-extended f32 loads are custom lowered, especially when
they're promoted for use in printf.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60438
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 2 Dec 2008 19:27:20 +0000 (19:27 +0000)]
Fix a typo in a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60434
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 2 Dec 2008 18:53:47 +0000 (18:53 +0000)]
Add support for folding spills into preceding defs when doing pre-alloc splitting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60433
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Tue, 2 Dec 2008 18:40:40 +0000 (18:40 +0000)]
One more transformation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60432
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Tue, 2 Dec 2008 18:40:09 +0000 (18:40 +0000)]
Make the code do what the comment says it does.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60431
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 2 Dec 2008 18:33:11 +0000 (18:33 +0000)]
Comment typeo fix, thanks Duncan!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60429
91177308-0d34-0410-b5e6-
96231b3b80d8
Tilmann Scheller [Tue, 2 Dec 2008 12:12:25 +0000 (12:12 +0000)]
make it possible to custom lower TRUNCATE (needed for the CellSPU target)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60409
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 2 Dec 2008 08:16:11 +0000 (08:16 +0000)]
Implement PRE of loads in the GVN pass with a pretty cheap and
straight-forward implementation. This does not require any extra
alias analysis queries beyond what we already do for non-local loads.
Some programs really really like load PRE. For example, SPASS triggers
this ~1000 times, ~300 times in 255.vortex, and ~1500 times on 403.gcc.
The biggest limitation to the implementation is that it does not split
critical edges. This is a huge killer on many programs and should be
addressed after the initial patch is enabled by default.
The implementation of this should incidentally speed up rejection of
non-local loads because it avoids creating the repl densemap in cases
when it won't be used for fully redundant loads.
This is currently disabled by default.
Before I turn this on, I need to fix a couple of miscompilations in
the testsuite, look at compile time performance numbers, and look at
perf impact. This is pretty close to ready though.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60408
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Tue, 2 Dec 2008 08:05:48 +0000 (08:05 +0000)]
Add a new SCEV representing signed division.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60407
91177308-0d34-0410-b5e6-
96231b3b80d8
Mon P Wang [Tue, 2 Dec 2008 07:35:08 +0000 (07:35 +0000)]
Removed some unnecessary code in widening.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60406
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 2 Dec 2008 07:16:45 +0000 (07:16 +0000)]
add a little helper function that does PHI translation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60405
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 2 Dec 2008 06:32:34 +0000 (06:32 +0000)]
add a note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60404
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 2 Dec 2008 06:24:20 +0000 (06:24 +0000)]
Remove some errors that crept in. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60403
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 2 Dec 2008 06:22:04 +0000 (06:22 +0000)]
Merge two if-statements into one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60402
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 2 Dec 2008 06:18:11 +0000 (06:18 +0000)]
More styalistic changes. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60401
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 2 Dec 2008 06:08:04 +0000 (06:08 +0000)]
add densemap range insertion method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60400
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 2 Dec 2008 05:12:47 +0000 (05:12 +0000)]
- Remove the buggy -X/C -> X/-C transform. This isn't valid when X isn't a
constant. If X is a constant, then this is folded elsewhere.
- Added a note to Target/README.txt to indicate that we'd like to implement
this when we're able.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60399
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 2 Dec 2008 05:09:00 +0000 (05:09 +0000)]
Improve comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60398
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 2 Dec 2008 05:06:43 +0000 (05:06 +0000)]
- Reduce nesting.
- No need to do a swap on a canonicalized pattern.
No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60397
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 2 Dec 2008 04:52:26 +0000 (04:52 +0000)]
some random comment improvements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60395
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 2 Dec 2008 04:25:42 +0000 (04:25 +0000)]
Add a test for my previous PRE fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60394
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 2 Dec 2008 04:09:22 +0000 (04:09 +0000)]
Fix an issue that Chris noticed, where local PRE was not properly instantiating
a new value numbering set after splitting a critical edge. This increases
the number of instances of PRE on 403.gcc from ~60 to ~570.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60393
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 2 Dec 2008 02:15:36 +0000 (02:15 +0000)]
Fix PR3124: overly strict assert.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60392
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Tue, 2 Dec 2008 01:30:54 +0000 (01:30 +0000)]
Add a few more transformations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60391
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 2 Dec 2008 01:06:39 +0000 (01:06 +0000)]
Second stab at target-dependent lowering of everyone's favorite nodes: [SU]ADDO
- LowerXADDO lowers [SU]ADDO into an ADD with an implicit EFLAGS define. The
EFLAGS are fed into a SETCC node which has the conditional COND_O or COND_C,
depending on the type of ADDO requested.
- LowerBRCOND now recognizes if it's coming from a SETCC node with COND_O or
COND_C set.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60388
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 2 Dec 2008 00:07:05 +0000 (00:07 +0000)]
Reapply r60382. This time, don't mark "ADC" nodes with "implicit EFLAGS".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60385
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 23:44:08 +0000 (23:44 +0000)]
Temporarily revert r60382. It caused CodeGen/X86/i2k.ll and others to fail.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60383
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 23:30:42 +0000 (23:30 +0000)]
- Have "ADD" instructions return an implicit EFLAGS.
- Add support for seto, setno, setc, and setnc instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60382
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 23:28:22 +0000 (23:28 +0000)]
Expand getVTList, getNodeValueTypes, and SelectNodeTo to handle more value types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60381
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 22:35:31 +0000 (22:35 +0000)]
Add rdar reference, make this actually fail when the patch isn't applied.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60376
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Mon, 1 Dec 2008 22:00:01 +0000 (22:00 +0000)]
Consider only references to an IV within the loop when
figuring out the base of the IV. This produces better
code in the example. (Addresses use (IV) instead of
(BASE,IV) - a significant improvement on low-register
machines like x86).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60374
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 21:11:25 +0000 (21:11 +0000)]
reenable array_pod_sort, this time hopefully happy on 64-bit
and big endian systems.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60371
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 21:06:30 +0000 (21:06 +0000)]
Don't rebuild RHSNeg. Just use the one that's already there.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60370
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 21:03:43 +0000 (21:03 +0000)]
Document what this check is doing. Also, no need to cast to ConstantInt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60369
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 19:46:27 +0000 (19:46 +0000)]
Use a simple comparison. Overflow on integer negation can only occur when the
integer is "minint".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60366
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 19:45:45 +0000 (19:45 +0000)]
don't #include <algorithm> into the llvm namespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60365
91177308-0d34-0410-b5e6-
96231b3b80d8
Scott Michel [Mon, 1 Dec 2008 17:56:02 +0000 (17:56 +0000)]
CellSPU:
- Fix v2[if]64 vector insertion code before IBM files a bug report.
- Ensure that zero (0) offsets relative to $sp don't trip an assert
(add $sp, 0 gets legalized to $sp alone, tripping an assert)
- Shuffle masks passed to SPUISD::SHUFB are now v16i8 or v4i32
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60358
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 17:00:08 +0000 (17:00 +0000)]
switch to std::sort until I have time to sort this out.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60354
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 16:55:19 +0000 (16:55 +0000)]
cleanups suggested by duncan, thanks!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60353
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 16:50:01 +0000 (16:50 +0000)]
define array_pod_sort in terms of operator< instead of my brain
damaged approximation. This should fix it on big endian platforms
and on 64-bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60352
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Mon, 1 Dec 2008 11:41:29 +0000 (11:41 +0000)]
There are no longer any places that require a
MERGE_VALUES node with only one operand, so get
rid of special code that only existed to handle
that possibility.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60349
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Mon, 1 Dec 2008 11:39:25 +0000 (11:39 +0000)]
Change the interface to the type legalization method
ReplaceNodeResults: rather than returning a node which
must have the same number of results as the original
node (which means mucking around with MERGE_VALUES,
and which is also easy to get wrong since SelectionDAG
folding may mean you don't get the node you expect),
return the results in a vector.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60348
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 08:32:40 +0000 (08:32 +0000)]
Generalize the FoldOrWithConstant method to fold for any two constants which
don't have overlapping bits.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60344
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 08:23:25 +0000 (08:23 +0000)]
Reduce copy-and-paste code by splitting out the code into its own function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60343
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 08:09:47 +0000 (08:09 +0000)]
Use m_Specific() instead of double matching.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60341
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 07:47:02 +0000 (07:47 +0000)]
Move pattern check outside of the if-then statement. This prevents us from fiddling with constants unless we have to.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60340
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 07:35:54 +0000 (07:35 +0000)]
Rename some variables, only increment BI once at the start of the loop instead of throughout it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60339
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 07:29:03 +0000 (07:29 +0000)]
pull the predMap densemap out of the inner loop of performPRE, so
that it isn't reallocated all the time. This is a tiny speedup for
GVN: 3.90->3.88s
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60338
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 06:52:57 +0000 (06:52 +0000)]
switch a couple more calls to use array_pod_sort.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60337
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 06:50:46 +0000 (06:50 +0000)]
don't assume iterators implicitly convert to pointers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60336
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 06:49:59 +0000 (06:49 +0000)]
Introduce a new array_pod_sort function and switch LSR to use it
instead of std::sort. This shrinks the release-asserts LSR.o file
by 1100 bytes of code on my system.
We should start using array_pod_sort where possible.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60335
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 06:27:41 +0000 (06:27 +0000)]
Eliminate use of setvector for the DeadInsts set, just use a smallvector.
This is a lot cheaper and conceptually simpler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60332
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 06:14:28 +0000 (06:14 +0000)]
DeleteTriviallyDeadInstructions is always passed the
DeadInsts ivar, just use it directly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60330
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 06:11:32 +0000 (06:11 +0000)]
simplify DeleteTriviallyDeadInstructions again, unlike my previous
buggy rewrite, this notifies ScalarEvolution of a pending instruction
about to be removed and then erases it, instead of erasing it then
notifying.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60329
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 05:16:26 +0000 (05:16 +0000)]
simplify these patterns using m_Specific. No need to grep for
xor in testcase (or is a substring).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60328
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 04:48:07 +0000 (04:48 +0000)]
Teach jump threading to clean up after itself, DCE and constfolding the
new instructions it simplifies. Because we're threading jumps on edges
with constants coming in from PHI's, we inherently are exposing a lot more
constants to the new block. Folding them and deleting dead conditions
allows the cost model in jump threading to be more accurate as it iterates.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60327
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 03:58:38 +0000 (03:58 +0000)]
The PreVerifier pass preserves everything. In practice, this
prevents the passmgr from adding yet-another domtree invocation
for Verifier if there is already one live.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60326
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 03:42:51 +0000 (03:42 +0000)]
Change instcombine to use FoldPHIArgGEPIntoPHI to fold two operand PHIs
instead of using FoldPHIArgBinOpIntoPHI. In addition to being more
obvious, this also fixes a problem where instcombine wouldn't merge two
phis that had different variable indices. This prevented instcombine
from factoring big chunks of code in 403.gcc. For example:
insn_cuid.exit:
- %tmp336 = load i32** @uid_cuid, align 4
- %tmp337 = getelementptr %struct.rtx_def* %insn_addr.0.ph.i, i32 0, i32 3
- %tmp338 = bitcast [1 x %struct.rtunion]* %tmp337 to i32*
- %tmp339 = load i32* %tmp338, align 4
- %tmp340 = getelementptr i32* %tmp336, i32 %tmp339
br label %bb62
bb61:
- %tmp341 = load i32** @uid_cuid, align 4
- %tmp342 = getelementptr %struct.rtx_def* %insn, i32 0, i32 3
- %tmp343 = bitcast [1 x %struct.rtunion]* %tmp342 to i32*
- %tmp344 = load i32* %tmp343, align 4
- %tmp345 = getelementptr i32* %tmp341, i32 %tmp344
br label %bb62
bb62:
- %iftmp.62.0.in = phi i32* [ %tmp345, %bb61 ], [ %tmp340, %insn_cuid.exit ]
+ %insn.pn2 = phi %struct.rtx_def* [ %insn, %bb61 ], [ %insn_addr.0.ph.i, %insn_cuid.exit ]
+ %tmp344.pn.in.in = getelementptr %struct.rtx_def* %insn.pn2, i32 0, i32 3
+ %tmp344.pn.in = bitcast [1 x %struct.rtunion]* %tmp344.pn.in.in to i32*
+ %tmp341.pn = load i32** @uid_cuid
+ %tmp344.pn = load i32* %tmp344.pn.in
+ %iftmp.62.0.in = getelementptr i32* %tmp341.pn, i32 %tmp344.pn
%iftmp.62.0 = load i32* %iftmp.62.0.in
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60325
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 02:34:36 +0000 (02:34 +0000)]
Teach inst combine to merge GEPs through PHIs. This is really
important because it is sinking the loads using the GEPs, but
not the GEPs themselves. This triggers 647 times on 403.gcc
and makes the .s file much much nicer. For example before:
je LBB1_87 ## bb78
LBB1_62: ## bb77
leal 84(%esi), %eax
LBB1_63: ## bb79
movl (%eax), %eax
...
LBB1_87: ## bb78
movl $0, 4(%esp)
movl %esi, (%esp)
call L_make_decl_rtl$stub
jmp LBB1_62 ## bb77
after:
jne LBB1_63 ## bb79
LBB1_62: ## bb78
movl $0, 4(%esp)
movl %esi, (%esp)
call L_make_decl_rtl$stub
LBB1_63: ## bb79
movl 84(%esi), %eax
The input code was (and the GEPs are merged and
the PHI is now eliminated by instcombine):
br i1 %tmp233, label %bb78, label %bb77
bb77:
%tmp234 = getelementptr %struct.tree_node* %t_addr.3, i32 0, i32 0, i32 22
br label %bb79
bb78:
call void @make_decl_rtl(%struct.tree_node* %t_addr.3, i8* null) nounwind
%tmp235 = getelementptr %struct.tree_node* %t_addr.3, i32 0, i32 0, i32 22
br label %bb79
bb79:
%iftmp.12.0.in = phi %struct.rtx_def** [ %tmp235, %bb78 ], [ %tmp234, %bb77 ]
%iftmp.12.0 = load %struct.rtx_def** %iftmp.12.0.in
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60322
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 01:42:03 +0000 (01:42 +0000)]
testcase for my previous commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60315
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 01:31:36 +0000 (01:31 +0000)]
Make GVN be more intelligent about redundant load
elimination: when finding dependent load/stores, realize that
they are the same if aliasing claims must alias instead of relying
on the pointers to be exactly equal. This makes load elimination
more aggressive. For example, on 403.gcc, we had:
< 68 gvn - Number of instructions PRE'd
< 152718 gvn - Number of instructions deleted
< 49699 gvn - Number of loads deleted
< 6153 memdep - Number of dirty cached non-local responses
< 169336 memdep - Number of fully cached non-local responses
< 162428 memdep - Number of uncached non-local responses
now we have:
> 64 gvn - Number of instructions PRE'd
> 153623 gvn - Number of instructions deleted
> 49856 gvn - Number of loads deleted
> 5022 memdep - Number of dirty cached non-local responses
> 159030 memdep - Number of fully cached non-local responses
> 162443 memdep - Number of uncached non-local responses
That's an extra 157 loads deleted and extra 905 other instructions nuked.
This slows down GVN very slightly, from 3.91 to 3.96s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60314
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 01:15:42 +0000 (01:15 +0000)]
Reimplement the non-local dependency data structure in terms of a sorted
vector instead of a densemap. This shrinks the memory usage of this thing
substantially (the high water mark) as well as making operations like
scanning it faster. This speeds up memdep slightly, gvn goes from
3.9376 to 3.9118s on 403.gcc
This also splits out the statistics for the cached non-local case to
differentiate between the dirty and clean cached case. Here's the stats
for 403.gcc:
6153 memdep - Number of dirty cached non-local responses
169336 memdep - Number of fully cached non-local responses
162428 memdep - Number of uncached non-local responses
yay for caching :)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60313
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 1 Dec 2008 01:07:11 +0000 (01:07 +0000)]
Implement ((A|B)&1)|(B&-2) -> (A&1) | B transformation. This also takes care of
permutations of this pattern.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60312
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Mon, 1 Dec 2008 00:43:48 +0000 (00:43 +0000)]
Fix bogus assertion using getSExtValue for legitimate values, like -1 in
an 128-bit-wide integer. No testcase; the issue I ran into depends on
local changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60311
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 1 Dec 2008 00:40:32 +0000 (00:40 +0000)]
Cache analyses in ivars and add some useful DEBUG output.
This speeds up GVN from 4.0386s to 3.9376s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60310
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 23:39:23 +0000 (23:39 +0000)]
improve indentation, do cheap checks before expensive ones,
remove some fixme's. This speeds up GVN very slightly on 403.gcc
(4.06->4.03s)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60309
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 23:17:19 +0000 (23:17 +0000)]
Eliminate the DepResultTy abstraction. It is now completely
redundant with MemDepResult, and MemDepResult has a nicer interface.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60308
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Sun, 30 Nov 2008 22:48:49 +0000 (22:48 +0000)]
Minor cleanup: use getTrue and getFalse where appropriate. No
functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60307
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Sun, 30 Nov 2008 21:09:11 +0000 (21:09 +0000)]
Some minor cleanups to instcombine; no functionality change.
Note that the FoldOpIntoPhi call is dead because it's impossible for the
first operand of a subtraction to be both a ConstantInt and a PHINode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60306
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 19:24:31 +0000 (19:24 +0000)]
Cache TargetData/AliasAnalysis in the pass instead of calling
getAnalysis<>. getAnalysis<> is apparently extremely expensive.
Doing this speeds up GVN on 403.gcc by 16%!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60304
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 19:10:41 +0000 (19:10 +0000)]
add the rest of the comparison routines.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60303
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 30 Nov 2008 13:52:49 +0000 (13:52 +0000)]
Add instruction combining for ((A&~B)|(~A&B)) -> A^B and all permutations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60291
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 30 Nov 2008 13:08:13 +0000 (13:08 +0000)]
Implement (A&((~A)|B)) -> A&B transformation in the instruction combiner. This
takes care of all permutations of this pattern.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60290
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 30 Nov 2008 12:41:09 +0000 (12:41 +0000)]
Forgot one remaining call to getSExtValue().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60289
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 30 Nov 2008 12:38:24 +0000 (12:38 +0000)]
getSExtValue() doesn't work for ConstantInts with bitwidth > 64 bits. Use all
APInt calls instead.
This fixes PR3144.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60288
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Sun, 30 Nov 2008 08:32:11 +0000 (08:32 +0000)]
Optimize memmove and memset into the LLVM builtins. Note that these
only show up in code from front-ends besides llvm-gcc, like clang.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60287
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Sun, 30 Nov 2008 07:52:27 +0000 (07:52 +0000)]
A couple small cleanups, plus a new potential optimization.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60286
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Sun, 30 Nov 2008 07:36:04 +0000 (07:36 +0000)]
Moving potential optimizations out of PR2330 into lib/Target/README.txt.
Hopefully this isn't too much stuff to dump into this file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60285
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Sun, 30 Nov 2008 06:35:39 +0000 (06:35 +0000)]
Followup to r60283: optimize arbitrary width signed divisions as well
as unsigned divisions. Same caveats as before.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60284
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Sun, 30 Nov 2008 06:02:26 +0000 (06:02 +0000)]
Fix for PR2164: allow transforming arbitrary-width unsigned divides into
multiplies.
Some more cleverness would be nice, though. It would be nice if we
could do this transformation on illegal types. Also, we would
prefer a narrower constant when possible so that we can use a narrower
multiply, which can be cheaper.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60283
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 30 Nov 2008 05:29:33 +0000 (05:29 +0000)]
Don't make TwoToExp signed by default.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60279
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 30 Nov 2008 05:01:05 +0000 (05:01 +0000)]
From Hacker's Delight:
"For signed integers, the determination of overflow of x*y is not so simple. If
x and y have the same sign, then overflow occurs iff xy > 2**31 - 1. If they
have opposite signs, then overflow occurs iff xy < -2**31."
In this case, x == -1.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60278
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Sun, 30 Nov 2008 04:59:26 +0000 (04:59 +0000)]
APIntify a test which is potentially unsafe otherwise, and fix the
nearby FIXME.
I'm not sure what the right way to fix the Cell test was; if the
approach I used isn't okay, please let me know.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60277
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 30 Nov 2008 04:33:53 +0000 (04:33 +0000)]
Strengthen check for div inst-combining.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60276
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 30 Nov 2008 03:42:12 +0000 (03:42 +0000)]
Instcombine was illegally transforming -X/C into X/-C when either X or C
overflowed on negation. This commit checks to make sure that neithe C nor X
overflows. This requires that the RHS of X (a subtract instruction) be a
constant integer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60275
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 02:52:26 +0000 (02:52 +0000)]
Two changes: Make getDependency remove QueryInst for a dirty record's
ReverseLocalDeps when we update it. This fixes a regression test
failure from my last commit.
Second, for each non-local cached information structure, keep a bit that
indicates whether it is dirty or not. This saves us a scan over the whole
thing in the common case when it isn't dirty.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60274
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Sun, 30 Nov 2008 02:42:05 +0000 (02:42 +0000)]
Fix a link issue I ran into trying compiling LLVM on MinGW with CMake.
Hopefully this doesn't break anyone else's build... it shouldn't unless
the MinGW variable means something other than compiling with MinGW.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60273
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 02:30:50 +0000 (02:30 +0000)]
introduce a typedef, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60272
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 02:28:25 +0000 (02:28 +0000)]
Change NonLocalDeps to be a densemap of pointers to densemap
instead of containing them by value. This increases the density
(!) of NonLocalDeps as well as making the reallocation case
faster. This speeds up gvn on 403.gcc by 2% and makes room for
future improvements.
I'm not super thrilled with having to explicitly manage the new/delete
of the map, but it is necesary for the next change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60271
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 01:44:00 +0000 (01:44 +0000)]
calls never depend on allocations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60268
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 01:39:32 +0000 (01:39 +0000)]
Fix a fixme by making memdep's handling of allocations more logical.
If we see that a load depends on the allocation of its memory with no
intervening stores, we now return a 'None' depedency instead of "Normal".
This tweaks GVN to do its optimization with the new result.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60267
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 01:26:32 +0000 (01:26 +0000)]
implement a fixme by introducing a new getDependencyFromInternal
method that returns its result as a DepResultTy instead of as a
MemDepResult. This reduces conversion back and forth.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60266
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 01:18:27 +0000 (01:18 +0000)]
Move the getNonLocalDependency method to a more logical place in
the file, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60265
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 01:17:08 +0000 (01:17 +0000)]
REmove an old fixme, resolve another fixme by adding liberal
comments about what this class does.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60264
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 01:09:30 +0000 (01:09 +0000)]
remove a bit of incorrect code that tried to be tricky about speeding up
dependencies. The basic situation was this: consider if we had:
store1
...
store2
...
store3
Where memdep thinks that store3 depends on store2 and store2 depends
on store1. The problem happens when we delete store2: The code in
question was updating dep info for store3 to be store1. This is a
spiffy optimization, but is not safe at all, because aliasing isn't
transitive. This bug isn't exposed today with DSE because DSE will only
zap store2 if it is identifical to store 3, and in this case, it is
safe to update it to depend on store1. However, memcpyopt is not so
fortunate, which is presumably why the "dropInstruction" code used to
exist.
Since this doesn't actually provide a speedup in practice, just rip the
code out.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60263
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 30 Nov 2008 00:50:20 +0000 (00:50 +0000)]
fix indentation. std::pair is "isPod" if the first/second are both isPod.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60262
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 30 Nov 2008 00:36:34 +0000 (00:36 +0000)]
Remove warning about declaration does not declare anything. This class was
already declared in the other headers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60261
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 29 Nov 2008 23:30:39 +0000 (23:30 +0000)]
Eliminate the dropInstruction method, which is not needed any more.
Fix a subtle iterator invalidation bug I introduced in the last commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60258
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sat, 29 Nov 2008 22:49:59 +0000 (22:49 +0000)]
Add protected visibility to libLTO.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60257
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 29 Nov 2008 22:02:15 +0000 (22:02 +0000)]
implement some fixme's: when deleting an instruction with
an entry in the nonlocal deps map, don't reset entries
referencing that instruction to [dirty, null], instead, set
them to [dirty,next] where next is the instruction after the
deleted one. Use this information in the non-local deps
code to avoid rescanning entire blocks.
This speeds up GVN slightly by avoiding pointless work. On
403.gcc this makes GVN 1.5% faster.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60256
91177308-0d34-0410-b5e6-
96231b3b80d8