visited set before they are used. If used, their blocks need to be
added to the visited set so that subsequent queries don't use conflicting
pointer values in the cache result blocks.
llvm-svn: 61080
computation code. Also, avoid adding output-depenency edges when both
defs are dead, which frequently happens with EFLAGS defs.
Compute Depth and Height lazily, and always in terms of edge latency
values. For the schedulers that don't care about latency, edge latencies
are set to 1.
Eliminate Cycle and CycleBound, and LatencyPriorityQueue's Latencies array.
These are all subsumed by the Depth and Height fields.
llvm-svn: 61073
alignment attribute such that 0 means unaligned.
This will probably require a rebuild of llvm-gcc because of the change to
Attributes.h. If you see many test failures on "make check", please rebuild
your llvm-gcc.
llvm-svn: 61030
CFG when there is exactly one predecessor where the load is not available.
This is designed to not increase code size but still eliminate partially
redundant loads. This fires 1765 times on 403.gcc even though it doesn't
do critical edge splitting yet (the most common reason for it to fail).
llvm-svn: 61027
cleans up the generated code a bit. This should have the added benefit of
not randomly renaming functions/globals like my previous patch did. :)
llvm-svn: 61023
memdep keeps track of how PHIs affect the pointer in dep queries, which
allows it to eliminate the load in cases like rle-phi-translate.ll, which
basically end up being:
BB1:
X = load P
br BB3
BB2:
Y = load Q
br BB3
BB3:
R = phi [P] [Q]
load R
turning "load R" into a phi of X/Y. In addition to additional exposed
opportunities, this makes memdep safe in many cases that it wasn't before
(which is required for load PRE) and also makes it substantially more
efficient. For example, consider:
bb1: // has many predecessors.
P = some_operator()
load P
In this example, previously memdep would scan all the predecessors of BB1
to see if they had something that would mustalias P. In some cases (e.g.
test/Transforms/GVN/rle-must-alias.ll) it would actually find them and end
up eliminating something. In many other cases though, it would scan and not
find anything useful. MemDep now stops at a block if the pointer is defined
in that block and cannot be phi translated to predecessors. This causes it
to miss the (rare) cases like rle-must-alias.ll, but makes it faster by not
scanning tons of stuff that is unlikely to be useful. For example, this
speeds up GVN as a whole from 3.928s to 2.448s (60%)!. IMO, scalar GVN
should be enhanced to simplify the rle-must-alias pointer base anyway, which
would allow the loads to be eliminated.
In the future, this should be enhanced to phi translate through geps and
bitcasts as well (as indicated by FIXMEs) making memdep even more powerful.
llvm-svn: 61022
callee will not introduce any new aliases of that pointer.
The attributes had all bits allocated already, so I decided to collapse
alignment. Alignment was previously stored as a 16-bit integer from bits 16 to
32 of the attribute, but it was required to be a power of 2. Now it's stored in
log2 encoded form in five bits from 16 to 21. That gives us 11 more bits of
space.
You may have already noticed that you only need four bits to encode a 16-bit
power of two, so why five bits? Because the AsmParser accepted 32-bit
alignments, even though we couldn't store them (they were silently discarded).
Now we can store them in memory, but not in the bitcode.
The bitcode format was already storing these as 64-bit VBR integers. So, the
bitcode format stays the same, keeping the alignment values stored as 16 bit
raw values. There's some hideous code in the reader and writer that deals with
this, waiting to be ripped out the moment we run out of bits again and have to
replace the parameter attributes table encoding.
llvm-svn: 61019
llvm[2]: Linking Release executable opt (without symbols)
...
Undefined symbols:
"llvm::APFloat::IEEEsingle", referenced from:
__ZN4llvm7APFloat10IEEEsingleE$non_lazy_ptr in libLLVMCore.a(Constants.o)
__ZN4llvm7APFloat10IEEEsingleE$non_lazy_ptr in libLLVMCore.a(AsmWriter.o)
__ZN4llvm7APFloat10IEEEsingleE$non_lazy_ptr in libLLVMCore.a(ConstantFold.o)
"llvm::APFloat::IEEEdouble", referenced from:
__ZN4llvm7APFloat10IEEEdoubleE$non_lazy_ptr in libLLVMCore.a(Constants.o)
__ZN4llvm7APFloat10IEEEdoubleE$non_lazy_ptr in libLLVMCore.a(AsmWriter.o)
__ZN4llvm7APFloat10IEEEdoubleE$non_lazy_ptr in libLLVMCore.a(ConstantFold.o)
ld: symbol(s) not found
This is in release mode. To replicate, compile llvm and llvm-gcc in optimized
mode. Then build llvm, in optimized mode, with the newly created compiler.
llvm-svn: 60977
width register load followed by a truncating
store for the copy, since the load will not place
the value in the lower bits. Probably partial
loads/stores can never happen here, but fix it
anyway.
llvm-svn: 60972
use of illegal integer types: instead, use a stack slot
and copying via integer registers. The existing code
is still used if the bitconvert is to a legal integer
type.
This fires on the PPC testcases 2007-09-08-unaligned.ll
and vec_misaligned.ll. It looks like equivalent code
is generated with these changes, just permuted, but
it's hard to tell.
With these changes, nothing in LegalizeDAG produces
illegal integer types anymore. This is a prerequisite
for removing the LegalizeDAG type legalization code.
While there I noticed that the existing code doesn't
handle trunc store of f64 to f32: it turns this into
an i64 store, which represents a 4 byte stack smash.
I added a FIXME about this. Hopefully someone more
motivated than I am will take care of it.
llvm-svn: 60964
which are identical to the original patterns.
- Change the multiply with overflow so that we distinguish between signed and
unsigned multiplication. Currently, unsigned multiplication with overflow
isn't working!
llvm-svn: 60963
ISD::ADD to emit an implicit EFLAGS. This was horribly broken. Instead, replace
the intrinsic with an ISD::SADDO node. Then custom lower that into an
X86ISD::ADD node with a associated SETCC that checks the correct condition code
(overflow or carry). Then that gets lowered into the correct X86::ADDOvf
instruction.
Similar for SUB and MUL instructions.
llvm-svn: 60915
for promoted integer types, eg: i16 on ppc-32, or
i24 on any platform. Complete support for arbitrary
precision integers would require handling expanded
integer types, eg: i128, but I couldn't be bothered.
llvm-svn: 60834
parallel, allowing it to decide that P/Q must alias if A/B
must alias in things like:
P = gep A, 0, i, 1
Q = gep B, 0, i, 1
This allows GVN to delete 62 more instructions out of 403.gcc.
llvm-svn: 60820
node latencies. Use CalcLatency instead of manual code in
CalculatePriorities to keep it consistent. Previously it
computed slightly different results.
llvm-svn: 60817
- Fix bug 3185, with misc other cleanups.
- Needed to implement SPUInstrInfo::InsertBranch(). CAUTION: Not sure what
gets or needs to get passed to InsertBranch() to insert a conditional
branch. This will abort for now until a good test case shows up.
llvm-svn: 60811
overflow/carry from the "arithmetic with overflow" intrinsics. It searches the
machine basic block from bottom to top to find the SETO/SETC instruction that is
its conditional. If an instruction modifies EFLAGS before it reaches the
SETO/SETC instruction, then it defaults to the normal instruction emission.
llvm-svn: 60807
The Cost field is removed. It was only being used in a very limited way,
to indicate when the scheduler should attempt to protect a live register,
and it isn't really needed to do that. If we ever want the scheduler to
start inserting copies in non-prohibitive situations, we'll have to
rethink some things anyway.
A Latency field is added. Instead of giving each node a single
fixed latency, each edge can have its own latency. This will eventually
be used to model various micro-architecture properties more accurately.
The PointerIntPair class and an internal union are now used, which
reduce the overall size.
llvm-svn: 60806
of a pointer. This allows is to catch more equivalencies. For example,
the type_lists_compatible_p function used to require two iterations of
the gvn pass (!) to delete its 18 redundant loads because the first pass
would CSE all the addressing computation cruft, which would unblock the
second memdep/gvn passes from recognizing them. This change allows
memdep/gvn to catch all 18 when run just once on the function (as is
typical :) instead of just 3.
On all of 403.gcc, this bumps up the # reundandancies found from:
63 gvn - Number of instructions PRE'd
153991 gvn - Number of instructions deleted
50069 gvn - Number of loads deleted
to:
63 gvn - Number of instructions PRE'd
154137 gvn - Number of instructions deleted
50185 gvn - Number of loads deleted
+120 loads deleted isn't bad.
llvm-svn: 60799
essential problem was that the DAG can contain
random unused nodes which were never analyzed.
When remapping a value of a node being processed,
such a node may become used and need to be analyzed;
however due to operands being transformed during
analysis the node may morph into a different one.
Users of the morphing node need to be updated, and
this wasn't happening. While there I added a bunch
of documentation and sanity checks, so I (or some
other poor soul) won't have to scratch their head
over this stuff so long trying to remember how it
was all supposed to work next time some obscure
problem pops up! The extra sanity checking exposed
a few places where invariants weren't being preserved,
so those are fixed too. Since some of the sanity
checking is expensive, I added a flag to turn it
on. It is also turned on when building with
ENABLE_EXPENSIVE_CHECKS=1.
llvm-svn: 60797
tricks based on readnone/readonly functions.
Teach memdep to look past readonly calls when analyzing
deps for a readonly call. This allows elimination of a
few more calls from 403.gcc:
before:
63 gvn - Number of instructions PRE'd
153986 gvn - Number of instructions deleted
50069 gvn - Number of loads deleted
after:
63 gvn - Number of instructions PRE'd
153991 gvn - Number of instructions deleted
50069 gvn - Number of loads deleted
5 calls isn't much, but this adds plumbing for the next change.
llvm-svn: 60794
load dependence queries. This allows GVN to eliminate a few more
instructions on 403.gcc:
152598 gvn - Number of instructions deleted
49240 gvn - Number of loads deleted
after:
153986 gvn - Number of instructions deleted
50069 gvn - Number of loads deleted
llvm-svn: 60786
MemDep::getNonLocalPointerDependency method. There are
some open issues with this (missed optimizations) and
plenty of future work, but this does allow GVN to eliminate
*slightly* more loads (49246 vs 49033).
Switching over now allows simplification of the other code
path in memdep.
llvm-svn: 60780
the first block of a query specially. This makes the "complete query
caching" subsystem more effective, avoiding predecessor queries. This
speeds up GVN another 4%.
llvm-svn: 60752
Fix the shift amount when unrolling a vector shift into scalar shifts.
Fix problem in getShuffleScalarElt where it assumes that the input of
a bit convert must be a vector.
llvm-svn: 60740
- Change default scheduling preference to list-burr, which produces somewhat
better code than the default. Could also use list-tdrr, but need to ask
dev list about the appropriate handy mnemonic before commiting.
llvm-svn: 60738
complete. For instance, it lowers the common case into this less-than-optimal
code:
addl %ecx, %eax
seto %cl
testb %cl, %cl
jne LBB1_2 ## overflow
instead of:
addl %ecx, %eax
jo LBB1_2 ## overflow
That will come in a future commit.
llvm-svn: 60737
and use it in x86 address mode folding. Also, make
getRegForValue return 0 for illegal types even if it has a
ValueMap for them, because Argument values are put in the
ValueMap. This fixes PR3181.
llvm-svn: 60696
track of whether the CachedNonLocalPointerInfo for a block is specific
to a block. If so, just return it without any pred scanning. This is
good for a 6% speedup on GVN (when it uses this lookup method, which
it doesn't right now).
llvm-svn: 60695
method. This will eventually take over load/store dep
queries from getNonLocalDependency. For now it works
fine, but is incredibly slow because it does no caching.
Lets not switch GVN to use it until that is fixed :)
llvm-svn: 60649
clobber with the current implementation. Instead of returning
a "precise clobber" just return a fuzzy one. This doesn't
matter to any clients anyway and should speed up analysis time
very very slightly.
llvm-svn: 60641
doesn't do its own local caching, and is slightly more aggressive about
free/store dse (see testcase). This eliminates the last external client
of MemDep::getDependenceFrom().
llvm-svn: 60619
since %p isn't formatted consistently, but obviously plain %x is wrong.
PRIxPTR with a cast to uintptr_t would work here, but that requires
inconvenient build-system changes. %lu works on all current and
foreseable future hosts.
llvm-svn: 60616
loops when they can be subsumed into addressing modes.
Change X86 addressing mode check to realize that
some PIC references need an extra register.
(I believe this is correct for Linux, if not, I'm sure
someone will tell me.)
llvm-svn: 60608
1. Merge the 'None' result into 'Normal', making loads
and stores return their dependencies on allocations as Normal.
2. Split the 'Normal' result into 'Clobber' and 'Def' to
distinguish between the cases when memdep knows the value is
produced from when we just know if may be changed.
3. Move some of the logic for determining whether readonly calls
are CSEs into memdep instead of it being in GVN. This still
leaves verification that the arguments are hte same to GVN to
let it know about value equivalences in different contexts.
4. Change memdep's call/call dependency analysis to use
getModRefInfo(CallSite,CallSite) instead of doing something
very weak. This only really matters for things like DSA, but
someday maybe we'll have some other decent context sensitive
analyses :)
5. This reimplements the guts of memdep to handle the new results.
6. This simplifies GVN significantly:
a) readonly call CSE is slightly simpler
b) I eliminated the "getDependencyFrom" chaining for load
elimination and load CSE doesn't have to worry about
volatile (they are always clobbers) anymore.
c) GVN no longer does any 'lastLoad' caching, leaving it to
memdep.
7. The logic in DSE is simplified a bit and sped up. A potentially
unsafe case was eliminated.
llvm-svn: 60607
1. GlobalBaseReg may have been spilled.
2. It may not be live at the use.
3. Spiller doesn't know this is happening so it won't prevent GlobalBaseReg from being spilled later (That by itself is a nasty hack. It's needed because we don't insert the reload until later).
llvm-svn: 60595
While they appear to provide a normal clobbering def, they don't
in the case of the awkward IMPLICIT_DEF+INSERT_SUBREG idiom. It
would be good to change INSERT_SUBREG; until then, this change
allows post-regalloc scheduling to cope in a mildly conservative
way.
llvm-svn: 60583
This fixes many bugs. I will add more test cases in a separate check-in.
Some day, the code that manipulates CFG and updates dom. info could use refactoring help.
llvm-svn: 60554
- Add v4f32, v2f64 to LowerVECTOR_SHUFFLE
- Look for vector rotate in shuffle elements, generate a vector rotate
instead of a full-blown shuffle when opportunity presents itself.
- Generate larger test harness and fix a few interesting but obscure bugs.
llvm-svn: 60552
- First patch from Nehal Desai, a new contributor at Aerospace. Nehal's patch
fixes sign/zero/any-extending loads for integers and floating point. Example
code, compiled w/o debugging or optimization where he first noticed the bug:
int main(void) {
float a = 99.0;
printf("%d\n", a);
return 0;
}
Verified that this code actually works on a Cell SPU.
Changes by Scott Michel:
- Fix bug in the value type list constructed by SPUISD::LDRESULT to include
both the load result's result and chain, not just the chain alone.
- Simplify LowerLOAD and remove extraneous and unnecessary chains.
- Remove unused SPUISD pseudo instructions.
llvm-svn: 60526
1) have it fold "br undef", which does occur with
surprising frequency as jump threading iterates.
2) teach j-t to delete dead blocks. This removes the successor
edges, reducing the in-edges of other blocks, allowing
recursive simplification.
3) Fold things like:
br COND, BBX, BBY
BBX:
br COND, BBZ, BBW
which also happens because jump threading iterates.
llvm-svn: 60470
foldMemoryOperand how to "fold" them, by converting them into constant-pool
loads. When they aren't folded, they use xorps/cmpeqd, but for example when
register pressure is high, they may now be folded as memory operands, which
reduces register pressure.
Also, mark V_SET0 isAsCheapAsAMove so that two-address-elimination will
remat it instead of copying zeros around (V_SETALLONES was already marked).
llvm-svn: 60461
is set but mayLoad is not set. Fix all the problems this turned up.
Change code to not use isSimpleLoad instead of mayLoad unless it
really wants isSimpleLoad.
llvm-svn: 60459
delegates to the regular x86-32 convention which handles byval, but only
after it handles a few cases, and it's necessary to handle byval before
handling those cases. This fixes PR3122 (and rdar://6400815), llvm-gcc
miscompiling LLVM.
llvm-svn: 60453
1. ppcf128 select is expanded to f64 select's.
2. f64 select operand 0 is an i1 truncate, it's promoted to i32 zero_extend.
3. f64 select is updated. It's changed back to a "NewNode" and being re-analyzed.
4. f64 select operands are being processed. Operand 0 is a "NewNode". It's being expunged out of ReplacedValues map.
5. ExpungeNode tries to remap f64 select and notice it's a "NewNode" and assert.
Duncan, please take a look. Thanks.
llvm-svn: 60443
- 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.
llvm-svn: 60438
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.
llvm-svn: 60408
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.
llvm-svn: 60399
