While the .td entry is nice and all, it takes a pretty gross hack in
ARMAsmParser::ParseInstruction() because of handling of other "subs"
instructions to get it to match. Ran it by Jim Grosbach and he said it was
about what he expected to make this work given the existing code.
rdar://14214063
llvm-svn: 187530
If we merge vector when a vector is used, it will generate an artificial
antidependency that can prevent 2 tex/vtx instructions to use the same
clause and thus generate extra clauses that reduce performance.
There is no test case as such situation is really hard to predict.
llvm-svn: 187516
There are a lot of restrictions on instruction groups that contain
LDS instructions, so for now we will be conservative and not packetize
anything else with them.
llvm-svn: 187513
We were using two instructions for similar purpose : break and
predicated break. Only predicated_break was emitted and it was
lowered at R600ControlFlowFinalizer to JUMP;CF_BREAK;POP.
This commit simplify the situation by making AMDILCFGStructurizer
emit IF_PREDICATE;BREAK;ENDIF; instead of predicated_break (which
is now removed).
There is no functionality change.
llvm-svn: 187510
The loop optimizers were assuming that scales > 1 were OK. I think this
is actually a bug in TargetLoweringBase::isLegalAddressingMode(),
since it seems to be trying to reject anything that isn't r+i or r+r,
but it has no default case for scales other than 0, 1 or 2. Implementing
the hook for z means that z can no longer test any change there though.
llvm-svn: 187497
Extend r187495 to conditional loads. I split this out because the
easiest way seemed to be to force a particular operand order in
SystemZISelDAGToDAG.cpp.
llvm-svn: 187496
System z branches have a mask to select which of the 4 CC values should
cause the branch to be taken. We can invert a branch by inverting the mask.
However, not all instructions can produce all 4 CC values, so inverting
the branch like this can lead to some oddities. For example, integer
comparisons only produce a CC of 0 (equal), 1 (less) or 2 (greater).
If an integer EQ is reversed to NE before instruction selection,
the branch will test for 1 or 2. If instead the branch is reversed
after instruction selection (by inverting the mask), it will test for
1, 2 or 3. Both are correct, but the second isn't really canonical.
This patch therefore keeps track of which CC values are possible
and uses this when inverting a mask.
Although this is mostly cosmestic, it fixes undefined behavior
for the CIJNLH in branch-08.ll. Another fix would have been
to mask out bit 0 when generating the fused compare and branch,
but the point of this patch is that we shouldn't need to do that
in the first place.
The patch also makes it easier to reuse CC results from other instructions.
llvm-svn: 187495
r187116 moved compare-and-branch generation from the instruction-selection
pass to the peephole optimizer (via optimizeCompare). It turns out that even
this is a bit too early. Fused compare-and-branch instructions don't
interact well with predication, where a CC result is needed. They also
make it harder to reuse the CC side-effects of earlier instructions
(not yet implemented, but the subject of a later patch).
Another problem was that the AnalyzeBranch family of routines weren't
handling compares and branches, so we weren't able to reverse the fused
form in cases where we would reverse a separate branch. This could have
been fixed by extending AnalyzeBranch, but given the other problems,
I've instead moved the fusing to the long-branch pass, which is also
responsible for the opposite transformation: splitting out-of-range
compares and branches into separate compares and long branches.
I've added a test for the AnalyzeBranch problem. A test for the
predication problem is included in the next patch, which fixes a bug
in the choice of CC mask.
llvm-svn: 187494
r186399 aggressively used the RISBG instruction for immediate ANDs,
both because it can handle some values that AND IMMEDIATE can't,
and because it allows the destination register to be different from
the source. I realized later while implementing the distinct-ops
support that it would be better to leave the choice up to
convertToThreeAddress() instead. The AND IMMEDIATE form is shorter
and is less likely to be cracked.
This is a problem for 32-bit ANDs because we assume that all 32-bit
operations will leave the high word untouched, whereas RISBG used in
this way will either clear the high word or copy it from the source
register. The patch uses the z196 instruction RISBLG for this instead.
This means that z10 will be restricted to NILL, NILH and NILF for
32-bit ANDs, but I think that should be OK for now. Although we're
using z10 as the base architecture, the optimization work is going
to be focused more on z196 and zEC12.
llvm-svn: 187492
All insertf*/extractf* functions replaced with insert/extract since we have insertf and inserti forms.
Added lowering for INSERT_VECTOR_ELT / EXTRACT_VECTOR_ELT for 512-bit vectors.
Added lowering for EXTRACT/INSERT subvector for 512-bit vectors.
Added a test.
llvm-svn: 187491
This fix is very lightweight. The same fix already existed for AddRec
but was missing for NAry expressions.
This is obviously an improvement and I'm unsure how to test compile
time problems.
Patch by Xiaoyi Guo!
llvm-svn: 187475
For a testcase like the following:
typedef unsigned long uint64_t;
typedef struct {
uint64_t lo;
uint64_t hi;
} blob128_t;
void add_128_to_128(const blob128_t *in, blob128_t *res) {
asm ("PAND %1, %0" : "+Q"(*res) : "Q"(*in));
}
where we'll fail to allocate the register for the output constraint,
our matching input constraint will not find a register to match,
and could try to search past the end of the current operands array.
On the idea that we'd like to attempt to keep compilation going
to find more errors in the module, change the error cases when
we're visiting inline asm IR to return immediately and avoid
trying to create a node in the DAG. This leaves us with only
a single error message per inline asm instruction, but allows us
to safely keep going in the general case.
llvm-svn: 187470
One form would accept a vector of pointers, and the other did not.
Make both accept vectors of pointers, and add an assertion
for the number of elements.
llvm-svn: 187464
The unix one was returning no_such_file_or_directory, but the windows one
was return success.
Update the one one caller that was depending on the old behavior.
llvm-svn: 187463
This avoids constant folding bitcast/ptrtoint/inttoptr combinations
that have illegal bitcasts between differently sized address spaces.
llvm-svn: 187455
Call into ComputeMaskedBits to figure out which bits are set on both add
operands and determine if the value is a power-of-two-or-zero or not.
llvm-svn: 187445
It will now only convert the arguments / return value and call
the underlying function if the types are able to be bitcasted.
This avoids using fp<->int conversions that would occur before.
llvm-svn: 187444
When registers must be live throughout the scheduling region, increase
the limit for the register class. Once we exceed the original limit,
they will be spilled, and there's no point further reducing pressure.
This isn't a perfect heuristics but avoids a situation where the
scheduler could become trapped by trying to achieve the impossible.
llvm-svn: 187436
When simplifying a (or (and B A) (and C ~A)) to a (VBSL A B C) ensure that the
bitwidth of the second operands to both ands match before comparing the negation
of the values.
Split the check of the value of the second operands to the ands. Move the cast
and variable declaration slightly higher to make it slightly easier to follow.
Bug-Id: 16700
Signed-off-by: Saleem Abdulrasool <compnerd@compnerd.org>
llvm-svn: 187404
This is the first of many upcoming patches for PowerPC fast
instruction selection support. This patch implements the minimum
necessary for a functional (but extremely limited) FastISel pass. It
allows the table-generated portions of the selector to be created and
used, but in most cases selection will fall back to the DAG selector.
None of the block terminator instructions are implemented yet, and
most interesting instructions require some special handling.
Therefore there aren't any new test cases with this patch. There will
be quite a few tests coming with future patches.
This patch adds the make/CMake support for the new code (including
tablegen -gen-fast-isel) and creates the FastISel object for PPC64 ELF
only. It instantiates the necessary virtual functions
(TargetSelectInstruction, TargetMaterializeConstant,
TargetMaterializeAlloca, tryToFoldLoadIntoMI, and FastLowerArguments),
but of these, only TargetMaterializeConstant contains any useful
implementation. This is present since the table-generated code
requires the ability to materialize integer constants for some
instructions.
This patch has been tested by building and running the
projects/test-suite code with -O0. All tests passed with the
exception of a couple of long-running tests that time out using -O0
code generation.
llvm-svn: 187399
This patch prevents the following combine when the input vector is used more
than once.
insert_vector_elt (build_vector elt0, ..., eltN), NewEltIdx, idx
=>
build_vector elt0, ..., NewEltIdx, ..., eltN
The reasons are:
- Building a vector may be expensive, so try to reuse the existing part of a
vector instead of creating a new one (think big vectors).
- elt0 to eltN now have two users instead of one. This may prevent some other
optimizations.
llvm-svn: 187396
update testcase to make sure we generate debug info for walrus
by adding a non-trivial constructor and verify that we don't
emit an ODR signature for the type.
llvm-svn: 187393
32-bit symbols have "_" as global prefix, but when forming the name of
COMDAT sections this prefix is ignored. The current behavior assumes that
this prefix is always present which is not the case for 64-bit and names
are truncated.
llvm-svn: 187356
If no other operation is specified, 's' becomes an operation instead of an
modifier. The s operation just creates a symbol table. It is the same as
running ranlib.
We assume the archive was created by a sane ar (like llvm-ar or gnu ar) and
if the symbol table is present, then it is current. We use that to optimize
the most common case: a broken build system that thinks it has to run ranlib.
llvm-svn: 187353
infrastructure to do promotion without a domtree the same smarts about
looking through GEPs, bitcasts, etc., that I just taught mem2reg about.
This way, if SROA chooses to promote an alloca which still has some
noisy instructions this code can cope with them.
I've not used as principled of an approach here for two reasons:
1) This code doesn't really need it as we were already set up to zip
through the instructions used by the alloca.
2) I view the code here as more of a hack, and hopefully a temporary one.
The SSAUpdater path in SROA is a real sore point for me. It doesn't make
a lot of architectural sense for many reasons:
- We're likely to end up needing the domtree anyways in a subsequent
pass, so why not compute it earlier and use it.
- In the future we'll likely end up needing the domtree for parts of the
inliner itself.
- If we need to we could teach the inliner to preserve the domtree. Part
of the re-work of the pass manager will allow this to be very powerful
even in large SCCs with many functions.
- Ultimately, computing a domtree has gotten significantly faster since
the original SSAUpdater-using code went into ScalarRepl. We no longer
use domfrontiers, and much of domtree is lazily done based on queries
rather than eagerly.
- At this point keeping the SSAUpdater-based promotion saves a total of
0.7% on a build of the 'opt' tool for me. That's not a lot of
performance given the complexity!
So I'm leaving this a bit ugly in the hope that eventually we just
remove all of this nonsense.
I can't even readily test this because this code isn't reachable except
through SROA. When I re-instate the patch that fast-tracks allocas
already suitable for promotion, I'll add a testcase there that failed
before this change. Before that, SROA will fix any test case I give it.
llvm-svn: 187347
standards for LLVM. Remove duplicated comments on the interface from the
implementation file (implementation comments are left there of course).
Also clean up, re-word, and fix a few typos and errors in the commenst
spotted along the way.
This is in preparation for changes to these files and to keep the
uninteresting tidying in a separate commit.
llvm-svn: 187335
uses of an alloca, we can pre-compute promotability while analyzing an
alloca for splitting in SROA. That lets us short-circuit the common case
of a bunch of trivially promotable allocas. This cuts 20% to 30% off the
run time of SROA for typical frontend-generated IR sequneces I'm seeing.
It gets the new SROA to within 20% of ScalarRepl for such code. My
current benchmark for these numbers is PR15412, but it fits the general
pattern of IR emitted by Clang so it should be widely applicable.
llvm-svn: 187323
The tests !defined(__ppc__) && !defined(__powerpc__) are not needed
or helpful when verifying that code is being compiled for a 64-bit
target. The simpler test provided by this revision is sufficient to
tell if the target is 64-bit.
llvm-svn: 187318
IEEE-754R 1.4 Exclusions states that IEEE-754R does not specify the
interpretation of the sign of NaNs. In order to remove an irrelevant
variable that most floating point implementations do not use,
standardize add, sub, mul, div, mod so that operating anything with
NaN always yields a positive NaN.
In a later commit I am going to update the APIs for creating NaNs so
that one can not even create a negative NaN.
llvm-svn: 187314
Zeroing the significand of a floating point number does not necessarily cause a
floating point number to become finite non zero. For instance, if one has a NaN,
zeroing the significand will cause it to become +/- infinity.
llvm-svn: 187313
do in the SDag when lowering references to the GOT: use
ARMConstantPoolSymbol rather than creating a dummy global variable. The
computation of the alignment still feels weird (it uses IR types and
datalayout) but it preserves the exact previous behavior. This change
fixes the memory leak of the global variable detected on the valgrind
leak checking bot.
Thanks to Benjamin Kramer for pointing me at ARMConstantPoolSymbol to
handle this use case.
llvm-svn: 187303
me) should start watching this bot more as its catching lots of bugs.
The fix here is to not construct the global if we aren't going to need
it. That's cheaper anyways, and globals have highly predictable types in
practice. I've added an assert to catch skew between our manual testing
of the type and the actual type just for paranoia's sake.
Note that this pattern is actually fine in most globals because when you
build a global with a module it automatically is moved to be owned by
that module. But here, we're in isel and don't really want to do that.
The solution of not creating a global is simpler anyways.
llvm-svn: 187302
There doesn't appear to be any reason to put this variable on the heap.
I'm suspicious of the LexicalScope above that we stuff in a map and then
delete afterward, but I'm just trying to get the valgrind bot clean.
llvm-svn: 187301
than once, and the second time through we leaked memory. Found thanks to
the vg-leak bot, but I can't locally reproduce it with valgrind. The
debugger confirms that it is in fact leaking here.
This whole code is totally gross. Why is initialize being called on each
runOnFunction??? Why aren't these OwningPtr<>s, and why aren't their
lifetimes better defined? Anyways, this is just a surgical change to
help out the leak checking bots.
llvm-svn: 187299
their being optimized out in debug mode. Realistically, this just isn't
going to be the slow part anyways. This also fixes unused variable
warnings that are breaking LLD build bots. =/ I didn't see these at
first, and kept losing track of the fact that they were broken.
llvm-svn: 187297
analysis of the alloca. We don't need to visit all the users twice for
this. We build up a kill list during the analysis and then just process
it afterward. This recovers the tiny bit of performance lost by moving
to the visitor based analysis system as it removes one entire use-list
walk from mem2reg. In some cases, this is now faster than mem2reg was
previously.
llvm-svn: 187296
Also always add DIType, DISubprogram and DIGlobalVariable to the list
in DebugInfoFinder without checking them, so we can verify them later
on.
llvm-svn: 187285
Adds unit tests for it too.
Split BasicBlockUtils into an analysis-half and a transforms-half, and put the
analysis bits into a new Analysis/CFG.{h,cpp}. Promote isPotentiallyReachable
into llvm::isPotentiallyReachable and move it into Analysis/CFG.
llvm-svn: 187283
Merge consecutive if-regions if they contain identical statements.
Both transformations reduce number of branches. The transformation
is guarded by a target-hook, and is currently enabled only for +R600,
but the correctness has been tested on X86 target using a variety of
CPU benchmarks.
Patch by: Mei Ye
llvm-svn: 187278
Both GCC and LLVM will implicitly define __ppc__ and __powerpc__ for
all PowerPC targets, whether 32- or 64-bit. They will both implicitly
define __ppc64__ and __powerpc64__ for 64-bit PowerPC targets, and not
for 32-bit targets. We cannot be sure that all other possible
compilers used to compile Clang/LLVM define both __ppc__ and
__powerpc__, for example, so it is best to check for both when relying
on either inside the Clang/LLVM code base.
This patch makes sure we always check for both variants. In addition,
it fixes one unnecessary check in lib/Target/PowerPC/PPCJITInfo.cpp.
(At least one of __ppc__ and __powerpc__ should always be defined when
compiling for a PowerPC target, no matter which compiler is used, so
testing for them is unnecessary.)
There are some places in the compiler that check for other variants,
like __POWERPC__ and _POWER, and I have left those in place. There is
no need to add them elsewhere. This seems to be in Apple-specific
code, and I won't take a chance on breaking it.
There is no intended change in behavior; thus, no test cases are
added.
llvm-svn: 187248
We used to call Verify before adding DICompileUnit to the list, and now we
remove the check and always add DICompileUnit to the list in DebugInfoFinder,
so we can verify them later on.
llvm-svn: 187237
These were reverted in r167222 along with the rest
of the last different address space pointer size attempt.
These will be used in later commits.
llvm-svn: 187223
type units.
Initially this support is used in the computation of an ODR checker
for C++. For now we're attaching it to the DIE, but in the future
it will be attached to the type unit.
This also starts breaking out types into the separation for type
units, but without actually splitting the DIEs.
In preparation for hashing the DIEs this adds a DIEString type
that contains a StringRef with the string contained at the label.
llvm-svn: 187213
On Windows, this improves clean cmake configuration time on my
workstation from 1m58s to 1m32s, which is pretty significant. There's
probably more that can be done here, but this is the low hanging fruit.
Eric volunteered to regenerate ./configure for me.
llvm-svn: 187209
* Remove LLVM_ENABLE_CRT_REPORT. LLVM_DISABLE_CRASH_REPORT made it redundant.
* set Return to 1, so that we get a stack trace on failure.
* don't call _exit, so that we get a negative exit value and "not --crash"
correctly differentiates crashes and regular errors.
This is a bit experimental since the documentation on this interface is sparse.
It doesn't bring up a dialog on my windows setup, but feel free to revert
if it causes problem for your setup (and let me know what it is so that I
can try to fix this patch).
llvm-svn: 187206
CustomLowerNode was not being called during SplitVectorOperand,
meaning custom legalization could not be used by targets.
This also adds a test case for NVPTX that depends on this custom
legalization.
Differential Revision: http://llvm-reviews.chandlerc.com/D1195
Attempt to fix the buildbots by making the X86 test I just added platform independent
llvm-svn: 187202
This reverts commit 187198. It broke the bots.
The soft float test probably needs a -triple because of name differences.
On the hard float test I am getting a "roundss $1, %xmm0, %xmm0", instead of
"vroundss $1, %xmm0, %xmm0, %xmm0".
llvm-svn: 187201
CustomLowerNode was not being called during SplitVectorOperand,
meaning custom legalization could not be used by targets.
This also adds a test case for NVPTX that depends on this custom
legalization.
Differential Revision: http://llvm-reviews.chandlerc.com/D1195
llvm-svn: 187198
robust. It now uses an InstVisitor and worklist to actually walk the
uses of the Alloca transitively and detect the pattern which we can
directly promote: loads & stores of the whole alloca and instructions we
can completely ignore.
Also, with this new implementation teach both the predicate for testing
whether we can promote and the promotion engine itself to use the same
code so we no longer have strange divergence between the two code paths.
I've added some silly test cases to demonstrate that we can handle
slightly more degenerate code patterns now. See the below for why this
is even interesting.
Performance impact: roughly 1% regression in the performance of SROA or
ScalarRepl on a large C++-ish test case where most of the allocas are
basically ready for promotion. The reason is because of silly redundant
work that I've left FIXMEs for and which I'll address in the next
commit. I wanted to separate this commit as it changes the behavior.
Once the redundant work in removing the dead uses of the alloca is
fixed, this code appears to be faster than the old version. =]
So why is this useful? Because the previous requirement for promotion
required a *specific* visit pattern of the uses of the alloca to verify:
we *had* to look for no more than 1 intervening use. The end goal is to
have SROA automatically detect when an alloca is already promotable and
directly hand it to the mem2reg machinery rather than trying to
partition and rewrite it. This is a 25% or more performance improvement
for SROA, and a significant chunk of the delta between it and
ScalarRepl. To get there, we need to make mem2reg actually capable of
promoting allocas which *look* promotable to SROA without have SROA do
tons of work to massage the code into just the right form.
This is actually the tip of the iceberg. There are tremendous potential
savings we can realize here by de-duplicating work between mem2reg and
SROA.
llvm-svn: 187191
The bitcode representation attribute kinds are encoded into / decoded from
should be independent of the current set of LLVM attributes and their position
in the AttrKind enum. This patch explicitly encodes attributes to fixed bitcode
values.
With this patch applied, LLVM does not silently misread attributes written by
LLVM 3.3. We also enhance the decoding slightly such that an error message is
printed if an unknown AttrKind encoding was dected.
Bonus: Dropping bitcode attributes from AttrKind is now easy, as old AttrKinds
do not need to be kept to support the Bitcode reader.
llvm-svn: 187186
This patch provides basic support for powerpc64le as an LLVM target.
However, use of this target will not actually generate little-endian
code. Instead, use of the target will cause the correct little-endian
built-in defines to be generated, so that code that tests for
__LITTLE_ENDIAN__, for example, will be correctly parsed for
syntax-only testing. Code generation will otherwise be the same as
powerpc64 (big-endian), for now.
The patch leaves open the possibility of creating a little-endian
PowerPC64 back end, but there is no immediate intent to create such a
thing.
The LLVM portions of this patch simply add ppc64le coverage everywhere
that ppc64 coverage currently exists. There is nothing of any import
worth testing until such time as little-endian code generation is
implemented. In the corresponding Clang patch, there is a new test
case variant to ensure that correct built-in defines for little-endian
code are generated.
llvm-svn: 187179
The previous change to local live range allocation also suppressed
eviction of local ranges. In rare cases, this could result in more
expensive register choices. This commit actually revives a feature
that I added long ago: check if live ranges can be reassigned before
eviction. But now it only happens in rare cases of evicting a local
live range because another local live range wants a cheaper register.
The benefit is improved code size for some benchmarks on x86 and armv7.
I measured no significant compile time increase and performance
changes are noise.
llvm-svn: 187140
Also avoid locals evicting locals just because they want a cheaper register.
Problem: MI Sched knows exactly how many registers we have and assumes
they can be colored. In cases where we have large blocks, usually from
unrolled loops, greedy coloring fails. This is a source of
"regressions" from the MI Scheduler on x86. I noticed this issue on
x86 where we have long chains of two-address defs in the same live
range. It's easy to see this in matrix multiplication benchmarks like
IRSmk and even the unit test misched-matmul.ll.
A fundamental difference between the LLVM register allocator and
conventional graph coloring is that in our model a live range can't
discover its neighbors, it can only verify its neighbors. That's why
we initially went for greedy coloring and added eviction to deal with
the hard cases. However, for singly defined and two-address live
ranges, we can optimally color without visiting neighbors simply by
processing the live ranges in instruction order.
Other beneficial side effects:
It is much easier to understand and debug regalloc for large blocks
when the live ranges are allocated in order. Yes, global allocation is
still very confusing, but it's nice to be able to comprehend what
happened locally.
Heuristics could be added to bias register assignment based on
instruction locality (think late register pairing, banks...).
Intuituvely this will make some test cases that are on the threshold
of register pressure more stable.
llvm-svn: 187139
Before the patch we took advantage of the fact that the compare and
branch are glued together in the selection DAG and fused them together
(where possible) while emitting them. This seemed to work well in practice.
However, fusing the compare so early makes it harder to remove redundant
compares in cases where CC already has a suitable value. This patch
therefore uses the peephole analyzeCompare/optimizeCompareInstr pair of
functions instead.
No behavioral change intended, but it paves the way for a later patch.
llvm-svn: 187116
These instructions are allowed to trap even if the condition is false,
so for now they are only used for "*ptr = (cond ? x : *ptr)"-style
constructs.
llvm-svn: 187111
Make sure the context and type fields are MDNodes. We will generate
verification errors if those fields are non-empty strings.
Fix testing cases to make them pass the verifier.
llvm-svn: 187106
The language reference says that:
"If a symbol appears in the @llvm.used list, then the compiler,
assembler, and linker are required to treat the symbol as if there is
a reference to the symbol that it cannot see"
Since even the linker cannot see the reference, we must assume that
the reference can be using the symbol table. For example, a user can add
__attribute__((used)) to a debug helper function like dump and use it from
a debugger.
llvm-svn: 187103
Kevin Enderby