Adds an interface to get ProfileSummary for a module and makes InlineCost use ProfileSummary to get max function count.
Differential Revision: http://reviews.llvm.org/D18622
llvm-svn: 266477
Divisions by a constant can be converted into multiplies which are usually
cheaper, but this isn't possible if the constant gets separated (particularly
in loops). Fix this by telling ConstantHoisting that the immediate in a DIV is
cheap.
I considered making the check generic, but neither AArch64 (strangely) nor x86
showed any benefit on the tests I had.
llvm-svn: 266464
InstCombine wants to optimize compares of calls to fabs with zero.
However, we didn't have the necessary legality checking to verify that
the function call had the same behavior as fabs.
llvm-svn: 266452
Currently each Function points to a DISubprogram and DISubprogram has a
scope field. For member functions the scope is a DICompositeType. DIScopes
point to the DICompileUnit to facilitate type uniquing.
Distinct DISubprograms (with isDefinition: true) are not part of the type
hierarchy and cannot be uniqued. This change removes the subprograms
list from DICompileUnit and instead adds a pointer to the owning compile
unit to distinct DISubprograms. This would make it easy for ThinLTO to
strip unneeded DISubprograms and their transitively referenced debug info.
Motivation
----------
Materializing DISubprograms is currently the most expensive operation when
doing a ThinLTO build of clang.
We want the DISubprogram to be stored in a separate Bitcode block (or the
same block as the function body) so we can avoid having to expensively
deserialize all DISubprograms together with the global metadata. If a
function has been inlined into another subprogram we need to store a
reference the block containing the inlined subprogram.
Attached to https://llvm.org/bugs/show_bug.cgi?id=27284 is a python script
that updates LLVM IR testcases to the new format.
http://reviews.llvm.org/D19034
<rdar://problem/25256815>
llvm-svn: 266446
This is almost identical to:
http://reviews.llvm.org/rL264527
This doesn't solve PR27344; it just allows the profile weights to survive.
To solve the bug, we need to use the profile weights in the backend.
llvm-svn: 266442
Summary:
This lets us add this pass to the IR pass manager unconditionally; it
will simply not do anything on targets without branch divergence.
Reviewers: tra
Subscribers: llvm-commits, jingyue, rnk, chandlerc
Differential Revision: http://reviews.llvm.org/D18625
llvm-svn: 266398
If the size of an AST entry changes, we also need to make sure we perform
necessary alias set merges, as the new size may overlap pointers in other sets.
We happen to run into this with memset, because memset allows an entry for a
i8* pointer to have a decidedly non-i8 size.
This fixes PR27262.
Differential Revision: http://reviews.llvm.org/D18939
llvm-svn: 266381
Some SIMD implementations are not IEEE-754 compliant, for example ARM's NEON.
This patch teaches the loop vectorizer to only allow transformations of loops
that either contain no floating-point operations or have enough allowance
flags supporting lack of precision (ex. -ffast-math, Darwin).
For that, the target description now has a method which tells us if the
vectorizer is allowed to handle FP math without falling into unsafe
representations, plus a check on every FP instruction in the candidate loop
to check for the safety flags.
This commit makes LLVM behave like GCC with respect to ARM NEON support, but
it stops short of fixing the underlying problem: sub-normals. Neither GCC
nor LLVM have a flag for allowing sub-normal operations. Before this patch,
GCC only allows it using unsafe-math flags and LLVM allows it by default with
no way to turn it off (short of not using NEON at all).
As a first step, we push this change to make it safe and in sync with GCC.
The second step is to discuss a new sub-normal's flag on both communitues
and come up with a common solution. The third step is to improve the FastMath
flags in LLVM to encode sub-normals and use those flags to restrict NEON FP.
Fixes PR16275.
llvm-svn: 266363
https://llvm.org/bugs/show_bug.cgi?id=27105
We can check if all bits outside of a constant mask are set with a
single constant.
As noted in the bug report, although this form should be considered the
canonical IR, backends may want to transform this into an 'andn' / 'andc'
comparison against zero because that could be a single machine instruction.
Differential Revision: http://reviews.llvm.org/D18842
llvm-svn: 266362
At some point, ARM stopped getting any benefit from ConstantHoisting because
the pass called a different variant of getIntImmCost. Reimplementing the
correct variant revealed some problems, however:
+ ConstantHoisting was modifying switch statements. This is simply invalid,
the cases must remain integer constants no matter the notional cost.
+ ConstantHoisting was mangling alloca instructions in the entry block. These
should be handled by FrameLowering, so constants actually have a cost of 0.
Worse, the resulting bitcasts meant they became dynamic allocas.
rdar://25707382
llvm-svn: 266260
LLVM optimization passes may reduce a profiled target expression
to a constant. Removing runtime calls at such instrumentation points
would help speedup the runtime of the instrumented program.
llvm-svn: 266229
Summary:
To be able to work accurately on the reference graph when taking decision
about internalizing, promoting, renaming, etc. We need to have the alias
information explicit.
Reviewers: tejohnson
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D18836
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 266214
This patch fixes calculating of builtin_object_size if it depends on a
condition. Before this patch compiler did not know how to calculate the
object size when it finds a condition that cannot be eliminated.
This patch enables calculating of builtin_object_size even in case when
condition cannot be eliminated by choosing minimum or maximum value as a
result from condition. Choosing minimum or maximum value from condition
is based on the second argument of __builtin_object_size function.
Patch by Strahinja Petrovic.
Differential Revision: http://reviews.llvm.org/D18438
llvm-svn: 266193
Remove an ad-hoc transform in InstCombine and replace it with more
general machinery (ValueTracking, InstructionSimplify and VectorUtils).
This fixes PR27332.
llvm-svn: 266175
This bug was introduced with:
http://reviews.llvm.org/rL262269
AVX masked loads are specified to set vector lanes to zero when the high bit of the mask
element for that lane is zero:
"If the mask is 0, the corresponding data element is set to zero in the load form of these
instructions, and unmodified in the store form." --Intel manual
Differential Revision: http://reviews.llvm.org/D19017
llvm-svn: 266148
Summary:
For correct handling of alias to nameless
function, we need to be able to refer them through a GUID in the summary.
Here we name them using a hash of the non-private global names in the module.
Reviewers: tejohnson
Subscribers: joker.eph, llvm-commits
Differential Revision: http://reviews.llvm.org/D18883
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 266132
Summary:
Let keep llvm-as "dumb": it converts textual IR to bitcode. This
commit removes the dependency from llvm-as to libLLVMAnalysis.
We'll add back summary in llvm-as if we get to a textual
representation for it at some point. In the meantime, opt seems
like a better place for that.
Reviewers: tejohnson
Subscribers: joker.eph, llvm-commits
Differential Revision: http://reviews.llvm.org/D19032
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 266131
(Recommit of r266002, with r266011, r266016, and not accidentally
including an extra unused/uninitialized element in LibcallRoutineNames)
AtomicExpandPass can now lower atomic load, atomic store, atomicrmw, and
cmpxchg instructions to __atomic_* library calls, when the target
doesn't support atomics of a given size.
This is the first step towards moving all atomic lowering from clang
into llvm. When all is done, the behavior of __sync_* builtins,
__atomic_* builtins, and C11 atomics will be unified.
Previously LLVM would pass everything through to the ISelLowering
code. There, unsupported atomic instructions would turn into __sync_*
library calls. Because of that behavior, Clang currently avoids emitting
llvm IR atomic instructions when this would happen, and emits __atomic_*
library functions itself, in the frontend.
This change makes LLVM able to emit __atomic_* libcalls, and thus will
eventually allow clang to depend on LLVM to do the right thing.
It is advantageous to do the new lowering to atomic libcalls in
AtomicExpandPass, before ISel time, because it's important that all
atomic operations for a given size either lower to __atomic_*
libcalls (which may use locks), or native instructions which won't. No
mixing and matching.
At the moment, this code is enabled only for SPARC, as a
demonstration. The next commit will expand support to all of the other
targets.
Differential Revision: http://reviews.llvm.org/D18200
llvm-svn: 266115
This is a resubmittion of 263158 change.
This patch fixes the problem which occurs when loop-vectorize tries to use @llvm.masked.load/store intrinsic for a non-default addrspace pointer. It fails with "Calling a function with a bad signature!" assertion in CallInst constructor because it tries to pass a non-default addrspace pointer to the pointer argument which has default addrspace.
The fix is to add pointer type as another overloaded type to @llvm.masked.load/store intrinsics.
Reviewed By: reames
Differential Revision: http://reviews.llvm.org/D17270
llvm-svn: 266086
They broke the msan bot.
Original message:
Add __atomic_* lowering to AtomicExpandPass.
AtomicExpandPass can now lower atomic load, atomic store, atomicrmw,and
cmpxchg instructions to __atomic_* library calls, when the target
doesn't support atomics of a given size.
This is the first step towards moving all atomic lowering from clang
into llvm. When all is done, the behavior of __sync_* builtins,
__atomic_* builtins, and C11 atomics will be unified.
Previously LLVM would pass everything through to the ISelLowering
code. There, unsupported atomic instructions would turn into __sync_*
library calls. Because of that behavior, Clang currently avoids emitting
llvm IR atomic instructions when this would happen, and emits __atomic_*
library functions itself, in the frontend.
This change makes LLVM able to emit __atomic_* libcalls, and thus will
eventually allow clang to depend on LLVM to do the right thing.
It is advantageous to do the new lowering to atomic libcalls in
AtomicExpandPass, before ISel time, because it's important that all
atomic operations for a given size either lower to __atomic_*
libcalls (which may use locks), or native instructions which won't. No
mixing and matching.
At the moment, this code is enabled only for SPARC, as a
demonstration. The next commit will expand support to all of the other
targets.
Differential Revision: http://reviews.llvm.org/D18200
llvm-svn: 266062
`allocsize` is a function attribute that allows users to request that
LLVM treat arbitrary functions as allocation functions.
This patch makes LLVM accept the `allocsize` attribute, and makes
`@llvm.objectsize` recognize said attribute.
The review for this was split into two patches for ease of reviewing:
D18974 and D14933. As promised on the revisions, I'm landing both
patches as a single commit.
Differential Revision: http://reviews.llvm.org/D14933
llvm-svn: 266032
r237193 fix handling of alloca size / align in MergeFunctions, but only tested one and didn't follow FunctionComparator::cmpOperations's usual comparison pattern. It also didn't update Instruction.cpp:haveSameSpecialState which I'll do separately.
llvm-svn: 266022
This is more robust to changes in the link ordering.
Differential Revision: http://reviews.llvm.org/D18946
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 266018
Add StackProtector to SafeStack. This adds limited protection against
data corruption in the caller frame. Current implementation treats
all stack protector levels as -fstack-protector-all.
llvm-svn: 266004
AtomicExpandPass can now lower atomic load, atomic store, atomicrmw, and
cmpxchg instructions to __atomic_* library calls, when the target
doesn't support atomics of a given size.
This is the first step towards moving all atomic lowering from clang
into llvm. When all is done, the behavior of __sync_* builtins,
__atomic_* builtins, and C11 atomics will be unified.
Previously LLVM would pass everything through to the ISelLowering
code. There, unsupported atomic instructions would turn into __sync_*
library calls. Because of that behavior, Clang currently avoids emitting
llvm IR atomic instructions when this would happen, and emits __atomic_*
library functions itself, in the frontend.
This change makes LLVM able to emit __atomic_* libcalls, and thus will
eventually allow clang to depend on LLVM to do the right thing.
It is advantageous to do the new lowering to atomic libcalls in
AtomicExpandPass, before ISel time, because it's important that all
atomic operations for a given size either lower to __atomic_*
libcalls (which may use locks), or native instructions which won't. No
mixing and matching.
At the moment, this code is enabled only for SPARC, as a
demonstration. The next commit will expand support to all of the other
targets.
Differential Revision: http://reviews.llvm.org/D18200
llvm-svn: 266002
This patch ensures that when we detect first-order recurrences, we reject a phi
node if its previous value is also a phi node. During vectorization the initial
and previous values of the recurrence are shuffled together to create the value
for the current iteration. However, phi nodes are not widened like other
instructions. This fixes PR27246.
Differential Revision: http://reviews.llvm.org/D18971
llvm-svn: 265983
This is the straightforward fix for PR26760:
https://llvm.org/bugs/show_bug.cgi?id=26760
But we still need to make some changes to generalize this helper function
and then send the lshr case into here.
llvm-svn: 265960
Summary:
This is the first step in also serializing the index out to LLVM
assembly.
The per-module summary written to bitcode is moved out of the bitcode
writer and to a new analysis pass (ModuleSummaryIndexWrapperPass).
The pass itself uses a new builder class to compute index, and the
builder class is used directly in places where we don't have a pass
manager (e.g. llvm-as).
Because we are computing summaries outside of the bitcode writer, we no
longer can use value ids created by the bitcode writer's
ValueEnumerator. This required changing the reference graph edge type
to use a new ValueInfo class holding a union between a GUID (combined
index) and Value* (permodule index). The Value* are converted to the
appropriate value ID during bitcode writing.
Also, this enables removal of the BitWriter library's dependence on the
Analysis library that was previously required for the summary computation.
Reviewers: joker.eph
Subscribers: joker.eph, llvm-commits
Differential Revision: http://reviews.llvm.org/D18763
llvm-svn: 265941
Summary:
If we can prove that an op.with.overflow intrinsic does not overflow, we
can get rid of the intrinsic, and replace it with non-wrapping
arithmetic.
Reviewers: atrick, regehr
Subscribers: sanjoy, mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D18685
llvm-svn: 265913
Vectorization cost of uniform load wasn't correctly calculated.
As a result, a simple loop that loads a uniform value wasn't vectorized.
Differential Revision: http://reviews.llvm.org/D18940
llvm-svn: 265901
Summary:
The llvm cos intrinsic currently does not propagate undef's. This change
transforms cos(undef) to null value or 0.
There are 2 test cases added as well.
Patch by Anna Thomas!
Reviewers: sanjoy
Subscribers: majnemer, llvm-commits
Differential Revision: http://reviews.llvm.org/D18863
llvm-svn: 265825
We had a select of a cast of a select but attempted to replace the outer
select with the inner select dispite their incompatible types.
Patch by Anton Korobeynikov!
This fixes PR27236.
llvm-svn: 265805
InstCombine cannot effectively remove redundant assumptions without them
registered in the assumption cache. The vectorizer can create identical
assumptions but doesn't register them with the cache, resulting in
slower compile times because InstCombine tries to reason about a lot
more assumptions.
Fix this by registering the cloned assumptions.
llvm-svn: 265800
This re-commits r265535 which was reverted in r265541 because it
broke the windows bots. The problem was that we had a PointerIntPair
which took a pointer to a struct allocated with new. The problem
was that new doesn't provide sufficient alignment guarantees.
This pattern was already present before r265535 and it just happened
to work. To fix this, we now separate the PointerToIntPair from the
ExitNotTakenInfo struct into a pointer and a bool.
Original commit message:
Summary:
When the backedge taken codition is computed from an icmp, SCEV can
deduce the backedge taken count only if one of the sides of the icmp
is an AddRecExpr. However, due to sign/zero extensions, we sometimes
end up with something that is not an AddRecExpr.
However, we can use SCEV predicates to produce a 'guarded' expression.
This change adds a method to SCEV to get this expression, and the
SCEV predicate associated with it.
In HowManyGreaterThans and HowManyLessThans we will now add a SCEV
predicate associated with the guarded backedge taken count when the
analyzed SCEV expression is not an AddRecExpr. Note that we only do
this as an alternative to returning a 'CouldNotCompute'.
We use new feature in Loop Access Analysis and LoopVectorize to analyze
and transform more loops.
Reviewers: anemet, mzolotukhin, hfinkel, sanjoy
Subscribers: flyingforyou, mcrosier, atrick, mssimpso, sanjoy, mzolotukhin, llvm-commits
Differential Revision: http://reviews.llvm.org/D17201
llvm-svn: 265786
This reverts commit r265765, reapplying r265759 after changing a call from
LocalAsMetadata::get to ValueAsMetadata::get (and adding a unit test). When a
local value is mapped to a constant (like "i32 %a" => "i32 7"), the new debug
intrinsic operand may no longer be pointing at a local.
http://lab.llvm.org:8080/green/job/clang-stage1-configure-RA_build/19020/
The previous coommit message follows:
--
This is a partial re-commit -- maybe more of a re-implementation -- of
r265631 (reverted in r265637).
This makes RF_IgnoreMissingLocals behave (almost) consistently between
the Value and the Metadata hierarchy. In particular:
- MapValue returns nullptr or "metadata !{}" for missing locals in
MetadataAsValue/LocalAsMetadata bridging paris, depending on
the RF_IgnoreMissingLocals flag.
- MapValue doesn't memoize LocalAsMetadata-related results.
- MapMetadata no longer deals with LocalAsMetadata or
RF_IgnoreMissingLocals at all. (This wasn't in r265631 at all, but
I realized during testing it would make the patch simpler with no
loss of generality.)
r265631 went too far, making both functions universally ignore
RF_IgnoreMissingLocals. This broke building (e.g.) compiler-rt.
Reassociate (and possibly other passes) don't currently maintain
dominates-use invariants for metadata operands, resulting in IR like
this:
define void @foo(i32 %arg) {
call void @llvm.some.intrinsic(metadata i32 %x)
%x = add i32 1, i32 %arg
}
If the inliner chooses to inline @foo into another function, then
RemapInstruction will call `MapValue(metadata i32 %x)` and assert that
the return is not nullptr.
I've filed PR27273 to add a Verifier check and fix the underlying
problem in the optimization passes.
As a workaround, return `!{}` instead of nullptr for unmapped
LocalAsMetadata when RF_IgnoreMissingLocals is unset. Otherwise, match
the behaviour of r265631.
Original commit message:
ValueMapper: Make LocalAsMetadata match function-local Values
Start treating LocalAsMetadata similarly to function-local members of
the Value hierarchy in MapValue and MapMetadata.
- Don't memoize them.
- Return nullptr if they are missing.
This also cleans up ConstantAsMetadata to stop listening to the
RF_IgnoreMissingLocals flag.
llvm-svn: 265768
Summary:
Fixes PR26774.
If you're aware of the issue, feel free to skip the "Motivation"
section and jump directly to "This patch".
Motivation:
I define "refinement" as discarding behaviors from a program that the
optimizer has license to discard. So transforming:
```
void f(unsigned x) {
unsigned t = 5 / x;
(void)t;
}
```
to
```
void f(unsigned x) { }
```
is refinement, since the behavior went from "if x == 0 then undefined
else nothing" to "nothing" (the optimizer has license to discard
undefined behavior).
Refinement is a fundamental aspect of many mid-level optimizations done
by LLVM. For instance, transforming `x == (x + 1)` to `false` also
involves refinement since the expression's value went from "if x is
`undef` then { `true` or `false` } else { `false` }" to "`false`" (by
definition, the optimizer has license to fold `undef` to any non-`undef`
value).
Unfortunately, refinement implies that the optimizer cannot assume
that the implementation of a function it can see has all of the
behavior an unoptimized or a differently optimized version of the same
function can have. This is a problem for functions with comdat
linkage, where a function can be replaced by an unoptimized or a
differently optimized version of the same source level function.
For instance, FunctionAttrs cannot assume a comdat function is
actually `readnone` even if it does not have any loads or stores in
it; since there may have been loads and stores in the "original
function" that were refined out in the currently visible variant, and
at the link step the linker may in fact choose an implementation with
a load or a store. As an example, consider a function that does two
atomic loads from the same memory location, and writes to memory only
if the two values are not equal. The optimizer is allowed to refine
this function by first CSE'ing the two loads, and the folding the
comparision to always report that the two values are equal. Such a
refined variant will look like it is `readonly`. However, the
unoptimized version of the function can still write to memory (since
the two loads //can// result in different values), and selecting the
unoptimized version at link time will retroactively invalidate
transforms we may have done under the assumption that the function
does not write to memory.
Note: this is not just a problem with atomics or with linking
differently optimized object files. See PR26774 for more realistic
examples that involved neither.
This patch:
This change introduces a new set of linkage types, predicated as
`GlobalValue::mayBeDerefined` that returns true if the linkage type
allows a function to be replaced by a differently optimized variant at
link time. It then changes a set of IPO passes to bail out if they see
such a function.
Reviewers: chandlerc, hfinkel, dexonsmith, joker.eph, rnk
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D18634
llvm-svn: 265762
This is a partial re-commit -- maybe more of a re-implementation -- of
r265631 (reverted in r265637).
This makes RF_IgnoreMissingLocals behave (almost) consistently between
the Value and the Metadata hierarchy. In particular:
- MapValue returns nullptr or "metadata !{}" for missing locals in
MetadataAsValue/LocalAsMetadata bridging paris, depending on
the RF_IgnoreMissingLocals flag.
- MapValue doesn't memoize LocalAsMetadata-related results.
- MapMetadata no longer deals with LocalAsMetadata or
RF_IgnoreMissingLocals at all. (This wasn't in r265631 at all, but
I realized during testing it would make the patch simpler with no
loss of generality.)
r265631 went too far, making both functions universally ignore
RF_IgnoreMissingLocals. This broke building (e.g.) compiler-rt.
Reassociate (and possibly other passes) don't currently maintain
dominates-use invariants for metadata operands, resulting in IR like
this:
define void @foo(i32 %arg) {
call void @llvm.some.intrinsic(metadata i32 %x)
%x = add i32 1, i32 %arg
}
If the inliner chooses to inline @foo into another function, then
RemapInstruction will call `MapValue(metadata i32 %x)` and assert that
the return is not nullptr.
I've filed PR27273 to add a Verifier check and fix the underlying
problem in the optimization passes.
As a workaround, return `!{}` instead of nullptr for unmapped
LocalAsMetadata when RF_IgnoreMissingLocals is unset. Otherwise, match
the behaviour of r265631.
Original commit message:
ValueMapper: Make LocalAsMetadata match function-local Values
Start treating LocalAsMetadata similarly to function-local members of
the Value hierarchy in MapValue and MapMetadata.
- Don't memoize them.
- Return nullptr if they are missing.
This also cleans up ConstantAsMetadata to stop listening to the
RF_IgnoreMissingLocals flag.
llvm-svn: 265759
When GVN wants to re-interpret an already available value in a smaller
type, it needs to right-shift the value on big-endian systems to ensure
the correct bytes are accessed. The shift value is the difference of
the sizes of the two types.
This is correct as long as both types occupy multiples of full bytes.
However, when one of them is a sub-byte type like i1, this no longer
holds true: we still need to shift, but only to access the correct
*byte*. Accessing bits within the byte requires no shift in either
endianness; e.g. an i1 resides in the least-significant bit of its
containing byte on both big- and little-endian systems.
Therefore, the appropriate shift value to be used is the difference of
the *storage* sizes of the two types. This is already handled correctly
in one place where such a shift takes place (GetStoreValueForLoad), but
is incorrect in two other places: GetLoadValueForLoad and
CoerceAvailableValueToLoadType.
This patch changes both places to use the storage size as well.
Differential Revision: http://reviews.llvm.org/D18662
llvm-svn: 265684
Updating dominators for exit-blocks of the unrolled loops is not enough,
as shown in PR27157. The proper way is to update dominators for all
dominance-children of original loop blocks.
llvm-svn: 265605
Summary:
When the backedge taken codition is computed from an icmp, SCEV can
deduce the backedge taken count only if one of the sides of the icmp
is an AddRecExpr. However, due to sign/zero extensions, we sometimes
end up with something that is not an AddRecExpr.
However, we can use SCEV predicates to produce a 'guarded' expression.
This change adds a method to SCEV to get this expression, and the
SCEV predicate associated with it.
In HowManyGreaterThans and HowManyLessThans we will now add a SCEV
predicate associated with the guarded backedge taken count when the
analyzed SCEV expression is not an AddRecExpr. Note that we only do
this as an alternative to returning a 'CouldNotCompute'.
We use new feature in Loop Access Analysis and LoopVectorize to analyze
and transform more loops.
Reviewers: anemet, mzolotukhin, hfinkel, sanjoy
Subscribers: flyingforyou, mcrosier, atrick, mssimpso, sanjoy, mzolotukhin, llvm-commits
Differential Revision: http://reviews.llvm.org/D17201
llvm-svn: 265535
To quote the langref "Unlike sqrt in libm, however, llvm.sqrt has
undefined behavior for negative numbers other than -0.0 (which allows
for better optimization, because there is no need to worry about errno
being set). llvm.sqrt(-0.0) is defined to return -0.0 like IEEE sqrt."
This means that it's unsafe to replace sqrt with llvm.sqrt unless the
call is annotated with nnan.
Thanks to Hal Finkel for pointing this out!
llvm-svn: 265521
r265273 added Mapper::mapBlockAddress, which delays mapping a
blockaddress value until the function has a body. The condition was
backwards, and should be checking Function::empty instead of
GlobalValue::isDeclaration.
llvm-svn: 265508
Don't emit a gc.result for a statepoint lowered from
@llvm.experimental.deoptimize since the call into __llvm_deoptimize is
effectively noreturn. Instead follow the corresponding gc.statepoint
with an "unreachable".
llvm-svn: 265485
utils/update_test_checks.py was improved with:
http://reviews.llvm.org/rL265414
to CHECK-NEXT the first line of the IR function. This ensures that nothing bad
has happened before that.
llvm-svn: 265417
Presently, CodeGenPrepare deletes all nearly empty (only phi and branch)
basic blocks. This pass can delete loop preheaders which frequently creates
critical edges. A preheader can be a convenient place to spill registers to
the stack. If the entrance to a loop body is a critical edge, then spills
may occur in the loop body rather than immediately before it. This patch
protects loop preheaders from deletion in CodeGenPrepare even if they are
nearly empty.
Since the patch alters the CFG, it affects a large number of test cases.
In most cases, the changes are merely cosmetic (basic blocks have different
names or instruction orders change slightly). I am somewhat concerned about
the test/CodeGen/Mips/brdelayslot.ll test case. If the loop preheader is not
deleted, then the MIPS backend does not take advantage of a branch delay
slot. Consequently, I would like some close review by a MIPS expert.
The patch also partially subsumes D16893 from George Burgess IV. George
correctly notes that CodeGenPrepare does not actually preserve the dominator
tree. I think the dominator tree was usually not valid when CodeGenPrepare
ran, but I am using LoopInfo to mark preheaders, so the dominator tree is
now always valid before CodeGenPrepare.
Author: Tom Jablin (tjablin)
Reviewers: hfinkel george.burgess.iv vkalintiris dsanders kbarton cycheng
http://reviews.llvm.org/D16984
llvm-svn: 265397
Direct callees' that are cast to other function prototypes,
show up in the Call/Invoke instructions as ConstantExpr's.
Currently llvm::CallSite's getCalledFunction() fails
to return the callees in such expressions as direct calls.
Value profiling should avoid instrumenting such cases. Mostly NFC.
llvm-svn: 265330
Summary:
Useful for debugging since we lose this correlation after the permodule
summary/VST is read and until we later materialize source modules in the
function importer.
Reviewers: joker.eph
Subscribers: llvm-commits, joker.eph
Differential Revision: http://reviews.llvm.org/D18555
llvm-svn: 265327
Sinking comparisons in CGP can undo the job of hoisting them done
earlier by LICM, and soft-FP makes this an expensive mistake.
A common pattern that produces floating point comparisons uniform
over a loop is an explicit check for division by zero. If the divisor
is hoisted out of the loop, the comparison can also be, but hoisting
the function that unwinds is never legal, since it may cause side
effects in the loop body prior to the unwinding to not be executed.
Differential Revision: http://reviews.llvm.org/D18744
llvm-svn: 265264
Floating point intrinsics in LLVM are generally not speculatively
executed, since most of them are defined to behave the same as libm
functions, which set errno.
However, the only error that can happen when executing ceil, floor,
nearbyint, rint and round libm functions per POSIX.1-2001 is -ERANGE,
and that requires the maximum value of the exponent to be smaller
than the number of mantissa bits, which is not the case with any of
the floating point types supported by LLVM.
The trunc and copysign functions never set errno per per POSIX.1-2001.
Differential Revision: http://reviews.llvm.org/D18643
llvm-svn: 265262
A catchswitch cannot be preceded by another instruction in the same
basic block (other than a PHI node).
Instead, insert the extract element right after the materialization of
the vectorized value. This isn't optimal but is a reasonable compromise
given the constraints of WinEH.
This fixes PR27163.
llvm-svn: 265157
They're not necessary (since the stack pointer is trivially restored on
return), and the way LLVM inserts the stackrestore calls breaks the
IR (we get a stackrestore between the deoptimize call and the return).
llvm-svn: 265101
They're not necessary (since the lifetime of the alloca is trivially
over due to the return), and the way LLVM inserts the lifetime.end
markers breaks the IR (we get a lifetime end marker between the
deoptimize call and the return).
llvm-svn: 265100
This mostly cosmetic patch moves the DebugEmissionKind enum from DIBuilder
into DICompileUnit. DIBuilder is not the right place for this enum to live
in — a metadata consumer should not have to include DIBuilder.h.
I also added a Verifier check that checks that the emission kind of a
DICompileUnit is actually legal.
http://reviews.llvm.org/D18612
<rdar://problem/25427165>
llvm-svn: 265077
This patch simply mirrors the attributes we give to @llvm.nvvm.reflect
to the __nvvm_reflect libdevice call. This shaves about 30% of the code
in libdevice away because of CSE opportunities. It's also helps us
figure out that libdevice implementations of transcendental functions
don't have side-effects.
llvm-svn: 265060
Summary:
As discussed on llvm-dev[1].
This change adds the basic boilerplate code around having this intrinsic
in LLVM:
- Changes in Intrinsics.td, and the IR Verifier
- A lowering pass to lower @llvm.experimental.guard to normal
control flow
- Inliner support
[1]: http://lists.llvm.org/pipermail/llvm-dev/2016-February/095523.html
Reviewers: reames, atrick, chandlerc, rnk, JosephTremoulet, echristo
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D18527
llvm-svn: 264976
Widening a PHI requires us to insert a trunc.
The logical place for this trunc is in the same BB as the PHI.
This is not possible if the BB is terminated by a catchswitch.
This fixes PR27133.
llvm-svn: 264926
The TailDup transform was removed in r138841 in 2011, along with most
of the tests for it. This test, however, was missed. Probably because
it had already been XFAIL'd for 3 years at that point (since r52243!)
and continued to fail when the opt flag for -tailduplicate stopped
being valid.
llvm-svn: 264916
This change prevents the loop vectorizer from vectorizing when all of the vector
types it generates will be scalarized. I've run into this problem on the PPC's QPX
vector ISA, which only holds floating-point vector types. The loop vectorizer
will, however, happily vectorize loops with purely integer computation. Here's
an example:
LV: The Smallest and Widest types: 32 / 32 bits.
LV: The Widest register is: 256 bits.
LV: Found an estimated cost of 0 for VF 1 For instruction: %indvars.iv25 = phi i64 [ 0, %entry ], [ %indvars.iv.next26, %for.body ]
LV: Found an estimated cost of 0 for VF 1 For instruction: %arrayidx = getelementptr inbounds [1600 x i32], [1600 x i32]* %a, i64 0, i64 %indvars.iv25
LV: Found an estimated cost of 0 for VF 1 For instruction: %2 = trunc i64 %indvars.iv25 to i32
LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %2, i32* %arrayidx, align 4
LV: Found an estimated cost of 1 for VF 1 For instruction: %indvars.iv.next26 = add nuw nsw i64 %indvars.iv25, 1
LV: Found an estimated cost of 1 for VF 1 For instruction: %exitcond27 = icmp eq i64 %indvars.iv.next26, 1600
LV: Found an estimated cost of 0 for VF 1 For instruction: br i1 %exitcond27, label %for.cond.cleanup, label %for.body
LV: Scalar loop costs: 3.
LV: Found an estimated cost of 0 for VF 2 For instruction: %indvars.iv25 = phi i64 [ 0, %entry ], [ %indvars.iv.next26, %for.body ]
LV: Found an estimated cost of 0 for VF 2 For instruction: %arrayidx = getelementptr inbounds [1600 x i32], [1600 x i32]* %a, i64 0, i64 %indvars.iv25
LV: Found an estimated cost of 0 for VF 2 For instruction: %2 = trunc i64 %indvars.iv25 to i32
LV: Found an estimated cost of 2 for VF 2 For instruction: store i32 %2, i32* %arrayidx, align 4
LV: Found an estimated cost of 1 for VF 2 For instruction: %indvars.iv.next26 = add nuw nsw i64 %indvars.iv25, 1
LV: Found an estimated cost of 1 for VF 2 For instruction: %exitcond27 = icmp eq i64 %indvars.iv.next26, 1600
LV: Found an estimated cost of 0 for VF 2 For instruction: br i1 %exitcond27, label %for.cond.cleanup, label %for.body
LV: Vector loop of width 2 costs: 2.
LV: Found an estimated cost of 0 for VF 4 For instruction: %indvars.iv25 = phi i64 [ 0, %entry ], [ %indvars.iv.next26, %for.body ]
LV: Found an estimated cost of 0 for VF 4 For instruction: %arrayidx = getelementptr inbounds [1600 x i32], [1600 x i32]* %a, i64 0, i64 %indvars.iv25
LV: Found an estimated cost of 0 for VF 4 For instruction: %2 = trunc i64 %indvars.iv25 to i32
LV: Found an estimated cost of 4 for VF 4 For instruction: store i32 %2, i32* %arrayidx, align 4
LV: Found an estimated cost of 1 for VF 4 For instruction: %indvars.iv.next26 = add nuw nsw i64 %indvars.iv25, 1
LV: Found an estimated cost of 1 for VF 4 For instruction: %exitcond27 = icmp eq i64 %indvars.iv.next26, 1600
LV: Found an estimated cost of 0 for VF 4 For instruction: br i1 %exitcond27, label %for.cond.cleanup, label %for.body
LV: Vector loop of width 4 costs: 1.
...
LV: Selecting VF: 8.
LV: The target has 32 registers
LV(REG): Calculating max register usage:
LV(REG): At #0 Interval # 0
LV(REG): At #1 Interval # 1
LV(REG): At #2 Interval # 2
LV(REG): At #4 Interval # 1
LV(REG): At #5 Interval # 1
LV(REG): VF = 8
The problem is that the cost model here is not wrong, exactly. Since all of
these operations are scalarized, their cost (aside from the uniform ones) are
indeed VF*(scalar cost), just as the model suggests. In fact, the larger the VF
picked, the lower the relative overhead from the loop itself (and the
induction-variable update and check), and so in a sense, picking the largest VF
here is the right thing to do.
The problem is that vectorizing like this, where all of the vectors will be
scalarized in the backend, isn't really vectorizing, but rather interleaving.
By itself, this would be okay, but then the vectorizer itself also interleaves,
and that's where the problem manifests itself. There's aren't actually enough
scalar registers to support the normal interleave factor multiplied by a factor
of VF (8 in this example). In other words, the problem with this is that our
register-pressure heuristic does not account for scalarization.
While we might want to improve our register-pressure heuristic, I don't think
this is the right motivating case for that work. Here we have a more-basic
problem: The job of the vectorizer is to vectorize things (interleaving aside),
and if the IR it generates won't generate any actual vector code, then
something is wrong. Thus, if every type looks like it will be scalarized (i.e.
will be split into VF or more parts), then don't consider that VF.
This is not a problem specific to PPC/QPX, however. The problem comes up under
SSE on x86 too, and as such, this change fixes PR26837 too. I've added Sanjay's
reduced test case from PR26837 to this commit.
Differential Revision: http://reviews.llvm.org/D18537
llvm-svn: 264904
We already try not to truncate PHIs in computeMinimalBitwidths. LoopVectorize can't handle it and we really don't need to, because both induction and reduction PHIs are truncated by other means.
However, we weren't bailing out in all the places we should have, and we ended up by returning a PHI to be truncated, which has caused PR27018.
This fixes PR17018.
llvm-svn: 264852
Prior to this patch, the MemorySSA caching visitor would cache all
calls that it visited. When paired with phi optimization, this can be
problematic. Consider:
define void @foo() {
; 1 = MemoryDef(liveOnEntry)
call void @clobberFunction()
br i1 undef, label %if.end, label %if.then
if.then:
; MemoryUse(??)
call void @readOnlyFunction()
; 2 = MemoryDef(1)
call void @clobberFunction()
br label %if.end
if.end:
; 3 = MemoryPhi(...)
; MemoryUse(?)
call void @readOnlyFunction()
ret void
}
When optimizing MemoryUse(?), we visit defs 1 and 2, so we note to
cache them later. We ultimately end up not being able to optimize
passed the Phi, so we set MemoryUse(?) to point to the Phi. We then
cache the clobbering call for def 1 to be the Phi.
This commit changes this behavior so that we wipe out any calls
added to VisistedCalls while visiting the defs of a phi we couldn't
optimize.
Aside: With this patch, we now can bootstrap clang/LLVM without a
single MemorySSA verifier failure. Woohoo. :)
llvm-svn: 264820
This patch teaches the caching MemorySSA walker a few things:
1. Not to walk Phis we've walked before. It seems that we tried to do
this before, but it didn't work so well in cases like:
define void @foo() {
%1 = alloca i8
%2 = alloca i8
br label %begin
begin:
; 3 = MemoryPhi({%0,liveOnEntry},{%end,2})
; 1 = MemoryDef(3)
store i8 0, i8* %2
br label %end
end:
; MemoryUse(?)
load i8, i8* %1
; 2 = MemoryDef(1)
store i8 0, i8* %2
br label %begin
}
Because we wouldn't put Phis in Q.Visited until we tried to visit them.
So, when trying to optimize MemoryUse(?):
- We would visit 3 above
- ...Which would make us put {%0,liveOnEntry} in Q.Visited
- ...Which would make us visit {%0,liveOnEntry}
- ...Which would make us put {%end,2} in Q.Visited
- ...Which would make us visit {%end,2}
- ...Which would make us visit 3
- ...Which would realize we've already visited everything in 3
- ...Which would make us conservatively return 3.
In the added test-case, (@looped_visitedonlyonce) this behavior would
cause us to give incorrect results. Specifically, we'd visit 4 twice
in the same query, but on the second visit, we'd skip while.cond because
it had been visited, visit if.then/if.then2, and cache "1" as the
clobbering def on the way back.
2. If we try to walk the defs of a {Phi,MemLoc} and see it has been
visited before, just hand back the Phi we're trying to optimize.
I promise this isn't as terrible as it seems. :)
We now insert {Phi,MemLoc} pairs just before walking the Phi's upward
defs. So, we check the cache for the {Phi,MemLoc} pair before checking
if we've already walked the Phi.
The {Phi,MemLoc} pair is (almost?) always guaranteed to have a cache
entry if we've already fully walked it, because we cache as we go.
So, if the {Phi,MemLoc} pair isn't in cache, either:
(a) we must be in the process of visiting it (in which case, we can't
give a better answer in a cache-as-we-go DFS walker)
(b) we visited it, but didn't cache it on the way back (...which seems
to require `ModifyingAccess` to not dominate `StartingAccess`,
so I'm 99% sure that would be an error. If it's not an error, I
haven't been able to get it to happen locally, so I suspect it's
rare.)
- - - - -
As a consequence of this change, we no longer skip upward defs of phis,
so we can kill the `VisitedOnlyOne` check. This gives us better accuracy
than we had before, at the cost of potentially doing a bit more work
when we have a loop.
llvm-svn: 264814
This is effectively NFC, minus the renaming of the options
(-cyclone-prefetch-distance -> -prefetch-distance).
The change was requested by Tim in D17943.
llvm-svn: 264806
During ADCE, track which debug info scopes still have live references
from the code, and delete debug info intrinsics for the dead ones.
These intrinsics describe the locations of variables (in registers or
stack slots). If there's no code left corresponding to a variable's
scope, then there's no way to reference the variable in the debugger and
it doesn't matter what its value is.
I add a DEBUG printout when the described location in an SSA register,
in case it helps some trying to track down why locations get lost.
However, we still delete these; the scope itself isn't attached to any
real code, so the ship has already sailed.
llvm-svn: 264800
Add function soft attribute to the generation of Jump Tables in CodeGen
as initial step towards clang support of gcc's no-jump-table support
Reviewers: hans, echristo
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D18321
llvm-svn: 264756
When eliminating or merging almost empty basic blocks, the existence of non-trivial PHI nodes
is currently used to recognize potential loops of which the block is the header and keep the block.
However, the current algorithm fails if the loops' exit condition is evaluated only with volatile
values hence no PHI nodes in the header. Especially when such a loop is an outer loop of a nested
loop, the loop is collapsed into a single loop which prevent later optimizations from being
applied (e.g., transforming nested loops into simplified forms and loop vectorization).
The patch augments the existing PHI node-based check by adding a pre-test if the BB actually
belongs to a set of loop headers and not eliminating it if yes.
llvm-svn: 264697
A DICompileUnit that is not listed in llvm.dbg.cu will cause assertion
failures and/or crashes in the backend. The Verifier should reject this.
rdar://problem/25369499
llvm-svn: 264657
When eliminating or merging almost empty basic blocks, the existence of non-trivial PHI nodes
is currently used to recognize potential loops of which the block is the header and keep the block.
However, the current algorithm fails if the loops' exit condition is evaluated only with volatile
values hence no PHI nodes in the header. Especially when such a loop is an outer loop of a nested
loop, the loop is collapsed into a single loop which prevent later optimizations from being
applied (e.g., transforming nested loops into simplified forms and loop vectorization).
The patch augments the existing PHI node-based check by adding a pre-test if the BB actually
belongs to a set of loop headers and not eliminating it if yes.
llvm-svn: 264596
Summary:
Add a statistic to count the number of imported functions. Also, add a
new -print-imports option to emit a trace of imported functions, that
works even for an NDEBUG build.
Note that emitOptimizationRemark does not work for the above printing as
it expects a Function object and DebugLoc, neither of which we have
with summary-based importing.
This is part 2 of D18487, the first part was committed separately as
r264536.
Reviewers: joker.eph
Subscribers: llvm-commits, joker.eph
Differential Revision: http://reviews.llvm.org/D18487
llvm-svn: 264537
Reject the following IR as malformed (assuming that %entry, %next are
not in a loop):
next:
%y = phi i32 [ 0, %entry ]
%x = phi i32 [ %y, %entry ]
Such PHI nodes came up in PR26718. While there was no consensus on
whether or not this is valid IR, most opinions on that bug and in a
discussion on the llvm-dev mailing list tended towards a
"strict interpretation" (term by Joseph Tremoulet) of PHI node uses.
Also, the language reference explicitly states that "the use of each
incoming value is deemed to occur on the edge from the corresponding
predecessor block to the current block" and
`DominatorTree::dominates(Instruction*, Use&)` uses this definition as
well.
For the code mentioned in PR15384, clang does not compile to such PHIs
(anymore?). The test case still hangs when replacing `%tmp6` with `%tmp`
in revisions before r176366 (where PR15384 has been fixed). The
occurrence of %tmp6 therefore was probably unintentional. Its value is
not used except in other PHIs.
Reviewers: majnemer, reames, JosephTremoulet, bkramer, grosser, jdoerfert, kparzysz, sanjoy
Differential Revision: http://reviews.llvm.org/D18443
llvm-svn: 264528
Summary:
Now that the summary contains the full reference/call graph, we can
replace the existing function importer that loads and inspect the IR
to iteratively walk the call graph by a traversal based purely on the
summary information. Decouple the actual importing decision from any
IR manipulation.
Reviewers: tejohnson
Subscribers: llvm-commits, joker.eph
Differential Revision: http://reviews.llvm.org/D18343
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 264503
A release fence acts as a publication barrier for stores within the current thread to become visible to other threads which might observe the release fence. It does not require the current thread to observe stores performed on other threads. As a result, we can allow store-load and load-load forwarding across a release fence.
We choose to be much more conservative about stores. In theory, nothing prevents us from shifting a store from after a release fence to before it, and then eliminating the preceeding (previously fenced) store. Doing this without actually moving the second store is likely also legal, but we chose to be conservative at this time.
The LangRef indicates only atomic loads and stores are effected by fences. This patch chooses to be far more conservative then that.
This is the GVN companion to http://reviews.llvm.org/D11434 which applied the same logic in EarlyCSE and has been baking in tree for a while now.
Differential Revision: http://reviews.llvm.org/D11436
llvm-svn: 264472
I didn't notice any significant changes in the actual checks here;
all of these tests already used FileCheck, so a script can batch
update them in one shot.
This commit is just to show the value of automating this process:
We have uniform formatting as opposed to a mish-mash of check
structure that changes based on individual prefs and the current
fashion. This makes it simpler to update when we find a bug or
make an enhancement.
llvm-svn: 264457
This changes RS4GC to lower calls to ``@llvm.experimental.deoptimize``
to gc.statepoints wrapping ``__llvm_deoptimize``, and changes
``callsGCLeafFunction`` to recognize ``@llvm.experimental.deoptimize``
as a non GC leaf function.
I've had to hard code the ``"__llvm_deoptimize"`` name in
RewriteStatepointsForGC; since ``TargetLibraryInfo`` is available only
during codegen. This isn't without precedent in the codebase, so I'm
not overtly concerned.
llvm-svn: 264456
We try to hoist the insertion point as high as possible to encourage
sharing. However, we must be careful not to hoist into a catchswitch as
it is both an EHPad and a terminator.
llvm-svn: 264344
isDependenceDistanceOfOne asserts that the store and the load access
through the same type. This function is also used by
removeDependencesFromMultipleStores so we need to make sure we filter
out mismatching types before reaching this point.
Now we do this when the initial candidates are gathered.
This is a refinement of the fix made in r262267.
Fixes PR27048.
llvm-svn: 264313
There are a few bugs in the walker that this patch addresses.
Primarily:
- Caching can break when we have multiple BBs without phis
- We weren't optimizing some phis properly
- Because of how the DFS iterator works, there were times where we
wouldn't cache any results of our DFS
I left the test cases with FIXMEs in, because I'm not sure how much
effort it will take to get those to work (read: We'll probably
ultimately have to end up redoing the walker, or we'll have to come up
with some creative caching tricks), and more test coverage = better.
Differential Revision: http://reviews.llvm.org/D18065
llvm-svn: 264180
CGP modifies the domtree in some cases, so saying that it preserves the
domtree is a lie. We'll be able to selectively preserve it with the new
pass manager.
Differential Revision: http://reviews.llvm.org/D16893
llvm-svn: 264099
When you have multiple LCSSA (single-operand) PHIs that are converted
into two-operand PHIs due to versioning, only assert that the PHI
currently being converted has a single operand. I.e. we don't want to
check PHIs that were converted earlier in the loop.
Fixes PR27023.
Thanks to Karl-Johan Karlsson for the minimized testcase!
llvm-svn: 264081
It's a bug fix.
For rerolled loops SE trip count remains unchanged. It leads to incorrect work of the next passes.
My patch just resets SE info for rerolled loop forcing SE to re-evaluate it next time it requested.
I also added a verifier call in the exisitng test to be sure no invalid SE data remain. Without my fix this test would fail with -verify-scev.
Differential Revision: http://reviews.llvm.org/D18316
llvm-svn: 264051
If we have a BB with only MemoryDefs, live-in calculations will ignore
it. This means we get results like this:
define void @foo(i8* %p) {
; 1 = MemoryDef(liveOnEntry)
store i8 0, i8* %p
br i1 undef, label %if.then, label %if.end
if.then:
; 2 = MemoryDef(1)
store i8 1, i8* %p
br label %if.end
if.end:
; 3 = MemoryDef(1)
store i8 2, i8* %p
ret void
}
...When there should be a MemoryPhi in the `if.end` BB.
This patch fixes that behavior.
llvm-svn: 263991
Summary:
replaceCongruentIVs can break LCSSA when trying to replace IV increments
since it tries to replace all uses of a phi node with another phi node
while both of the phi nodes are not necessarily in the processed loop.
This will cause an assert in IndVars.
To fix this, we add a check to make sure that the replacement maintains
LCSSA.
Reviewers: sanjoy
Subscribers: mzolotukhin, llvm-commits
Differential Revision: http://reviews.llvm.org/D18266
llvm-svn: 263941
The sinpi/cospi can be replaced with sincospi to remove unnecessary
computations. However, we need to make sure that the calls are within
the same function!
This fixes PR26993.
llvm-svn: 263875
Summary:
ThinLTO is relying on linkInModule to import selected function.
However a lot of "magic" was hidden in linkInModule and the IRMover,
who would rename and promote global variables on the fly.
This is moving to an approach where the steps are decoupled and the
client is reponsible to specify the list of globals to import.
As a consequence some test are changed because they were relying on
the previous behavior which was importing the definition of *every*
single global without control on the client side.
Now the burden is on the client to decide if a global has to be imported
or not.
Reviewers: tejohnson
Subscribers: joker.eph, llvm-commits
Differential Revision: http://reviews.llvm.org/D18122
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 263863
While not strictly necessary, since we don't support large integer
types, this avoids bugs due to silent truncation from uint64_t to a
32-bit unsigned (e.g. DL.isLegalInteger(DL.getTypeSizeInBits(Ty) )
This fixes PR26972.
Differential Revision: http://reviews.llvm.org/D18258
llvm-svn: 263850
The loop on IVOperand's incoming values assumes IVOperand to be an
induction variable on the loop over which `S Pred X` is invariant;
otherwise loop invariant incoming values to IVOperand are not guaranteed
to dominate the comparision.
This fixes PR26973.
llvm-svn: 263827
Summary:
It can hurt performance to prefetch ahead too much. Be conservative for
now and don't prefetch ahead more than 3 iterations on Cyclone.
Reviewers: hfinkel
Subscribers: llvm-commits, mzolotukhin
Differential Revision: http://reviews.llvm.org/D17949
llvm-svn: 263772
Summary:
And use this TTI for Cyclone. As it was explained in the original RFC
(http://thread.gmane.org/gmane.comp.compilers.llvm.devel/92758), the HW
prefetcher work up to 2KB strides.
I am also adding tests for this and the previous change (D17943):
* Cyclone prefetching accesses with a large stride
* Cyclone not prefetching accesses with a small stride
* Generic Aarch64 subtarget not prefetching either
Reviewers: hfinkel
Subscribers: aemerson, rengolin, llvm-commits, mzolotukhin
Differential Revision: http://reviews.llvm.org/D17945
llvm-svn: 263771
Summary:
Use the new LoopVersioning facility (D16712) to add noalias metadata in
the vector loop if we versioned with memchecks. This can enable some
optimization opportunities further down the pipeline (see the included
test or the benchmark improvement quoted in D16712).
The test also covers the bug I had in the initial version in D16712.
The vectorizer did not previously use LoopVersioning. The reason is
that the vectorizer performs its transformations in single shot. It
creates an empty single-block vector loop that it then populates with
the widened, if-converted instructions. Thus creating an intermediate
versioned scalar loop seems wasteful.
So this patch (rather than bringing in LoopVersioning fully) adds a
special interface to LoopVersioning to allow the vectorizer to add
no-alias annotation while still performing its own versioning.
As the vectorizer propagates metadata from the instructions in the
original loop to the vector instructions we also check the pointer in
the original instruction and see if LoopVersioning can add no-alias
metadata based on the issued memchecks.
Reviewers: hfinkel, nadav, mzolotukhin
Subscribers: mzolotukhin, llvm-commits
Differential Revision: http://reviews.llvm.org/D17191
llvm-svn: 263744
Summary:
If we decide to version a loop to benefit a transformation, it makes
sense to record the now non-aliasing accesses in the newly versioned
loop. This allows non-aliasing information to be used by subsequent
passes.
One example is 456.hmmer in SPECint2006 where after loop distribution,
we vectorize one of the newly distributed loops. To vectorize we
version this loop to fully disambiguate may-aliasing accesses. If we
add the noalias markers, we can use the same information in a later DSE
pass to eliminate some dead stores which amounts to ~25% of the
instructions of this hot memory-pipeline-bound loop. The overall
performance improves by 18% on our ARM64.
The scoped noalias annotation is added in LoopVersioning. The patch
then enables this for loop distribution. A follow-on patch will enable
it for the vectorizer. Eventually this should be run by default when
versioning the loop but first I'd like to get some feedback whether my
understanding and application of scoped noalias metadata is correct.
Essentially my approach was to have a separate alias domain for each
versioning of the loop. For example, if we first version in loop
distribution and then in vectorization of the distributed loops, we have
a different set of memchecks for each versioning. By keeping the scopes
in different domains they can conveniently be defined independently
since different alias domains don't affect each other.
As written, I also have a separate domain for each loop. This is not
necessary and we could save some metadata here by using the same domain
across the different loops. I don't think it's a big deal either way.
Probably the best is to review the tests first to see if I mapped this
problem correctly to scoped noalias markers. I have plenty of comments
in the tests.
Note that the interface is prepared for the vectorizer which needs the
annotateInstWithNoAlias API. The vectorizer does not use LoopVersioning
so we need a way to pass in the versioned instructions. This is also
why the maps have to become part of the object state.
Also currently, we only have an AA-aware DSE after the vectorizer if we
also run the LTO pipeline. Depending how widely this triggers we may
want to schedule a DSE toward the end of the regular pass pipeline.
Reviewers: hfinkel, nadav, ashutosh.nema
Subscribers: mssimpso, aemerson, llvm-commits, mcrosier
Differential Revision: http://reviews.llvm.org/D16712
llvm-svn: 263743
Easwaran Raman