```
always_inline foo() { }
bar () {
noinline foo();
}
```
We should prefer call site attribute over attribute on decl. This is fix for AlwaysInliner, similar fix is needed for normal Inliner (follow up).
Related to https://reviews.llvm.org/D119061
Reviewed By: aeubanks
Differential Revision: https://reviews.llvm.org/D119553
In addition to the self-recursion check, also check whether there
is more than one node in the SCC, which implies that there is a
larger cycle. I believe checking SCC structure (rather than
something like norecurse) is the right thing to do here, because
this is specifically about preventing infinite loops over the SCC.
Fixes https://github.com/llvm/llvm-project/issues/42028.
Differential Revision: https://reviews.llvm.org/D119418
Minor efficiency fix. There is no reason to perform the same set lookup
repeatedly in the inner loop as it is invariant there.
Differential Revision: https://reviews.llvm.org/D119474
New users might want to check bins without a load or store instruction
at hand. Since we use those instructions only to find the offset and
size of the access anyway, we can expose an offset and size interface
to the outside world as well.
This commit mainly moves code around and exposes a class (OffsetAndSize)
as well as a method forallInterferingAccesses in AAPointerInfo.
Differential Revision: https://reviews.llvm.org/D119249
The oversight caused us to ignore call sites that are effectively dead
when we computed reachability (or more precise the call edges of a
function). The problem is that loads in the readonly callee might depend
on stores prior to the callee. If we do not track the call edge we
mistakenly assumed the store before the call cannot reach the load.
The problem is nicely visible in:
`llvm/test/Transforms/Attributor/ArgumentPromotion/basictest.ll`
Caused by D118673.
Fixes https://github.com/llvm/llvm-project/issues/53726
When we privatize a pointer (~argument promotion) we introduce new
private allocas as replacement. These need to be placed in the alloca
address space as later passes cannot properly deal with them otherwise.
Fixes https://github.com/llvm/llvm-project/issues/53725
This rewrites ArgPromotion to be based on offsets rather than GEP
structure. We inspect all loads at constant offsets and remember
which types are loaded at which offsets. Then we promote based on
those types.
This generalizes ArgPromotion to work with bitcasted loads, and
is compatible with opaque pointers.
This patch also fixes incorrect handling of alignment during
argument promotion. Previously, the implementation only checked
that the pointer is dereferenceable, but was happy to speculate
overaligned loads. (I would have fixed this separately in advance,
but I found this hard to do with the previous implementation
approach).
Differential Revision: https://reviews.llvm.org/D118685
This patch replaces the function we emit the remark on when we run into
the fix-point limit. Previously we got a function to emit a remark on
from the worklist's associated function. However, the worklist may not
always have an associated function in the case of global variables.
Replace this with the function set, and if there are no functions don't
emit the remark.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D119248
Before walking all the callers, check whether we have a
dereferenceable attribute directly on the argument.
Also make it clearer that the code currently does not treat
alignment correctly.
The helper `Attributor::checkForAllReturnedValuesAndReturnInsts`
simplifies the returned value optimistically. In `AAUndefinedBehavior`
we cannot use such optimistic values when deducing UB. As a result, we
assumed UB for the return value of a function because we initially
(=optimistically) thought the function return is `undef`. While we later
adjusted this properly, the `AAUndefinedBehavior` was under the
impression the return value is "known" (=fix) and could never change.
To correct this we use `Attributor::checkForAllInstructions` and then
manually to perform simplification of the return value, only allowing
known values to be used. This actually matches the other UB deductions.
Fixes#53647
I'm seeing ext-tsp helps CSSPGO for our intern large benchmarks so I'm turning on it for CSSPGO. For non-CS AutoFDO, ext-tsp doesn't seem to help, probably because of lower profile counts quality.
Reviewed By: wenlei
Differential Revision: https://reviews.llvm.org/D119048
Changes the remark to emit on the function call that captures the globalized
variable instead of the globalized variable itself. The user should be able to
see which variable it was in the argument list of the function.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106980
This is a fix for a use-after-free found by the address sanitizer when
compiling GCC: https://github.com/llvm/llvm-project/issues/52821
The Function Specialization pass may remove instructions, cached
inside the PredicateBase class, which are later being dereferenced
from the SCCPInstVisitor class. To prevent the dangling references
I am lazily deleting the dead instructions after the Solver has run.
Differential Revision: https://reviews.llvm.org/D118591
Based on the output of include-what-you-use.
This is a big chunk of changes. It is very likely to break downstream code
unless they took a lot of care in avoiding hidden ehader dependencies, something
the LLVM codebase doesn't do that well :-/
I've tried to summarize the biggest change below:
- llvm/include/llvm-c/Core.h: no longer includes llvm-c/ErrorHandling.h
- llvm/IR/DIBuilder.h no longer includes llvm/IR/DebugInfo.h
- llvm/IR/IRBuilder.h no longer includes llvm/IR/IntrinsicInst.h
- llvm/IR/LLVMRemarkStreamer.h no longer includes llvm/Support/ToolOutputFile.h
- llvm/IR/LegacyPassManager.h no longer include llvm/Pass.h
- llvm/IR/Type.h no longer includes llvm/ADT/SmallPtrSet.h
- llvm/IR/PassManager.h no longer includes llvm/Pass.h nor llvm/Support/Debug.h
And the usual count of preprocessed lines:
$ clang++ -E -Iinclude -I../llvm/include ../llvm/lib/IR/*.cpp -std=c++14 -fno-rtti -fno-exceptions | wc -l
before: 6400831
after: 6189948
200k lines less to process is no that bad ;-)
Discourse thread on the topic: https://llvm.discourse.group/t/include-what-you-use-include-cleanup
Differential Revision: https://reviews.llvm.org/D118652
Generalize D99629 for ELF. A default visibility non-local symbol is preemptible
in a -shared link. `isInterposable` is an insufficient condition.
Moreover, a non-preemptible alias may be referenced in a sub constant expression
which intends to lower to a PC-relative relocation. Replacing the alias with a
preemptible aliasee may introduce a linker error.
Respect dso_preemptable and suppress optimization to fix the abose issues. With
the change, `alias = 345` will not be rewritten to use aliasee in a `-fpic`
compile.
```
int aliasee;
extern int alias __attribute__((alias("aliasee"), visibility("hidden")));
void foo() { alias = 345; } // intended to access the local copy
```
While here, refine the condition for the alias as well.
For some binary formats like COFF, `isInterposable` is a sufficient condition.
But I think canonicalization for the changed case has little advantage, so I
don't bother to add the `Triple(M.getTargetTriple()).isOSBinFormatELF()` or
`getPICLevel/getPIELevel` complexity.
For instrumentations, it's recommended not to create aliases that refer to
globals that have a weak linkage or is preemptible. However, the following is
supported and the IR needs to handle such cases.
```
int aliasee __attribute__((weak));
extern int alias __attribute__((alias("aliasee")));
```
There are other places where GlobalAlias isInterposable usage may need to be
fixed.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D107249
A call base can be a floating value if we talk about the instruction and
not the return value. This distinction was not made before but is
important for liveness, e.g., a call site return value might be unused
(=dead) but the call site is not.
To make usage easier (compared to the many reachability related AAs),
this patch introduces a helper API, `AA::isPotentiallyReachable`, which
performs all the necessary steps. It also does the "backwards"
reachability (see D106720) as that simplifies the AA a lot (backwards
queries were somewhat different from the other query resolvers), and
ensures we use cached values in every stage.
To test inter-procedural reachability in a reasonable way this patch
includes an extension to `AAPointerInfo::forallInterferingWrites`.
Basically, we can exclude writes if they cannot reach a load "during the
lifetime" of the allocation. That is, we need to go up the call graph to
determine reachability until we can determine the allocation would be
dead in the caller. This leads to new constant propagations (through
memory) in `value-simplify-pointer-info-gpu.ll`.
Note: The new code contains plenty debug output to determine how
reachability queries are resolved.
Parts extracted from D110078.
Differential Revision: https://reviews.llvm.org/D118673
D106720 introduced features that did not work properly as we could add
new queries after a fixpoint was reached and which could not be answered
by the information gathered up to the fixpoint alone.
As an alternative to D110078, which forced eager computation where we
want to continue to be lazy, this patch fixes the problem.
QueryAAs are AAs that allow lazy queries during their lifetime. They are
never fixed if they have no outstanding dependences and always run as
part of the updates in an iteration. To determine if we are done, all
query AAs are asked if they received new queries, if not, we only need
to consider updated AAs, as before. If new queries are present we go for
another iteration.
Differential Revision: https://reviews.llvm.org/D118669
This patch implement instruction reachability for AAFunctionReachability
attribute. It is used to tell if a certain instruction can reach a function
transitively.
NOTE: I created a new commit based of D106720 and set the author back to
Kuter. Other metadata, etc. is wrong. I also addressed the
remaining review comments and fixed the unit test.
Differential Revision: https://reviews.llvm.org/D106720
We missed out on AANoRecurse in the module pass because we had no call
graph. With AAFunctionReachability we can simply ask if the function may
reach itself.
Differential Revision: https://reviews.llvm.org/D110099
genericValueTraversal can look through arguments and allow value
simplification across function boundaries. In fact, the latter already
happened unchecked. With this change we allow the user of
genericValueTraversal to opt-out of interprocedural traversal if
required. We explicitly look through arguments now which helps to do
various things, incl. the propagation of constants into OpenMP parallel
regions (on the host).
This fixes a conceptual problem with our AAIsDead usage which conflated
call site liveness with call site return value liveness. Without the
fix tests would obviously miscompile as we make genericValueTraversal
more powerful (in a follow up). The effects on the tests are mixed but
mostly marginal. The most prominent one is the lack of `noreturn` for
functions. The reason is that we make entire blocks live at the same
time (for time reasons). Now that we actually look at the block
liveness, which we need to do, the return instructions are live and
will survive. As an example, `noreturn_async.ll` has been modified
to retain the `noreturn` even with block granularity. We could address
this easily but there is little need in practice.
We have two attributes that can answer readnone queries. While there is
a dependence between them, it seems best to not force the users to know
what AA to ask. The helpers also allow to check for readonly nicely.
Test changes show where we now deduce readnone but haven't before,
mostly because we only asked AAMemoryBehavior and not AAMemoryLocation.
AANoAlias has not been ported to the new API yet.
Since D104432 we can look through memory by analyzing all writes that
might interfere with a load. This patch provides some logic to exclude
writes that cannot interfere with a location, due to CFG reasoning.
We make sure to avoid multi-thread write-read situations properly while
we ignore writes that cannot reach a load or writes that will be
overwritten before the load is reached.
Differential Revision: https://reviews.llvm.org/D106397
No-sync is a property that we need in more places as complex
transformations emerge. To simplify the query we provide an
`AA::isNoSyncInst` helper now and expose two existing helpers through
the `AANoSync` class.
Patch originally by Giorgis Georgakoudis (@ggeorgakoudis), typos and
bugs introduced later by me.
This patch allows us to remove redundant barriers if they are part
of a "consecutive" pair of barriers in a basic block with no impacted
memory effect (read or write) in-between them. Memory accesses to
local (=thread private) or constant memory are allowed to appear.
Technically we could also allow any other memory that is not used to
share information between threads, e.g., the result of a malloc that
is also not captured. However, it will be easier to do more reasoning
once the code is put into an AA. That will also allow us to look through
phis/selects reasonably. At that point we should also deal with calls,
barriers in different blocks, and other complexities.
Differential Revision: https://reviews.llvm.org/D118002
We used to remove noinline from known OpenMP runtime functions (which
are declared in OMPKinds.td). Now we remove noinline from all functions
with the proper prefixes: __kmpc, _ZN4_OMP (= namespace omp), omp_
Due to some complications with lifetime, and assume-like intrinsics, intrinsics were not included as outlinable instructions. This patch opens up most intrinsics, excluding lifetime and assume-like intrinsics, to be outlined. For similarity, it is required that the intrinsic IDs, and the intrinsics names match exactly, as well as the function type. This puts intrinsics in a different class than normal call instructions (https://reviews.llvm.org/D109448), where the name will no longer have to match.
This also adds an additional command line flag debug option to disable outlining intrinsics.
Recommit of: 8de76bd569
Adds extra checking of intrinsic function calls names to avoid taking the address of intrinsic calls when extracting function calls.
Reviewers: paquette, jroelofs
Differential Revision: https://reviews.llvm.org/D109450
Summary:
This patch modifies code generation in OpenMPIRBuilder to pass arguments
to the parallel region outlined function in an aggregate (struct),
besides the global_tid and bound_tid arguments. It depends on the
updated CodeExtractor (see D96854) for support. It mirrors functionality
of Clang codegen (see D102107).
Differential Revision: https://reviews.llvm.org/D110114
We use the same similarity scheme we used for branch instructions for phi nodes, and allow them to be outlined. There is not a lot of special handling needed for these phi nodes when outlining, as they simply act as outputs. The code extractor does not currently allow for non entry blocks within the extracted region to have predecessors, so there are not conflicts to handle with respect to predecessors no longer contained in the function.
Recommit of 515eec3553
Reviewers: paquette
Differential Revision: https://reviews.llvm.org/D106997
Due to some complications with lifetime, and assume-like intrinsics, intrinsics were not included as outlinable instructions. This patch opens up most intrinsics, excluding lifetime and assume-like intrinsics, to be outlined. For similarity, it is required that the intrinsic IDs, and the intrinsics names match exactly, as well as the function type. This puts intrinsics in a different class than normal call instructions (https://reviews.llvm.org/D109448), where the name will no longer have to match.
This also adds an additional command line flag debug option to disable outlining intrinsics.
Reviewers: paquette, jroelofs
Differential Revision: https://reviews.llvm.org/D109450
Problem: Migration to new PM broke flatten attribute.
This is one use case why LLVM should support inlining call-site with alwaysinline. The flatten attribute is nowdays broken, so we should either land patch like this one or remove everything related to flatten attribute from Clang.
Second use case is something like "per call site inlining intrinsics" to control inlining even more; mentioned in
https://lists.llvm.org/pipermail/cfe-dev/2018-September/059232.html
Fixes https://github.com/llvm/llvm-project/issues/53360
Reviewed By: aeubanks
Differential Revision: https://reviews.llvm.org/D117965
The outliner currently requires that function calls not be indirect calls, and have that the function name, and function type must match, as well as other attributes such as calling conventions. This patch treats called functions as values, and just another operand, and named function calls as constants. This allows functions to be treated like any other constant, or input and output into the outlined functions.
There are also debugging flags added to enforce the old behaviors where indirect calls not be allowed, and to enforce the old rule that function calls names must also match.
Reviewers: paquette, jroelofs
Differential Revision: https://reviews.llvm.org/D109448
In addition to having multiple exit locations, there can be multiple blocks leading to the same exit location, which results in a potential phi node. If we find that multiple blocks within the region branch to the same block outside the region, resulting in a phi node, the code extractor pulls this phi node into the function and uses it as an output.
We make sure that this phi node is given an output slot, and that the two values are removed from the outputs if they are not used anywhere else outside of the region. Across the extracted regions, the phi nodes are combined into a single block for each potential output block, similar to the previous patch.
Reviewers: paquette
Differential Revision: https://reviews.llvm.org/D106995
Instead use either Type::getPointerElementType() or
Type::getNonOpaquePointerElementType().
This is part of D117885, in preparation for deprecating the API.
Currenly we push some variables to a global constant containing shared
memory as an optimization. This generated constant had internal linkage
and should not have collided with any known identifiers in the
translation unit. However, there have been observed cases of this
optimiztaion unintentionally colliding with undocumented PTX
identifiers. This patch adds a suffix to the created globals to
hopefully bypass this.
Depends on D118059
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D118068
Previously in OpenMPOpt we did not correctly inherit the calling
convention of the callee when creating new OpenMP runtime calls. This
created issues when the calling convention was changed during
`GlobalOpt` but a new call was creating without the correct calling
convention. This lead to the call being replaced with a poison value in
`InstCombine` due to undefined behaviour and causing large portions of
the program to be incorrectly eliminated. This patch correctly inherits
the existing calling convention from the callee.
Reviewed By: tianshilei1992, jdoerfert
Differential Revision: https://reviews.llvm.org/D118059
This relies on existing APIs and avoids accessing the pointer
element type. The alternative would be to extend getPointerOperand()
to also return the accessed type, but I figured going through
MemoryLocation would be cleaner.
Differential Revision: https://reviews.llvm.org/D117868
The old method to avoid unconstrained expansion of the constant range in
a loop did not work as soon as there were multiple instructions in
between the phi and its input. We now take a generic approach and limit
the number of updates as a fallback. The old method is kept as it
catches "the common case" early.
While we might know the value if an ICV at a getter position it is not
always clear that we can simply use it. Verify the value is valid first
to avoid invalid IR.
Fixes#53300.
Fixing ICE when partial inline tries to deal with blockaddress uses of function which is typical for asm-goto/callbr. We ran into this with PGO multi-region partial inline.
Differential Revision: https://reviews.llvm.org/D117509
AAPointerInfo currently bails on constant expression GEPs with
notional overindexing. I don't think this is necessary, as the
following code handling GEPOperator will deal with arbitrary
indices appropriately.
Differential Revision: https://reviews.llvm.org/D117203
The global state refers to the number of the nodes currently in the
module, and the number of direct calls between nodes, across the
module.
Node counts are not a problem; edge counts are because we want strictly
the kind of edges that affect inlining (direct calls), and that is not
easily obtainable without iteration over the whole module.
This patch avoids relying on analysis invalidation because it turned out
to be too aggressive in some cases. It leverages the fact that Node
objects are stable - they do not get deleted while cgscc passes are
run over the module; and cgscc pass manager invariants.
Reviewed By: aeubanks
Differential Revision: https://reviews.llvm.org/D115847
We can generalize the malloc-to-global transform for other allocation functions which are both a) removable, and b) have a known initialization value.
One subtlety that I want to point out - mostly because I hadn't realized it was true until I took a closer look - is that the existing code doesn't prove that initialization/malloc happens only once. The initialization function can be called multiple times. This is correct without special handling for malloc as undef can map to any value previously written, but a non-undef initializing allocation it means we may end up memseting the new global repeatedly. In particular, this means it's not legal to fold the memset into the initializer of the global.
Differential Revision: https://reviews.llvm.org/D117503
This was a last-minute addition to D117249, and of course I ended
up inverting the condition in a way that caused an uninitialized
memory read.
I've dropped it entirely, as I don't think we actually care whether
the size is zero or not here. The previous code wasn't checking
this either.
The malloc to global transform currently determines the type of the
global by looking at bitcasts of the malloc. This is limited (the
transform fails if there are multiple different types) and
incompatible with opaque pointers.
My initial approach was to construct an appropriate struct type
based on usage in loads/stores. What this patch does instead is
to always create an [i8 x AllocSize] global, without trying to
guess types at all.
This does mean that other transforms that require a certain global
type may break. I fixed two of these in D117034 and D117223, which
I believe should be sufficient to avoid regressions. In particular,
the global SRA change should end up splitting the global into
naturally-typed sub-globals, at which point all other optimizations
should work.
Differential Revision: https://reviews.llvm.org/D117092
Currently global SRA uses the GEP structure to determine how to
split the global. This patch instead analyses the loads and stores
that are performed on the global, and collects which types are used
at which offset, and then splits the global according to those.
This is both more general, and works fine with opaque pointers.
This is also closer to how ordinary SROA is performed.
Differential Revision: https://reviews.llvm.org/D117223
This completes removal of the isXLike queries, and depends on a whole series of earlier patches which have already landed.
Differential Revision: https://reviews.llvm.org/D117242
The basic idea is that we can parameterize the getObjectSize implementation with a callback which lets us replace the operand before analysis if desired. This is what Attributor is doing during it's abstract interpretation, and allows us to have one copy of the code.
Note this is not NFC for two reasons:
* The existing attributor code is wrong. (Well, this is under-specified to be honest, but at least inconsistent.) The intermediate math needs to be done in the index type of the pointer space. Imagine e.g. i64 arguments in a 32 bit address space.
* I did not preserve the behavior in getAPInt where we return 0 for a partially analyzed value. This looks simply wrong in the original code, and nothing test wise contradicts that.
Differential Revision: https://reviews.llvm.org/D117241
The existing code duplicated the same concern in two places, and (weirdly) changed the inference of the allocation size based on whether we could meet the alignment requirement. Instead, just directly check the allocation requirement.
Rewrite the calloc specific handling in heap-to-stack to allow arbitrary init values. The basic problem being solved is that if an allocation is initilized to anything other than zero, this must be explicitly done for the formed alloca as well.
This covers the calloc case today, but once a couple of earlier guards are removed in this code, downstream allocators with other init values could also be handled.
Inspired by discussion on D116971
GlobalOpt can optimize a global with undef initializer and a single
store to put the stored value into the initializer instead. Currently,
this requires the type of the global and the store to match.
This patch extends support to cases with different types (but same
size), in which case we create a new global to replace the old one.
Differential Revision: https://reviews.llvm.org/D117034
This avoids the InlineAdvisor carrying the responsibility of deleting
Function objects. We use LazyCallGraph::Node objects instead, which are
stable in memory for the duration of the Module-wide performance of CGSCC
passes started under the same ModuleToPostOrderCGSCCPassAdaptor (which
is the case here)
Differential Revision: https://reviews.llvm.org/D116964
If we look at potentially interfering accesses we need to ensure the
"IsExact" flag is set appropriately. Accesses that have an "unknown"
size or offset cannot be exact matches and we missed to flag that.
Error and test reported by Serguei N. Dmitriev.
We currently have two similar implementations of this concept:
isNoAliasCall() only checks for the noalias return attribute.
isNoAliasFn() also checks for allocation functions.
We should switch to only checking the attribute. SLC is responsible
for inferring the noalias return attribute for non-new allocation
functions (with a missing case fixed in
348bc76e35).
For new, clang is responsible for setting the attribute,
if -fno-assume-sane-operator-new is not passed.
Differential Revision: https://reviews.llvm.org/D116800
If we have multiple references into a map we need to ensure the ones
created late do not invalidate the ones created early. To do that we
need to make sure all but the first are not modifying the map, hence
for them the keys have to be present already.
Fixes#52875.
This is a reoccuring pattern, we can consolidate three copies into one. The main motivation is to reduce usages of isMallocLike.
The original commit (which was quickly reverted) didn't account for the allocation function could be an invoke, test coverage for that case added in this commit.
The original commit was expected to be NFC, but I didn't account for the fact that invokes could be considered allocation functions. Interestingly, only one builder caught the problem.
The naming has come up as a source of confusion in several recent reviews. onlyWritesMemory is consist with onlyReadsMemory which we use for the corresponding readonly case as well.
There are a number of places that specially handle loads from a
uniform value where all the bits are the same (zero, one, undef,
poison), because we a) don't care about the load offset in that
case b) it bypasses casts that might not be legal generally but
do work with uniform values.
We had multiple implementations of this, with a different set of
supported values each time. This replaces two usages with a more
complete helper. Other usages will be replaced separately, because
they have larger impact.
This is part of D115924.
This builds on the code from D114963, and extends it to handle calls both direct and indirect. With the revised code structure (from series of previously landed NFCs), this is pretty straight forward.
One thing to note is that we can not infer writeonly for arguments which might be captured. If the pointer can be read back by the caller, and then read through, we have no way to track that. This is the same restriction we have for readonly, except that we get no mileage out of the "callee can be readonly" exception since a writeonly param on a readonly function is either a) readnone or b) UB. This means we can't actually infer much unless nocapture has already been inferred.
Differential Revision: https://reviews.llvm.org/D115003
This class is solely used as a lightweight and clean way to build a set of
attributes to be removed from an AttrBuilder. Previously AttrBuilder was used
both for building and removing, which introduced odd situation like creation of
Attribute with dummy value because the only relevant part was the attribute
kind.
Differential Revision: https://reviews.llvm.org/D116110
Global ctor evaluation currently models memory as a map from Constant*
to Constant*. For this to be correct, it is required that there is
only a single Constant* referencing a given memory location. The
Evaluator tries to ensure this by imposing certain limitations that
could result in ambiguities (by limiting types, casts and GEP formats),
but ultimately still fails, as can be seen in PR51879. The approach
is fundamentally fragile and will get more so with opaque pointers.
My original thought was to instead store memory for each global as an
offset => value representation. However, we also need to make sure
that we can actually rematerialize the modified global initializer
into a Constant in the end, which may not be possible if we allow
arbitrary writes.
What this patch does instead is to represent globals as a MutableValue,
which is either a Constant* or a MutableAggregate*. The mutable
aggregate exists to allow efficient mutation of individual aggregate
elements, as mutating an element on a Constant would require interning
a new constant. When a write to the Constant* is made, it is converted
into a MutableAggregate* as needed.
I believe this should make the evaluator more robust, compatible
with opaque pointers, and a bit simpler as well.
Fixes https://github.com/llvm/llvm-project/issues/51221.
Differential Revision: https://reviews.llvm.org/D115530
This reverts commit fd4808887e.
This patch causes gcc to issue a lot of warnings like:
warning: base class ‘class llvm::MCParsedAsmOperand’ should be
explicitly initialized in the copy constructor [-Wextra]
AAPointerInfo, and thereby other places, can look already through
internal global and stack memory. This patch enables them to look
through heap memory returned by functions with a `noalias` return.
In the future we can look through `noalias` arguments as well but that
will require AAIsDead to learn that such memory can be inspected by the
caller later on. We also need teach AAPointerInfo about dominance to
actually deal with memory that might not be `null` or `undef`
initialized. D106397 is a first step in that direction already.
Reviewed By: kuter
Differential Revision: https://reviews.llvm.org/D109170
Similar to loads, we want to be aggressive when it comes to store
simplification. Not everything in LLVM handles dead stores well when
address space casts are involved, we can simply ask the Attributor to do
it for us though.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D109998
One of the unused ident_t fields now holds the size of the string
(=const char *) field so we have an easier time dealing with those
in the future.
Differential Revision: https://reviews.llvm.org/D113126
While we skipped uses in stores if we can find all copies of the value
when the memory is loaded, we did not correlate the use in the store
with the use in the load. So far this lead to less precise results in the
offset calculations which prevented deductions. With the new
EquivalentUseCB callback argument the user of checkForAllUses can be
informed of the correlation and act on it appropriately.
Differential Revision: https://reviews.llvm.org/D109662
This patch uses the return alignment attribute now present in the
`__kmpc_alloc_shared` runtime call to set the alignment of the shared
memory global created to replace it.
Depends on D115971
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D116319
This patch changes the HeapToStack optimization to attach the return alignment
attribute information to the created alloca instruction. This would cause
problems when replacing the heap allocation with an alloca did not respect the
alignment of the original heap allocation, which would typically be aligned on
an 8 or 16 byte boundary. Malloc calls now contain alignment attributes,
so we can use that information here.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D115888
Instead of using the ArgumentPromotion implementation, we now walk
call sites using checkForAllCallSites() and directly call
areTypesABICompatible() using the replacement types. I believe
that resolves the TODO in the code.
Differential Revision: https://reviews.llvm.org/D116033
Arguments to an indirect call is by definition outside the SCC, but there's no reason we can't use locally defined facts on the call site. This also has the nice effect of further simplifying the code.
Differential Revision: https://reviews.llvm.org/D116118
The areFunctionArgsABICompatible() hook currently accepts a list of
pointer arguments, though what we're actually interested in is the
ABI compatibility after these pointer arguments have been converted
into value arguments.
This means that a) the current API is incompatible with opaque
pointers (because it requires inspection of pointee types) and
b) it can only be used in the specific context of ArgPromotion.
I would like to reuse the API when inspecting calls during inlining.
This patch converts it into an areTypesABICompatible() hook, which
accepts a list of types. This makes the method more generally usable,
and compatible with opaque pointers from an API perspective (the
actual usage in ArgPromotion/Attributor is still incompatible,
I'll follow up on that in separate patches).
Differential Revision: https://reviews.llvm.org/D116031
In regular LTO, analyze IR and discard unreachable functions when finding virtual call targets.
Differential Revision: https://reviews.llvm.org/D116056
This change allows us to infer access attributes (readnone, readonly) on arguments passed to vararg functions. Since there isn't a formal argument corresponding to the parameter, they'll never be considered part of the speculative SCC, but they can still benefit from attributes on the call site or the callee function.
The main motivation here is just to simplify the code, and remove some special casing. Previously, an indirect vararg call could return more precise results than an direct vararg call which is just weird.
Differential Revision: https://reviews.llvm.org/D115964
This fixes a bug where we would infer readnone/readonly for a function which passed a value to a function which could capture it. With the value captured in memory, the function could reload the value from memory after the call, and write to it. Inferring the argument as readnone or readonly is unsound.
@jdoerfert apparently noticed this about two years ago, and tests were checked in with 76467c4, but the issue appears to have never gotten fixed.
Since this seems like this issue should break everything, let me explain why the case is actually fairly narrow. The main inference loop over the argument SCCs only analyzes nocapture arguments. As such, we can only hit this when construction the partial SCCs. Due to that restriction, we can only hit this when we have either a) a function declaration with a manually annotated argument, or b) an immediately self recursive call.
It's also worth highlighting that we do have cases we can infer readonly/readnone on a capturing argument validly. The easiest example is a function which simply returns its argument without ever accessing it.
Differential Revision: https://reviews.llvm.org/D115961
Rename option MaxConstantsThreshold to MaxClonesThreshold. Not only is this
more descriptive, this is also in preparation of introducing another threshold
to analyse more than just 1 constant argument as we currently do, and to better
distinguish these options/thresholds.
This fixes the assertion failure reported at
https://reviews.llvm.org/D114889#3198921 with a straightforward
check, until the cleaner fix in D115924 can be reapplied.
With Control-Flow Integrity (CFI), the LowerTypeTests pass replaces
function references with CFI jump table references, which is a problem
for low-level code that needs the address of the actual function body.
For example, in the Linux kernel, the code that sets up interrupt
handlers needs to take the address of the interrupt handler function
instead of the CFI jump table, as the jump table may not even be mapped
into memory when an interrupt is triggered.
This change adds the no_cfi constant type, which wraps function
references in a value that LowerTypeTestsModule::replaceCfiUses does not
replace.
Link: https://github.com/ClangBuiltLinux/linux/issues/1353
Reviewed By: nickdesaulniers, pcc
Differential Revision: https://reviews.llvm.org/D108478
This reverts commit 9fd4f80e33.
This breaks SingleSource/Regression/C/gcc-c-torture/execute/pr19687.c
in test-suite. Either the test is incorrect, or clang is generating
incorrect union initialization code. I've submitted
https://reviews.llvm.org/D115994 to fix the test, assuming my
interpretation is correct. Reverting this in the meantime as it
may take some time to resolve.
Instead of having the speculative path be the untaken path in the branch, explicitly have it return. This does require tail duplicating one call, but the resulting code is shorter and easier to understand. Also rewrite the condition using appropriate accessors.
We were being wildly inconsistent about what memory access was implied by an indirect function call. Depending on the call site attributes, you could get anything from a read, to unknown, to none at all. (The last was a miscompile.)
We were also always traversing the uses of a readonly indirect call. This is entirely unneeded as the indirect call does not capture. The callee might capture itself internally, but that has no implications for this caller. (See the nice explanation in the CaptureTracking comments if that case is confusing.)
Note that elsewhere in the same file, we were correctly computing the nocapture attribute for indirect calls. The changed case only resulted in conservatism when computing memory attributes if say the return value was written to.
Differential Revision: https://reviews.llvm.org/D115916
There are a number of places that specially handle loads from a
uniform value where all the bits are the same (zero, one, undef,
poison), because we a) don't care about the load offset in that
case and b) it bypasses casts that might not be legal generally
but do work with uniform values.
We had multiple implementations of this, with a different set of
supported values each time, as well as incomplete type checks in
some cases. In particular, this fixes the assertion reported in
https://reviews.llvm.org/D114889#3198921, as well as a similar
assertion that could be triggered via constant folding.
Differential Revision: https://reviews.llvm.org/D115924
This is a follow up of D115458 and truncates the worklist of actual arguments
that can be specialised to 'MaxConstantsThreshold' candidates if
MaxConstantsThreshold was exceeded. Thus, this changes the behaviour of option
-func-specialization-max-constants. Before it didn't specialise at all when
this threshold was exceeded, but now it specialises up to MaxConstantsThreshold
candidates from the sorted worklist.
Differential Revision: https://reviews.llvm.org/D115509
This mostly is the same code that is refactored to decouple the cost and
benefit analysis. The biggest change is top-level function specializeFunctions
that now drives the transformation more like this:
specializeFunctions() {
Cost = getSpecializationCost(F);
calculateGains(F, Cost);
specializeFunction(F);
}
while this is just a restructuring, it helps the functional change in
calculateGains. I.e., we now sort the candidates based on the expected
specialisation gain, which we didn't do before. For this, a book keeping struct
ArgInfo was introduced. If we have a list of N candidates, but we only want
specialise less than N as set by option -func-specialization-max-constants, we
sort the list and discard the candidates that give the least benefit.
Given a formal argument, this change results in selecting the best actual
argument(s). This is NFC'ish in that this shouldn't change the current output
(hence no test change here), but in follow ups starting with D115509, it
should and I want to go one step further and compare all functions and all
arguments, which will mostly build on top of this refactoring and change.
Differential Revision: https://reviews.llvm.org/D115458
Modules that are not compiled with pseudo probe enabled can still be compiled with a sample profile input, such as in LTO postlink where other modules are probed. Since the profile is unrelated to the current modules, we should warn instead of error out the compilation.
Reviewed By: wenlei
Differential Revision: https://reviews.llvm.org/D115642
CSSPGO currently employs a flat profile format for context-sensitive profiles. Such a flat profile allows for precisely manipulating contexts that is either inlined or not inlined. This is a benefit over the nested profile format used by non-CS AutoFDO. A downside of this is the longer build time due to parsing the indexing the full CS contexts.
For a CS flat profile, though only the context profiles relevant to a module are loaded when that module is compiled, the cost to figure out what profiles are relevant is noticeably high when there're many contexts, since the sample reader will need to scan all context strings anyway. On the contrary, a nested function profile has its related inline subcontexts isolated from other unrelated contexts. Therefore when compiling a set of functions, unrelated contexts will never need to be scanned.
In this change we are exploring using nested profile format for CSSPGO. This is expected to work based on an assumption that with a preinliner-computed profile all contexts are precomputed and expected to be inlined by the compiler. Contexts not expected to be inlined will be cut off and returned to corresponding base profiles (for top-level outlined functions). This naturally forms a nested profile where all nested contexts are expected to be inlined. The compiler will less likely optimize on derived contexts that are not precomputed.
A CS-nested profile will look exactly the same with regular nested profile except that each nested profile can come with an attributes. With pseudo probes, a nested profile shown as below can also have a CFG checksum.
```
main:1968679:12
2: 24
3: 28 _Z5funcAi:18
3.1: 28 _Z5funcBi:30
3: _Z5funcAi:1467398
0: 10
1: 10 _Z8funcLeafi:11
3: 24
1: _Z8funcLeafi:1467299
0: 6
1: 6
3: 287884
4: 287864 _Z3fibi:315608
15: 23
!CFGChecksum: 138828622701
!Attributes: 2
!CFGChecksum: 281479271677951
!Attributes: 2
```
Specific work included in this change:
- A recursive profile converter to convert CS flat profile to nested profile.
- Extend function checksum and attribute metadata to be stored in nested way for text profile and extbinary profile.
- Unifiy sample loader inliner path for CS and preinlined nested profile.
- Changes in the sample loader to support probe-based nested profile.
I've seen promising results regarding build time. A nested profile can result in a 20% shorter build time than a CS flat profile while keep an on-par performance. This is with -duplicate-contexts-into-base=1.
Test Plan:
Reviewed By: wenlei
Differential Revision: https://reviews.llvm.org/D115205
Inline assembly refererences to static functions with ThinLTO+CFI were
fixed in D104058 by creating aliases for promoted functions. Creating
the aliases unconditionally resulted in an unexpected size increase in
a Chrome helper binary:
https://bugs.chromium.org/p/chromium/issues/detail?id=1261715
This is caused by the compiler being unable to drop unused code now
referenced by the alias in module-level inline assembly. This change
adds a .set_conditional assembly extension, which emits an assignment
only if the target symbol is also emitted, avoiding phantom references
to functions that could have otherwise been dropped.
This is an alternative to the solution proposed in D112761.
Reviewed By: pcc, nickdesaulniers, MaskRay
Differential Revision: https://reviews.llvm.org/D113613
Reduction functions were guarded before which was wrong, these are SPMD
compatible.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D115159
Before SPMDzation it was sufficient to add an incompatible instruction
to the SPMDCompatibilityTracker. However, now adding instructions means
they need guarding. As calls cannot be guarded in general we need to
explicitly prevent SPMD mode.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D115158
This fixes a bug in 740057d. There's two ways to describe the issue:
* One caller hasn't yet proven nocapture on the argument. Given that, the inference routine is responsible for bailing out on a potential capture.
* Even if we know the argument is nocapture, the access inference needs to traverse the exact set of users the capture tracking would (or exit conservatively). Even if capture tracking can prove a store is non-capturing (e.g. to a local alloc which doesn't escape), we still need to track the copy of the pointer to see if it's later reloaded and accessed again.
Note that all the test changes except the newly added ones appear to be false negatives. That is, cases where we could prove writeonly, but the current code isn't strong enough. That's why I didn't spot this originally.
This change extends the current logic for inferring readonly and readnone argument attributes to also infer writeonly.
This change is deliberately minimal; there's a couple of areas for follow up.
* I left out all call handling and thus any benefit from the SCC walk. When examining the test changes, I realized the existing code is imprecise, and am going to fix that in it's own revision before adding in the writeonly handling. (Mostly because updating the tests is hard when I, the human, can't figure out whether the result is correct.)
* I left out handling for storing a value (as opposed to storing to a pointer). This should benefit readonly/readnone as well, and applies to a bunch of other instructions. Seemed worth having as a separate review.
Differential Revision: https://reviews.llvm.org/D114963
The benefits of sampling-based PGO crucially depends on the quality of profile
data. This diff implements a flow-based algorithm, called profi, that helps to
overcome the inaccuracies in a profile after it is collected.
Profi is an extended and significantly re-engineered classic MCMF (min-cost
max-flow) approach suggested by Levin, Newman, and Haber [2008, Complementing
missing and inaccurate profiling using a minimum cost circulation algorithm]. It
models profile inference as an optimization problem on a control-flow graph with
the objectives and constraints capturing the desired properties of profile data.
Three important challenges that are being solved by profi:
- "fixing" errors in profiles caused by sampling;
- converting basic block counts to edge frequencies (branch probabilities);
- dealing with "dangling" blocks having no samples in the profile.
The main implementation (and required docs) are in SampleProfileInference.cpp.
The worst-time complexity is quadratic in the number of blocks in a function,
O(|V|^2). However a careful engineering and extensive evaluation shows that
the running time is (slightly) super-linear. In particular, instances with
1000 blocks are solved within 0.1 second.
The algorithm has been extensively tested internally on prod workloads,
significantly improving the quality of generated profile data and providing
speedups in the range from 0% to 5%. For "smaller" benchmarks (SPEC06/17), it
generally improves the performance (with a few outliers) but extra work in
the compiler might be needed to re-tune existing optimization passes relying on
profile counts.
UPD Dec 1st 2021:
- synced the declaration and definition of the option `SampleProfileUseProfi ` to use type `cl::opt<bool`;
- added `inline` for `SampleProfileInference<BT>::findUnlikelyJumps` and `SampleProfileInference<BT>::isExit` to avoid linking problems on windows.
Reviewed By: wenlei, hoy
Differential Revision: https://reviews.llvm.org/D109860
This bases the CleanupConstantGlobalUsers() implementation around
the ConstantFoldLoadFromConst() API. The general approach is that
we discover all users while looking through casts, and then
constant fold loads and drop stores and memintrinsics.
This avoids special cases and limitations in the previous
implementation, which is also incompatible with opaque pointers.
The result is a bit more powerful than before, because we now use
more general load folding logic which can for example look through
pointer bitcasts between different sizes. This is where the test
changes come from, as we now fold more loads and can thus remove
more globals.
Differential Revision: https://reviews.llvm.org/D114889
If two reaching kernels disagree on the execution mode we cannot guard a
function right now. Ensure we do not as we otherwise will cause a
deadlock.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D114866
For nodes that are in a cycle of a profiled call graph, the current order the underlying scc_iter computes purely depends on how those nodes are reached from outside the SCC and inside the SCC, based on the Tarjan algorithm. This does not honor profile edge hotness, thus does not gurantee hot callsites to be inlined prior to cold callsites. To mitigate that, I'm adding an extra sorter on top of scc_iter to sort scc functions in the order of callsite hotness, instead of changing the internal of scc_iter.
Sorting on callsite hotness can be optimally based on detecting cycles on a directed call graph, i.e, to remove the coldest edge until a cycle is broken. However, detecting cycles isn't cheap. I'm using an MST-based approach which is faster and appear to deliver some performance wins.
Reviewed By: wenlei
Differential Revision: https://reviews.llvm.org/D114204
This patch adds the `__kmpc_get_warp_size` OpenMP RTL function to the
externalization RAII struct. This was getting optimized out and then
being replaced with an undefined value once added back in, causing bugs
for complex reductions.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D114802
The benefits of sampling-based PGO crucially depends on the quality of profile
data. This diff implements a flow-based algorithm, called profi, that helps to
overcome the inaccuracies in a profile after it is collected.
Profi is an extended and significantly re-engineered classic MCMF (min-cost
max-flow) approach suggested by Levin, Newman, and Haber [2008, Complementing
missing and inaccurate profiling using a minimum cost circulation algorithm]. It
models profile inference as an optimization problem on a control-flow graph with
the objectives and constraints capturing the desired properties of profile data.
Three important challenges that are being solved by profi:
- "fixing" errors in profiles caused by sampling;
- converting basic block counts to edge frequencies (branch probabilities);
- dealing with "dangling" blocks having no samples in the profile.
The main implementation (and required docs) are in SampleProfileInference.cpp.
The worst-time complexity is quadratic in the number of blocks in a function,
O(|V|^2). However a careful engineering and extensive evaluation shows that
the running time is (slightly) super-linear. In particular, instances with
1000 blocks are solved within 0.1 second.
The algorithm has been extensively tested internally on prod workloads,
significantly improving the quality of generated profile data and providing
speedups in the range from 0% to 5%. For "smaller" benchmarks (SPEC06/17), it
generally improves the performance (with a few outliers) but extra work in
the compiler might be needed to re-tune existing optimization passes relying on
profile counts.
Reviewed By: wenlei, hoy
Differential Revision: https://reviews.llvm.org/D109860
The benefits of sampling-based PGO crucially depends on the quality of profile
data. This diff implements a flow-based algorithm, called profi, that helps to
overcome the inaccuracies in a profile after it is collected.
Profi is an extended and significantly re-engineered classic MCMF (min-cost
max-flow) approach suggested by Levin, Newman, and Haber [2008, Complementing
missing and inaccurate profiling using a minimum cost circulation algorithm]. It
models profile inference as an optimization problem on a control-flow graph with
the objectives and constraints capturing the desired properties of profile data.
Three important challenges that are being solved by profi:
- "fixing" errors in profiles caused by sampling;
- converting basic block counts to edge frequencies (branch probabilities);
- dealing with "dangling" blocks having no samples in the profile.
The main implementation (and required docs) are in SampleProfileInference.cpp.
The worst-time complexity is quadratic in the number of blocks in a function,
O(|V|^2). However a careful engineering and extensive evaluation shows that
the running time is (slightly) super-linear. In particular, instances with
1000 blocks are solved within 0.1 second.
The algorithm has been extensively tested internally on prod workloads,
significantly improving the quality of generated profile data and providing
speedups in the range from 0% to 5%. For "smaller" benchmarks (SPEC06/17), it
generally improves the performance (with a few outliers) but extra work in
the compiler might be needed to re-tune existing optimization passes relying on
profile counts.
Reviewed By: wenlei, hoy
Differential Revision: https://reviews.llvm.org/D109860
In a CGSCC pass manager, we may visit the same function multiple times
due to SCC mutations. In the inliner pipeline, this results in running
the function simplification pipeline on a function multiple times even
if it hasn't been changed since the last function simplification
pipeline run.
We use a newly introduced analysis to keep track of whether or not a
function has changed since the last time the function simplification
pipeline has run on it. If we see this analysis available for a function
in a CGSCCToFunctionPassAdaptor, we skip running the function passes on
the function. The analysis is queried at the end of the function passes
so that it's available after the first time the function simplification
pipeline runs on a function. This is a per-adaptor option so it doesn't
apply to every adaptor.
The goal of this is to improve compile times. However, currently we
can't turn this on by default at least for the higher optimization
levels since the function simplification pipeline is not robust enough
to be idempotent in many cases, resulting in performance regressions if
we stop running the function simplification pipeline on a function
multiple times. We may be able to turn this on for -O1 in the near
future, but turning this on for higher optimization levels would require
more investment in the function simplification pipeline.
Heavily inspired by D98103.
Example compile time improvements with flag turned on:
https://llvm-compile-time-tracker.com/compare.php?from=998dc4a5d3491d2ae8cbe742d2e13bc1b0cacc5f&to=5c27c913687d3d5559ef3ab42b5a3d513531d61c&stat=instructions
Reviewed By: asbirlea, nikic
Differential Revision: https://reviews.llvm.org/D113947
std::hash returns a 64bit hash code while previously we were using only lower 32 bits which caused hash collision for large workloads.
Reviewed By: wenlei, wlei
Differential Revision: https://reviews.llvm.org/D113688
Previously, any change in any function in an SCC would cause all
analyses for all functions in the SCC to be invalidated. With this
change, we now manually invalidate analyses for functions we modify,
then let the pass manager know that all function analyses should be
preserved since we've already handled function analysis invalidation.
So far this only touches the inliner, argpromotion, function-attrs, and
updateCGAndAnalysisManager(), since they are the most used.
This is part of an effort to investigate running the function
simplification pipeline less on functions we visit multiple times in the
inliner pipeline.
However, this causes major memory regressions especially on larger IR.
To counteract this, turn on the option to eagerly invalidate function
analyses. This invalidates analyses on functions immediately after
they're processed in a module or scc to function adaptor for specific
parts of the pipeline.
Within an SCC, if a pass only modifies one function, other functions in
the SCC do not have their analyses invalidated, so in later function
passes in the SCC pass manager the analyses may still be cached. It is
only after the function passes that the eager invalidation takes effect.
For the default pipelines this makes sense because the inliner pipeline
runs the function simplification pipeline after all other SCC passes
(except CoroSplit which doesn't request any analyses).
Overall this has mostly positive effects on compile time and positive effects on memory usage.
https://llvm-compile-time-tracker.com/compare.php?from=7f627596977624730f9298a1b69883af1555765e&to=39e824e0d3ca8a517502f13032dfa67304841c90&stat=instructionshttps://llvm-compile-time-tracker.com/compare.php?from=7f627596977624730f9298a1b69883af1555765e&to=39e824e0d3ca8a517502f13032dfa67304841c90&stat=max-rss
D113196 shows that we slightly regressed compile times in exchange for
some memory improvements when turning on eager invalidation. D100917
shows that we slightly improved compile times in exchange for major
memory regressions in some cases when invalidating less in SCC passes.
Turning these on at the same time keeps the memory improvements while
keeping compile times neutral/slightly positive.
Reviewed By: asbirlea, nikic
Differential Revision: https://reviews.llvm.org/D113304
ProfileCount could model invalid values, but a user had no indication
that the getCount method could return bogus data. Optional<ProfileCount>
addresses that, because the user must dereference the optional. In
addition, the patch removes concept duplication.
Differential Revision: https://reviews.llvm.org/D113839
This patch uses the abstract attributor introduced in D111054 to get the
assumption values instead of the `hasAssumption` function. This also
calls it so assumption information should propagate throug the device
where applicabile.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D111445
This patch introduces a new abstract attributor instance that propagates
assumption information from functions. Conceptually, if a function is
only called by functions that have certain assumptions, then we can
apply the same assumptions to that function. This problem is similar to
calculating the dominator set, but the assumptions are merged instead of
nodes.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D111054
This is in preparation for only invalidating analyses on changed
functions.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D113303
Function specialisation was running at all optimisation levels (if enabled on
the command line, it is not on by default). That was an oversight and not
something we want to do. Function specialisation duplicates functions when it
triggers, so the backend is processing more functions/instructions resulting in
compile-time increases, which seems more appropriate with -O3 and inline with
GCC. Please note that since function specialisation is not enabled by default,
this didn't require updating any pass manager tests.
Differential Revision: https://reviews.llvm.org/D112129
Coroutines have weird semantics that don't quite match normal LLVM functions,
so trying to infer even simple attributes based on thier contents can go wrong.
We already make sure to properly clear analyses for deleted functions.
This makes investigating some future potential compile time improvements easier.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D113032
If we assume SPMD-mode during the fixpoint iteration we have to execute
the kernel in SPMD-mode. If we change our mind during manifest there is
the chance of a mismatch between the simplification, e.g., of
`__kmpc_is_spmd_exec_mode` calls, and the execution mode. This problem
was introduced in D109438.
This patch is compromise to resolve the problem purely in OpenMP-opt
while trying to keep the benefits of D109438 around. This might not
always work, see `get_hardware_num_threads_in_block_fold` but it often
does. At the same time we do keep value specialization and execution
mode in sync.
Proper solutions to this problem should be considered. I believe a new
execution mode is the easiest way forward (Singleton-SPMD).
Alternatively, SPMD-mode execution can be used with a way to provide a
new thread_limit (here 1) to the runtime. This is more general and could
be useful if we see `num_threads` clauses or workshared loops with small
trip counts in the kernel. In either proposal we need to disable the
guarding for the kernel (which was the motivation for D109438).
Reviewed By: jhuber6
Differential Revision: https://reviews.llvm.org/D112894
Before we had aligned barriers the `__kmpc_barrier_simple_spmd` was
OK to be used in the custom state machine. Now that SPMD barriers are
assumed to be aligned we need to use a "generic" barrier in places
that are not aligned.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D112893
Global symbols cannot have any name so we need to sanitize the string
first. Also remove an assertion that is not actually necessary nor
true in general.
Reviewed By: ggeorgakoudis
Differential Revision: https://reviews.llvm.org/D112892
As noted in https://reviews.llvm.org/D90924#inline-1076197
apparently this is a pretty common pattern,
let's not repeat it yet again, but have it in a common place.
There may be some more places where it could be used,
but these are the most obvious ones.
Adds the following switches:
1. --sample-profile-inline-replay-fallback/--cgscc-inline-replay-fallback: controls what the replay advisor does for inline sites that are not present in the replay. Options are:
1. Original: defers to original advisor
2. AlwaysInline: inline all sites not in replay
3. NeverInline: inline no sites not in replay
2. --sample-profile-inline-replay-format/--cgscc-inline-replay-format: controls what format should be generated to match against the replay remarks. Options are:
1. Line
2. LineColumn
3. LineDiscriminator
4. LineColumnDiscriminator
Adds support for negative inlining decisions. These are denoted by "will not be inlined into" as compared to the positive "inlined into" in the remarks.
All of these together with the previous `--sample-profile-inline-replay-scope/--cgscc-inline-replay-scope` allow tweaking in how to apply replay. In my testing, I'm using:
1. --sample-profile-inline-replay-scope/--cgscc-inline-replay-scope = Function to only replay on a function
2. --sample-profile-inline-replay-fallback/--cgscc-inline-replay-fallback = NeverInline since I'm feeding in only positive remarks to the replay system
3. --sample-profile-inline-replay-format/--cgscc-inline-replay-format = Line since I'm generating the remarks from DWARF information from GCC which can conflict quite heavily in column number compared to Clang
An alternative configuration could be to do Function, AlwaysInline, Line fallback with negative remarks which closer matches the final call-sites. Note that this can lead to unbounded inlining if a negative remark doesn't match/exist for one reason or another.
Updated various tests to cover the new switches and negative remarks
Testing:
ninja check-all
Reviewed By: wenlei, mtrofin
Differential Revision: https://reviews.llvm.org/D112040
Replay in sample profiling needs to be asked on candidates that may not have counts or below the threshold. If replay is in effect for a function make sure these are captured and also imported during thinLTO.
Testing:
ninja check-all
Reviewed By: wenlei
Differential Revision: https://reviews.llvm.org/D112033
createReplacementInstr was a trivial wrapper around
ConstantExpr::getAsInstruction, which also inserted the new instruction
into a basic block. Implement this directly in getAsInstruction by
adding an InsertBefore parameter and change all callers to use it. NFC.
A follow-up patch will remove createReplacementInstr.
Differential Revision: https://reviews.llvm.org/D112791
Move the section collecting `AlwaysPreserved` up before any
`maybeInternalize` is called. Otherwise, functions in `AlwaysPreserved` (in this case, `__stack_chk_fail`)
are not preserved.
Reviewed By: MaskRay, tejohnson
Differential Revision: https://reviews.llvm.org/D112684
Even if we look for `nocapture` we need to bail on escaping pointers.
The crucial thing is that we might not look at a big enough scope when
we derive the memory behavior. Thus, it might be `nocapture` in a larger
context while it is "captured" in a smaller context.
When we strip and accumulate constant offsets we need to pick the right
address space such that the offset APInt has the right bit width.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D112544
A constant complaint we get is that the __typeid__ symbols in the CFI
jump tables causes confusing stack traces in applications. Emit the more
readable cfi_jt aliases regardless of function export (LTO vs Thin LTO).
Reviewed By: pcc, tejohnson
Differential Revision: https://reviews.llvm.org/D107934
Currently strip.invariant/launder.invariant are handled by
constructing constant expressions with the intrinsics skipped.
This takes an alternative approach of accumulating the offset
using stripAndAccumulateConstantOffsets(), with a flag to look
through invariant.group intrinsics.
Differential Revision: https://reviews.llvm.org/D112382
Make use of the getGEPIndicesForOffset() helper for creating GEPs.
This handles arrays as well, uses correct GEP index types and
reduces code duplication.
Differential Revision: https://reviews.llvm.org/D112263
As discussed in:
* https://reviews.llvm.org/D94166
* https://lists.llvm.org/pipermail/llvm-dev/2020-September/145031.html
The GlobalIndirectSymbol class lost most of its meaning in
https://reviews.llvm.org/D109792, which disambiguated getBaseObject
(now getAliaseeObject) between GlobalIFunc and everything else.
In addition, as long as GlobalIFunc is not a GlobalObject and
getAliaseeObject returns GlobalObjects, a GlobalAlias whose aliasee
is a GlobalIFunc cannot currently be modeled properly. Creating
aliases for GlobalIFuncs does happen in the wild (e.g. glibc). In addition,
calling getAliaseeObject on a GlobalIFunc will currently return nullptr,
which is undesirable because it should return the object itself for
non-aliases.
This patch refactors the GlobalIFunc class to inherit directly from
GlobalObject, and removes GlobalIndirectSymbol (while inlining the
relevant parts into GlobalAlias and GlobalIFunc). This allows for
calling getAliaseeObject() on a GlobalIFunc to return the GlobalIFunc
itself, making getAliaseeObject() more consistent and enabling
alias-to-ifunc to be properly modeled in the IR.
I exercised some judgement in the API clients of GlobalIndirectSymbol:
some were 'monomorphized' for GlobalAlias and GlobalIFunc, and
some remained shared (with the type adapted to become GlobalValue).
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D108872
The goal is to allow grafting an inline tree from Clang or GCC into a new compilation without affecting other functions. For GCC, we're doing this by extracting the inline tree from dwarf information and generating the equivalent remarks.
This allows easier side-by-side asm analysis and a trial way to see if a particular inlining setup provides benefits by itself.
Testing:
ninja check-all
Reviewed By: wenlei, mtrofin
Differential Revision: https://reviews.llvm.org/D110658
Even if there are no interesting functions, the SCCP solver would still run
before bailing. Now bail earlier, avoid running the solver for nothing.
Differential Revision: https://reviews.llvm.org/D111645
If another inlining session came after a ModuleInlinerWrapperPass, the
advisor alanysis would still be cached, but its Result would be cleared.
We need to clear both.
This addresses PR52118
Differential Revision: https://reviews.llvm.org/D111586
This patch continues unblocking optimizations that are blocked by pseudo probe instrumentation.
Not exactly like DbgIntrinsics, PseudoProbe intrinsic has other attributes (such as mayread, maywrite, mayhaveSideEffect) that can block optimizations. The issues fixed are:
- Flipped default param of getFirstNonPHIOrDbg API to skip pseudo probes
- Unblocked CSE by avoiding pseudo probe from clobbering memory SSA
- Unblocked induction variable simpliciation
- Allow empty loop deletion by treating probe intrinsic isDroppable
- Some refactoring.
Reviewed By: wenlei
Differential Revision: https://reviews.llvm.org/D110847
This is a follow up of D110529 that disallowed constexprs. That change
introduced a regression as this also disallowed constexprs that are function
pointers, which is actually one of the motivating use cases that we do want to
support.
Differential Revision: https://reviews.llvm.org/D111567
To better reflect the meaning of the now-disambiguated {GlobalValue,
GlobalAlias}::getBaseObject after breaking off GlobalIFunc::getResolverFunction
(D109792), the function is renamed to getAliaseeObject.
As described on D111049, we're trying to remove the <string> dependency from error handling and replace uses of report_fatal_error(const std::string&) with the Twine() variant which can be forward declared.
This also removes the need to disable the mandatory inlining phase in
tests.
In a departure from the previous remark, we don't output a 'cost' in
this case, because there's no such thing. We just report that inlining
happened because of the attribute.
Differential Revision: https://reviews.llvm.org/D110891
This patch adds a new command line option `openmp-opt-max-iterations`
that controls the maximum number of iterations the attributor will run
for when compiling OpenMP target device code. This patch also adds a
remark to indicate when the attributor failed because it did not run
for enough iterations.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D110749
Stop using APInt constructors and methods that were soft-deprecated in
D109483. This fixes all the uses I found in llvm, except for the APInt
unit tests which should still test the deprecated methods.
Differential Revision: https://reviews.llvm.org/D110807
Summary:
The RTL functions added in https://reviews.llvm.org/D110429 were
mistakenly left out from the list of safe runtime calls in AAKernelInfo.
This patch adds them in.
Thinlink provides an opportunity to propagate function attributes across modules, enabling additional propagation opportunities.
This change propagates (currently default off, turn on with `disable-thinlto-funcattrs=1`) noRecurse and noUnwind based off of function summaries of the prevailing functions in bottom-up call-graph order. Testing on clang self-build:
1. There's a 35-40% increase in noUnwind functions due to the additional propagation opportunities.
2. Throughput is measured at 10-15% increase in thinlink time which itself is 1.5% of E2E link time.
Implementation-wise this adds the following summary function attributes:
1. noUnwind: function is noUnwind
2. mayThrow: function contains a non-call instruction that `Instruction::mayThrow` returns true on (e.g. windows SEH instructions)
3. hasUnknownCall: function contains calls that don't make it into the summary call-graph thus should not be propagated from (e.g. indirect for now, could add no-opt functions as well)
Testing:
Clang self-build passes and 2nd stage build passes check-all
ninja check-all with newly added tests passing
Reviewed By: tejohnson
Differential Revision: https://reviews.llvm.org/D36850
Function specialization was crashing on poison values and constexpr values.
The problem is that these values are not added to the solver, so it crashes
when a lookup is performed for these values. This fixes that by not
specialising on these values. For poison that is obvious, but for constexpr
this is a change in behaviour. Thus, in one way this is a bit of a stopgap, but
specialising on constexpr values wasn't done very intentionally, and need some
more work and tests if we wanted to support this.
As a follow up, we need to look if the solver should exit more gracefully and
return a "don't know", or that it should really support these constexprs.
This should fix PR51600 (https://bugs.llvm.org/show_bug.cgi?id=51600).
Differential Revision: https://reviews.llvm.org/D110529
In ThinLTO for locals we normally compute the GUID from the name after
prepending the source path to get a unique global id. SamplePGO indirect
call profiles contain the target GUID without this uniquification,
however (unless compiling with -funique-internal-linkage-names).
In order to correctly handle the call edges added to the combined index
for these indirect calls, during importing and bitcode writing we
consult a map of original to full GUID to identify the actual callee.
However, for a large application this was consuming a lot of compile
time as we need to do this repeatedly (especially during importing where
we may traverse call edges multiple times).
To fix this implement a suggestion in one of the FIXME comments, and
actually modify the call edges during a single traversal after the index
is built to perform the fixups once. I combined this fixup with the dead
code analysis performed on the index in order to avoid adding an
additional walk of the index. The dead code analysis is the first
analysis performed on the index.
This reduced the time required for a large thin link with SamplePGO by
about 20%.
No new test added, but I confirmed that there are existing tests that
will fail when no fixup is performed.
Differential Revision: https://reviews.llvm.org/D110374
Bisecting and reducing opt pipelines that includes the
ModuleInlinerWrapperPass has turned out to be a bit problematic.
This is far from perfect (it still lacks information about inline
advisor params etc.), but it should give some kind of hint to what
the wrapped pipeline looks like when using -print-pipeline-passes.
Reviewed By: aeubanks, mtrofin
Differential Revision: https://reviews.llvm.org/D109878
This patch fixes a problem when the AAKernelInfo state was invalidated,
e.g., due to `optnone` for a kernel, but not all parts indicated the
invalidation properly. We further eliminate most full state
invalidations as they should never be necessary.
Differential Revision: https://reviews.llvm.org/D109468
This is a follow-up of D110029, which uses bitset to indicate execution mode. This patches makes the changes in the function call.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D110279
The execution mode of a kernel is stored in a global variable, whose value means:
- 0 - SPMD mode
- 1 - indicates generic mode
- 2 - SPMD mode execution with generic mode semantics
We are going to add support for SIMD execution mode. It will be come with another
execution mode, such as SIMD-generic mode. As a result, this value-based indicator
is not flexible.
This patch changes to bitset based solution to encode execution mode. Each
position is:
[0] - generic mode
[1] - SPMD mode
[2] - SIMD mode (will be added later)
In this way, `0x1` is generic mode, `0x2` is SPMD mode, and `0x3` is SPMD mode
execution with generic mode semantics. In the future after we add the support for
SIMD mode, `0b1xx` will be in SIMD mode.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D110029
Summary:
The thread ID function was reintroduced in D110195, but could
potentially be removed by the optimizer. Make the function noinline to
preserve the call sites and add it to the externalization RAII so its
definition is not removed by the attributor.
IR with matrix intrinsics is likely to also contain large vector
operations, which can benefit from early simplifications.
This is the last step in a series of changes to improve code-gen for
code using matrix subscript operators with the C/C++ matrix extension in
CLang, like
using matrix_t = double __attribute__((matrix_type(15, 15)));
void foo(unsigned i, matrix_t &A, matrix_t &B) {
for (unsigned j = 0; j < 4; ++j)
for (unsigned k = 0; k < i; k++)
B[k][j] -= A[k][j] * B[i][j];
}
https://clang.godbolt.org/z/6dKxK1Ed7
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D102496
Add a new LLVM switch `-profile-sample-block-accurate` to trust zero block counts for branches. Currently we leave out such zero counts when annotating branch weight metadata, which would lead to weights being considered as unknown.
Differential Revision: https://reviews.llvm.org/D110117
In default pipelines the ModuleInlinerWrapperPass is adding the
InlinerPass to the pipeline twice, once due to MandatoryFirst (passing
true in the ctor) and then a second time with false as argument.
To make it possible to bisect and reduce opt test cases for this
part of the pipeline we need to be able to choose between the two
different variants of the InlinerPass when running opt. This patch is
changing 'inline' to a CGSCC_PASS_WITH_PARAMS in the PassRegistry,
making it possible run opt with both -passes=cgscc(inline) and
-passes=cgscc(inline<only-mandatory>).
Reviewed By: aeubanks, mtrofin
Differential Revision: https://reviews.llvm.org/D109877
The AAExecutionDomain instance checks if a BB is executed by the main
thread only. Currently, this only checks the `__kmpc_kernel_init` call
for generic regions to indicate the path taken by the main thread. In
the new runtime, we want to be able to detect basic blocks even in SPMD
mode. For this we enable it to check thread-ID intrinsics being compared
to zero as well.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D109849
This introduces an option to allow specialising on the address of global
values. This option is off by default because it is likely not that profitable
to do so and needs more investigation. Before, we were specialising on addresses
and thus this changes the default behaviour.
Differential Revision: https://reviews.llvm.org/D109775
Do not call `TryToShrinkGlobalToBoolean` for address spaces
that don't allow initializers. It inserts an initializer value
while shrinking to bool. Used the target hook introduced with
D109337 to skip this call for the restricted address spaces.
Reviewed By: tra
Differential Revision: https://reviews.llvm.org/D109823
getMetadata() currently uses a weird API where it populates a
structure passed to it, and optionally merges into it. Instead,
we can return the AAMDNodes and provide a separate merge() API.
This makes usages more compact.
Differential Revision: https://reviews.llvm.org/D109852
Added '-print-pipeline-passes' printing of parameters for those passes
declared with *_WITH_PARAMS macro in PassRegistry.def.
Note that it only prints the parameters declared inside *_WITH_PARAMS as
in a few cases there appear to be additional parameters not parsable.
The following passes are now covered (i.e. all of those with *_WITH_PARAMS in
PassRegistry.def).
LoopExtractorPass - loop-extract
HWAddressSanitizerPass - hwsan
EarlyCSEPass - early-cse
EntryExitInstrumenterPass - ee-instrument
LowerMatrixIntrinsicsPass - lower-matrix-intrinsics
LoopUnrollPass - loop-unroll
AddressSanitizerPass - asan
MemorySanitizerPass - msan
SimplifyCFGPass - simplifycfg
LoopVectorizePass - loop-vectorize
MergedLoadStoreMotionPass - mldst-motion
GVN - gvn
StackLifetimePrinterPass - print<stack-lifetime>
SimpleLoopUnswitchPass - simple-loop-unswitch
Differential Revision: https://reviews.llvm.org/D109310
Patch by @dpalermo
The corrupt bitcode reported in https://bugs.llvm.org/show_bug.cgi?id=51647 seems to be a result of a later pass changing the workfn variable to addrspace(5) (thread private, on the stack). That seems reasonable for an alloca without an address space so it's an open question why that can crash the bitcode reader.
This change puts it in the thread private address space to begin with which means whatever misfired further down the pipeline does not break it. That matches the codegen from clang where stack variables are always annotated (5) and then addrspace cast prior to following use.
This therefore patches around whatever unsuccessfully moved the alloca variable to addrspace(5). That solves the problem of openmp opt producing code that crashes the bitcode reader. It should be possible to create a minimal repro for the underlying bug based on some handwritten IR that uses an alloca in a generic address space.
Reviewed By: ronlieb, jdoerfert, dpalermo-phab
Differential Revision: https://reviews.llvm.org/D109500
This patch fixes a error made in 2cc6f7c8e1. That patch
added a call site position but there was a small error with the way
the presence of a unknown call edge was being propagated from call site
to function. This patch fixes that error. This error was effecting some
AMDGPU tests.
This patch makes it possible to query callbase reachability
(Can a callbase reach a function Fn transitively).
The patch moves the reachability query handling logic to a member class,
this class will have more users within the AA once we add other function
reachability queries.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106402
This patch adds a call site position for AACallEdges, this
allows us to ask questions about which functions a specific
`CallBase` might call.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106208
We peform runtime folding, but do not currently emit remarks when it is
performed. This is because it comes from the runtime library and is
beyond the users control. However, people may still wish to view this
and similar information easily, so we can enable this behaviour using a
special flag to enable verbose remarks.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D109627
This reapplies commit 7dbba3376f, or, put
differently, this reverts commit d9a8d20827.
The test now requires the amdgpu and nvptx backend explicitly as it
won't work without properly.
This patch adds functionality to check assumption attributes on call
sites as well.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D109376
Not all address spaces support initializers for globals and we can
therefore not set them without checking if they are allowed. This
patch adds a hook into TTI to check if an AS allows non-undef
initializers. We disable it for all but address space 0 by default,
NVPTX and AMDGPU targets allow all but address space 3.
Reviewed By: tra
Differential Revision: https://reviews.llvm.org/D109337
When we guard side-effects as part of SPMDzation we do it for
consecutive instructions that need guarding. This patch will try to
reorder guarded side-effects in a block to decrease the number of
guarded regions we need. It does not use any smarts, e.g., alias
analysis, to move side-effects over non-interfering reads. Instead,
it only moves side-effects downwards to the next guarded side-effect
if there was nothing in between that could have possibly be affected.
Reviewed By: ggeorgakoudis
Differential Revision: https://reviews.llvm.org/D109070
The MinSize attribute can be attached to both the callee and the caller
in the callsite. Function specialisation was already skipped for function
declarations (callees) with MinSize. This also skips specialisations for
the callsite when it has MinSize set.
Differential Revision: https://reviews.llvm.org/D109441
D104143 introduced canonical value numbering between regions, which allows for the easy identification of items across a region, eliminating the need in the outliner to create parallel lists of instructions for each region, and replace output values in a less convoluted way.
Additionally, in a future commit, the output values will not necessarily be recorded values from the region itself, it could be a combination value where the actual value being output is a PHINode instead. This new method allows us to handle the replacement of the output value to the stored value with the corresponding item in the same place for both normal output values, and PHINode outputs instead of handling the different types of outputs in different locations.
Reviewers: paquette, roelofs
Differential Revision: https://reviews.llvm.org/D108656
This patch changes SPMDization to not trigger for regions with no
parallelism. Otherwise, this will introduce unnecessary barriers that
will slow the single-threaded region down.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D109438
When we start outlining across branches, there is the possibility that we will have two different blocks with different output locations, or a single branch that goes to two blocks outside of the region that is being outlined. While the CodeExtractor provides most of the mechanisms by using the return value of the extracted function as the input to a switch statement to correctly branch to the correct location, we need special handling for different output schemas to each location.
This is done by repeating the existing storing scheme for each different exit block. We have a map from the return values used, to the basic block that is used to store the outputs for that particular exit block within the outlined function. Then if needed, we create a switch statement for each return block to branch to the correct set of stored outputs.
Reviewers: paquette
Differential Revision: https://reviews.llvm.org/D106993
Using the similarity found from the IRSimilarity Identifier, we take regions with structural similarity, and deduplicate them into a separate function. The Code Extractor is able to provide most of this functionality.
For simplicity, we start by only outlining regions with a single entry and single exit branch, this reduces the complexity in handling phi nodes outside the region, and handling many sets of outputs for each of the different exit blocks.
Reviewer: paquette
Differential Revision: https://reviews.llvm.org/D106990
The current IRSimilarityIdentifier does not try to find similarity across blocks, this patch provides a mechanism to compare two branches against one another, to find similarity across basic blocks, rather than just within them.
This adds a step in the similarity identification process that labels all of the basic blocks so that we can identify the relative branching locations. Within an IRSimilarityCandidate we use these relative locations to determine whether if the branching to other relative locations in the same region is the same between branches. If they are, we consider them similar.
We do not consider the relative location of the branch if the target branch is outside of the region. In this case, both branches must exit to a location outside the region, but the exact relative location does not matter.
Reviewers: paquette, yroux
Differential Revision: https://reviews.llvm.org/D106989
Recommit of 707ce34b06. Don't introduce a
dependency to the LLVMPasses component, instead register the required
passes individually.
Add methods for loop unrolling to the OpenMPIRBuilder class and use them in Clang if `-fopenmp-enable-irbuilder` is enabled. The unrolling methods are:
* `unrollLoopFull`
* `unrollLoopPartial`
* `unrollLoopHeuristic`
`unrollLoopPartial` and `unrollLoopHeuristic` can use compiler heuristics to automatically determine the unroll factor. If possible, that is if no CanonicalLoopInfo is required to pass to another method, metadata for LLVM's LoopUnrollPass is added. Otherwise the unroll factor is determined using the same heurstics as user by LoopUnrollPass. Not requiring a CanonicalLoopInfo, especially with `unrollLoopHeuristic` allows greater flexibility.
With full unrolling and partial unrolling with known unroll factor, instead of duplicating instructions by the OpenMPIRBuilder, the full unroll is still delegated to the LoopUnrollPass. In case of partial unrolling the loop is first tiled using the existing `tileLoops` methods, then the inner loop fully unrolled using the same mechanism.
Reviewed By: jdoerfert, kiranchandramohan
Differential Revision: https://reviews.llvm.org/D107764
When pre-inliner decision is used for CSSPGO, we should take that into account for ThinLTO importing as well, so post-link sample loader inliner can favor that decision. This is handled by a small tweak in this patch. It also includes a change to transfer preinliner decision when merging context.
Differential Revision: https://reviews.llvm.org/D109088
When preinliner is used for CSSPGO, we try to honor global preinliner decision as much as we can except for uninlinable callees. We rely on InlineCost::Never to prevent us from illegal inlining.
However, it turns out that we use InlineCost::Never for both illeagle inlining and some of the "not-so-beneficial" inlining.
The most common one is recursive inlining, while it can bloat size a lot during CGSCC bottom-up inlining, it's less of a problem when recursive inlining is guided by profile and done in top-down manner.
Ideally it'd be better to have a clear separation between inline legality check vs cost-benefit check, but that requires a bigger change.
This change enables InlineCost computation to allow inlining recursive calls, controlled by InlineParams. In SampleLoader, we now enable recursive inlining for CSSPGO when global preinliner decision is used.
With this change, we saw a few perf improvements on SPEC2017 with CSSPGO and preinliner on: 2% for povray_r, 6% for xalancbmk_s, 3% omnetpp_s, while size is about the same (no noticeable perf change for all other benchmarks)
Differential Revision: https://reviews.llvm.org/D109104
When pre-inliner decision is used for CSSPGO, we should take that into account for ThinLTO importing as well, so post-link sample loader inliner can favor that decision. This is handled by a small tweak in this patch. It also includes a change to transfer preinliner decision when merging context.
Differential Revision: https://reviews.llvm.org/D109088
For CSSPGO, turn on `sample-profile-use-preinliner` by default. This simplifies the use of llvm-profgen preinliner as it's now simply driven by ContextShouldBeInlined flag for each context profile without needing extra compiler switch.
Note that llvm-profgen's preinliner is still off by default, under switch `csspgo-preinliner`.
Differential Revision: https://reviews.llvm.org/D109111
Adding the compiler support of MD5 CS profile based on pervious context split work D107299. A MD5 CS profile is about 40% smaller than the string-based extbinary profile. As a result, the compilation is 15% faster.
There are a few conversion from real names to md5 names that have been made on the sample loader and context tracker side to get it work.
Reviewed By: wenlei, wmi
Differential Revision: https://reviews.llvm.org/D108342
Performance on GPU targets can be highly variable, sometimes inlining
everything hurts performance and sometimes it greatly improves it. Add
an option to toggle this behaviour to better investigate it.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D109014
To support Virtual Function Elimination to Swift, this PR adds support for Swift
vtables which contain "relative pointers" instead of direct pointer references.
These are in the form of:
@symbol = ... {
i32 trunc (i64 sub (i64 ptrtoint (<type> @target to i64), i64 ptrtoint (... @symbol to i64)) to i32)
}
The PR extends GlobalDCE's way of looking up a vtable offset into a dependency
to be able to see through this expression and find the target symbol.
Differential Revision: https://reviews.llvm.org/D107645
Currently context strings contain a lot of duplicated function names and that significantly increase the profile size. This change split the context into a series of {name, offset, discriminator} tuples so function names used in the context can be replaced by the index into the name table and that significantly reduce the size consumed by context.
A follow-up improvement made in the compiler and profiling tools is to avoid reconstructing full context strings which is time- and memory- consuming. Instead a context vector of `StringRef` is adopted to represent the full context in all scenarios. As a result, the previous prevalent profile map which was implemented as a `StringRef` is now engineered as an unordered map keyed by `SampleContext`. `SampleContext` is reshaped to using an `ArrayRef` to represent a full context for CS profile. For non-CS profile, it falls back to use `StringRef` to represent a contextless function name. Both the `ArrayRef` and `StringRef` objects are underpinned by real array and string objects that are stored in producer buffers. For compiler, they are maintained by the sample reader. For llvm-profgen, they are maintained in `ProfiledBinary` and `ProfileGenerator`. Full context strings can be generated only in those cases of debugging and printing.
When it comes to profile format, nothing has changed to the text format, though internally CS context is implemented as a vector. Extbinary format is only changed for CS profile, with an additional `SecCSNameTable` section which stores all full contexts logically in the form of `vector<int>`, which each element as an offset points to `SecNameTable`. All occurrences of contexts elsewhere are redirected to using the offset of `SecCSNameTable`.
Testing
This is no-diff change in terms of code quality and profile content (for text profile).
For our internal large service (aka ads), the profile generation is cut to half, with a 20x smaller string-based extbinary format generated.
The compile time of ads is dropped by 25%.
Differential Revision: https://reviews.llvm.org/D107299
Currently, the IROutliner uses a simple metric to outline the largest amount
of IR possible to outline first if it fits the cost model. This is model
loses out on smaller blocks of code that have higher reductions in cost that
are contained within larger blocks of IR.
This reverses the order, where we calculate all of the costs first, and then
reorder and extract items based on the calculated results.
Reviewers: paquette
Differential Revision: https://reviews.llvm.org/D106440
Occasionally instructions are between the last instruction in a region,
and the following instruction as identified by the Candidate. This
adds an extra check right before splitting a candidate that excludes the region from being split/checked for outlining to remove errors.
Tests Added:
Tranforms/IROuutliner/outlining-extra-bitcasts.ll
Reviewer: paquette, jroelofs
Differential Revision: https://reviews.llvm.org/D104142
When the initial relationship between two pairs of values between
similar sections is ambiguous to commutativity, arguments to the
outlined functions can be passed in such that the order is incorrect,
causing miscompilations. This adds a canonical mapping to each
similarity section, so that we can maintain the relationship of global
value numbering from one section to another.
Added Tests:
Transforms/IROutliner/outlining-commutative-operands-opposite-order.ll
unittests/Analysis/IRSimilarityIdentifierTest.cpp - IRSimilarityCandidate:CanonicalNumbering
Reviewers: jroelofs, jpaquette, yroux
Differential Revision: https://reviews.llvm.org/D104143
Recursion can happen when we see a PHI use the second time or when we
look at a store value operand use again. We already visited the
potential copies and doing so again will just cause endless looping.
Reviewed By: kuter
Differential Revision: https://reviews.llvm.org/D108190
For now we do should not treat byval arguments as local copies performed
on the call edge, though, in general we should. To make that happen we
need to teach various passes, e.g., DSE, about the copy effect of a
byval. That would also allow us to mark functions only accessing byval
arguments as readnone again, atguably their acceses have no effect
outside of the function, like accesses to allocas.
Reviewed By: kuter
Differential Revision: https://reviews.llvm.org/D108140
We try to forward a stored-once-constant-value from one global access
to another, but that's not safe if the constant value is an expression
that can trap.
The tests are reduced from the miscompile examples in:
https://llvm.org/PR47578
Differential Revision: https://reviews.llvm.org/D108771
The change adds a switch to allow sample loader to use global pre-inliner's decision instead. The pre-inliner in llvm-profgen makes inline decision globally based on whole program profile and function byte size as cost proxy.
Since pre-inliner also adjusts/merges context profile based on its inline decision, honoring its inline decision in sample loader would lead to better post-inline profile quality especially for thinlto where cross module profile merging isn't possible without pre-inliner.
Minor fix in profile reader is also included. When pre-inliner is use, we now also turn off the default merging and trimming logic unless it's explicitly asked.
Differential Revision: https://reviews.llvm.org/D108677
This is a follow up diff for BinarySizeContextTracker to track zero size for fully optimized inlinee. When an inlinee is fully optimized away, we won't be able to get its size through symbolizing instructions, hence we will treat the corresponding context size as unknown. However by traversing the inlined probe forest, we know what're original inlinees regardless of optimization. If a context show up in inlined probes, but not during symbolization, we know that it's fully optimized away hence its size is zero instead of unknown. It should provide more accurate size cost estimation for pre-inliner to make better inline decisions in llvm-profgen.
Differential Revision: https://reviews.llvm.org/D108350
It would waste time to specialize a function which would inline finally.
This patch did two things:
- Don't specialize functions which are always-inline.
- Don't spescialize functions whose lines of code are less than threshold
(100 by default).
For spec2017int, this patch could reduce the number of specialized
functions by 33%. Then the compile time didn't increase for every
benchmark.
Reviewed By: SjoerdMeijer, xbolva00, snehasish
Differential Revision: https://reviews.llvm.org/D107897
This change enables llvm-profgen to use accurate context-sensitive post-optimization function byte size as a cost proxy to drive global preinline decisions.
To do this, BinarySizeContextTracker is introduced to track function byte size under different inline context during disassembling. In preinliner, we can not query context byte size under switch `context-cost-for-preinliner`. The tracker uses a reverse trie to keep size of functions under different context (callee as parent, caller as child), and it can give best/longest possible matching context size for given input context.
The new size cost is off by default. There're a few TODOs that needs to addressed: 1) avoid dangling string from `Offset2LocStackMap`, which will be addressed in split context work; 2) using inlinee's entry probe to make sure we have correct zero size for inlinee that's completely optimized away after inlining. Some tuning is also needed.
Differential Revision: https://reviews.llvm.org/D108180
This patch implements Flow Sensitive Sample FDO (FSAFDO) profile
loader. We have two profile loaders for FS profile,
one before RegAlloc and one before BlockPlacement.
To enable it, when -fprofile-sample-use=<profile> is specified,
add "-enable-fs-discriminator=true \
-disable-ra-fsprofile-loader=false \
-disable-layout-fsprofile-loader=false"
to turn on the FS profile loaders.
Differential Revision: https://reviews.llvm.org/D107878
Currently, AAKernelInfo will fail on an assertion if we attempt to run
it on a kernel without the init / deinit runtime calls. However, this
occurs for global constructors on the device. This will cause OpenMPOpt
to crash whenever global constructors are present. This patch removes
this assertion and just gives up instead.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D108258
In the provided test case, we were trying to set the global's
initializer to `i32* null` when the global's value type was `@0`.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D108232
This patch adds an extra option to print the module after running one of
the OpenMPOpt passes if debugging is enabled. This makes it much easier
to inspect the effects of this pass when doing debugging.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D108146
AttributeList::hasAttribute() is confusing, use clearer methods like
hasParamAttr()/hasRetAttr().
Add hasRetAttr() since it was missing from AttributeList.
Besides SPMDization, other analysis and optimization for original, frontend-generated SPMD regions uses information from the AAKernelInfoFunction attribute. This fix makes sure disabling SPMDization through the corresponding option applies only to generic mode regions, which should not be SPMDized, while it leaves unaffected the attribute state of original SPMD regions.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D108001
PHI nodes are not pass through but change their value, we have to
account for that to avoid missing stores.
Follow up for D107798 to fix PR51249 for good.
Differential Revision: https://reviews.llvm.org/D107808
AAPointerInfoFloating needs to visit all uses and some multiple times if
we go through PHI nodes. Attributor::checkForAllUses keeps a visited set
so we don't recurs endlessly. We now allow recursion for non-phi uses so
we track all pointer offsets via PHI nodes properly without endless
recursion.
This replaces the first attempt D107579.
Differential Revision: https://reviews.llvm.org/D107798
To avoid simplification with wrong constants we need to make sure we
know that we won't perform specific optimizations based on the users
request. The non-SPMDzation and non-CustomStateMachine flags did only
prevent the final transformation but allowed to value simplification
to go ahead.
Differential Revision: https://reviews.llvm.org/D107862
The may get changed before specialization by RunSCCPSolver. In other
words, the pass may change the function without specialization happens.
Add test and comment to reveal this.
And it may return No Changed if the function get changed by
RunSCCPSolver before the specialization. It looks like a potential bug.
Test Plan: check-all
Reviewed By: https://reviews.llvm.org/D107622
Differential Revision: https://reviews.llvm.org/D107622
Noticed that the computation for function specialization cost of a
function wouldn't change during the traversal of the arguments for the
function. We could hoist the computation out of the traversal. I
observed about ~1% improvement on compile time for spec2017. But I guess
it may not be precise. This should be NFC and fine.
Reviewed By: Sjoerd Meijer
Differential Revision: https://reviews.llvm.org/D107621
Now the recursive functions may get specialized many times when
`func-specialization-max-iters` increases. See discussion in
https://reviews.llvm.org/D106426 for details.
This patch expands SPMDization (converting generic execution mode to SPMD for target regions) by guarding code regions that should be executed only by the main thread. Specifically, it generates guarded regions, which only the main thread executes, and the synchronization with worker threads using simple barriers. For correctness, the patch aborts SPMDization for target regions if the same code executes in a parallel region, thus must be not be guarded. This check is implemented using the ParallelLevels AA.
Reviewed By: jhuber6
Differential Revision: https://reviews.llvm.org/D106892
This adds support for specialising recursive functions. For example:
int Global = 1;
void recursiveFunc(int *arg) {
if (*arg < 4) {
print(*arg);
recursiveFunc(*arg + 1);
}
}
void main() {
recursiveFunc(&Global);
}
After 3 iterations of function specialisation, followed by inlining of the
specialised versions of recursiveFunc, the main function looks like this:
void main() {
print(1);
print(2);
print(3);
}
To support this, the following has been added:
- Update the solver and state of the new specialised functions,
- An optimisation to propagate constant stack values after each iteration of
function specialisation, which is necessary for the next iteration to
recognise the constant values and trigger.
Specialising recursive functions is (at the moment) controlled by option
-func-specialization-max-iters and is opt-in for compile-time reasons. I.e.,
the default is -func-specialization-max-iters=1, but for the example above we
would need to use -func-specialization-max-iters=3. Future work is to see if we
can increase the default, or improve the cost-model/heuristics to control
compile-times.
Differential Revision: https://reviews.llvm.org/D106426
Currently, in OptimizeGlobalAddressOfMalloc, the transformation for global loads assumes that they have the same Type. With the support of ConstantExpr (https://reviews.llvm.org/D106589), this may not be true any more (as seen in the test case), and we miss the code to handle this, This is to fix that.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D107397
I'm working on extending the OptimizeGlobalAddressOfMalloc to handle some more general cases. This is to add support of the ConstantExpr use of the global variables. The function allUsesOfLoadedValueWillTrapIfNull is now iterative with the added CE use of GV. Also, the recursive function valueIsOnlyUsedLocallyOrStoredToOneGlobal is changed to iterative using a worklist with the GEP case added.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D106589
This work provides four flags to disable four different sets of OpenMP optimizations. These flags take effect in llvm/lib/Transforms/IPO/OpenMPOpt.cpp and include the following:
- openmp-opt-disable-deglobalization: Defaults to false, adding this flag sets the variable DisableOpenMPOptDeglobalization to true. This prevents AA registration for HeapToStack and HeapToShared.
- openmp-opt-disable-spmdization: Defaults to false, adding this flag sets the variable DisableOpenMPOptSPMDization to true. This indicates a pessimistic fixpoint in changeToSPMDMode.
- openmp-opt-disable-folding: Defaults to false, adding this flag sets the variable DisableOpenMPOptFolding to true. This indicates a pessimistic fixpoint in the attributor init for AAFoldRuntimeCall.
- openmp-opt-disable-state-machine-rewrite: Defaults to false, adding this flag sets the variable DisableOpenMPOptStateMachineRewrite to true. This first prevents changes to the state machine in rewriteDeviceCodeStateMachine by returning before changes are made, and if a custom state machine is built in buildCustomStateMachine, stops by returning a pessimistic fixpoint.
Reviewed By: jhuber6
Differential Revision: https://reviews.llvm.org/D106802
The current implementation of function internalization creats a copy of each
function and replaces every use. This has the downside that the external
versions of the functions will call into the internalized versions of the
functions. This prevents them from being fully independent of eachother. This
patch replaces the current internalization scheme with a method that creates
all the copies of the functions intended to be internalized first and then
replaces the uses as long as their caller is not already internalized.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106931
The device runtime contains several calls to `__kmpc_get_hardware_num_threads_in_block`
and `__kmpc_get_hardware_num_blocks`. If the thread_limit and the num_teams are constant,
these calls can be folded to the constant value.
In this patch we use the already introduced `AAFoldRuntimeCall` and the `NumTeams` and
`NumThreads` kernel attributes (to be introduced in a different patch) to fold these functions.
The code checks all the kernels, and if their attributes match, the functions are folded.
In the future we will explore specializing for multiple values of NumThreads and NumTeams.
Depends on D106390
Reviewed By: jdoerfert, JonChesterfield
Differential Revision: https://reviews.llvm.org/D106033
This reapplies commit cbb709e251 and
includes the use of the lookup method instead of operator[] to avoid
accidentally setting (empty) simplification callbacks.
This reverts commit aa27430a62.
AAValueSimplify, AAValueConstantRange, and AAPotentialValues all look at
the IR by default. If queried for a IR position which has a
simplification callback we should either look at the callback return, or
give up. We do the latter for now.
The current JumpThreading pass does not jump thread loops since it can
result in irreducible control flow that harms other optimizations. This
prevents switch statements inside a loop from being optimized to use
unconditional branches.
This code pattern occurs in the core_state_transition function of
Coremark. The state machine can be implemented manually with goto
statements resulting in a large runtime improvement, and this transform
makes the switch implementation match the goto version in performance.
This patch specifically targets switch statements inside a loop that
have the opportunity to be threaded. Once it identifies an opportunity,
it creates new paths that branch directly to the correct code block.
For example, the left CFG could be transformed to the right CFG:
```
sw.bb sw.bb
/ | \ / | \
case1 case2 case3 case1 case2 case3
\ | / / | \
latch.bb latch.2 latch.3 latch.1
br sw.bb / | \
sw.bb.2 sw.bb.3 sw.bb.1
br case2 br case3 br case1
```
Co-author: Justin Kreiner @jkreiner
Co-author: Ehsan Amiri @amehsan
Reviewed By: SjoerdMeijer
Differential Revision: https://reviews.llvm.org/D99205
Eliminating loads/stores in the device code is worth the extra effort,
especially for the new device runtime.
At the same time we do not compute AAExecutionDomain for non-device code
anymore, there is no point.
Differential Revision: https://reviews.llvm.org/D106845
When we simplify at least one operand in the Attributor simplification
we can use the InstSimplify to work on the simplified operands. This
allows us to avoid duplication of the logic.
Depends on D106189
Differential Revision: https://reviews.llvm.org/D106190
This change slightly relaxed the current ICP threshold in top-down inliner, specifically always allow one ICP for it. It shows some perf improvements on SPEC and our internal benchmarks. Also renamed the previous flag. We can also try to turn off PGO ICP in the future.
Reviewed By: wenlei, hoy, wmi
Differential Revision: https://reviews.llvm.org/D106588
D106185 allows us to determine if a store is needed easily. Using that
knowledge we can start to delete dead stores.
In AAIsDead we now track more state as an instruction can be dead (= the
old optimisitc state) or just "removable". A store instruction can be
removable while being very much alive, e.g., if it stores a constant
into an alloca or internal global. If we would pretend it was dead
instead of only removablewe we would ignore it when we determine what
values a load can see, so that is not what we want.
Differential Revision: https://reviews.llvm.org/D106188
This patch introduces `getPotentialCopiesOfStoredValue` which uses
AAPointerInfo to determine all "aliases" or "potential copies" of a
value that is stored into memory. This operation can fail but if it
succeeds it means we can visit all "uses" of a value even if it is
temporarily stored in memory.
There are two users for the function:
1) `Attributor::checkForAllUses` which will now ignore the value use
in a store if all "potential copies" can be identified and instead
be visited. This allows various AAs, including AAPointerInfo
itself, to look through memory.
2) `AANoCapture` which uses a custom use tracking through the
CaptureTracker interface and therefore needs to be thought
explicitly.
Differential Revision: https://reviews.llvm.org/D106185
Similar to D105787, this patch tries to fold `__kmpc_parallel_level` if possible.
Note that `__kmpc_parallel_level` doesn't take activeness into consideration,
based on current `deviceRTLs`, its return value can be such as 0, 1, 2, instead
of 0, 129, 130, etc. that also indicate activeness.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106154
While rewriteDeviceCodeStateMachine should probably be folded into
buildCustomStateMachine, we at least need the optimization to happen.
This was not reliably the case in the CGSCC pass but in the Module pass
it seems to work reliably.
This also ports a test to the new kernel encoding (target_init/deinit),
and makes sure we cannot run the kernel in SPMD mode.
Differential Revision: https://reviews.llvm.org/D106345
This caused us to rerun AAMemoryBehaviorFloating::updateImpl over and
over again. Unfortunately it turned out to be hard to reproduce the
behavior in a reasonable way.
If we add a new live edge we need to indicate a change or otherwise the
new live block is not shown to users. Similarly, new known dead ends and
a changed `ToBeExploredFrom` set need to cause us to return CHANGED.
Summary:
There was an unnecessary variable assigned to the information cache when we
only need it in the constructor to extract the function declaration.
This patch introduces a new RAII struct that will temporarily make an OpenMP
RTL function have external linkage. This is done before the attributor is
invoked to prevent it from incorrectly removing some function definitions that
we will use later. For example, if we determine all calls to one function are
dead, because it has internal linkage it can safely be removed. Later when we
try to get an instance to that function to modify the source using
`getOrCreateRuntimeFunction` we will then get an empty declaration for that
function that won't be defined anywhere. This patch prevents this from
occurring.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106707
Rather than adding methods for dropping these attributes in
various places, add a function that returns an AttrBuilder with
these attributes, which can then be used with existing methods
for dropping attributes. This is with an eye on D104641, which
also needs to drop them from returns, not just parameters.
Also be more explicit about the semantics of the method in the
documentation. Refer to UB rather than Undef, which is what this
is actually about.
This patch introduces a pass that uses the Attributor to deduce AMDGPU specific attributes.
Reviewed By: jdoerfert, arsenm
Differential Revision: https://reviews.llvm.org/D104997
checkForAllInstructions was not handling declarations correctly.
It should have been returning false when it gets called on a declaration
The patch also fixes a test case for AAFunctionReachability for it to be able
to pass after the changes to the checkForAllinstructions.
Differential Revision: https://reviews.llvm.org/D106625
Since we are using assumed information now, the logic should be refined to avoid
unncessary assertion.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106630
Deduplication in OpenMPOpt finds redundant OpenMP runtime calls and replaces them with a single call placed in the earliest safe location in the IR. When deduplication happens in a target region this patch makes sure replacement calls are put after target_init.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106556
A simplification callback can mean that the IR value is modified beyond
the apparent IR semantics. That is, a `i1 true` could be replaced by an
`i1 false` based on high-level domain-specific information. If a user
provides a simplification callback we will not look at the IR but
instead give up if the callback returns a nullptr.
SPMDization D102307 detects incompatible OpenMP runtime calls to abort converting a target region to SPMD mode. Calls to memory allocation/de-allocation routines kmpc_alloc_shared, kmpc_free_shared are incompatible unless they are removed by AAHeapToStack/AAHeapToShared analysis. This patch extends SPMDization detection to include AAHeapToStack/AAHeapToShared analysis results for enlarging the scope of possible SPMDized regions detected.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D105634
Fixing a typo in SampleContextTracker to use debug name when debug linkage name is no present. This should only affect C programs.
Saw 0.6% perf win on Cinder which is mostly C code.
Reviewed By: wenlei, wmi
Differential Revision: https://reviews.llvm.org/D106599
D101977 added `BooleanStateWithPtrSetVector` to store pointers to a set meanwhile
tracking boolean state. One of the limitation is that it can only store pointer.
We might want it to store other types of values, such as integer for parallel
level. This patch generalizes the idea and create `BooleanStateWithSetVector`.
`BooleanStateWithPtrSetVector` therefore becomes a type alias of `BooleanStateWithSetVector`.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106149
If we remove a non-intrinsic instruction we need to tell the (old) call
graph about it. This caused problems with some features down the line as
they allowed to removed calls more aggressively.
If we have a recursive function we could create multiple instantiations
of an SSA value, one per recursive invocation of the function. This is a
problem as we use SSA value equality in various places. The basic idea
follows from this test:
```
static int r(int c, int *a) {
int X;
return c ? r(false, &X) : a == &X;
}
int test(int c) {
return r(c, undef);
}
```
If we look through the argument `a` we will end up with `X`. Using SSA
value equality we will fold `a == &X` to true and return true even
though it should have been false because `a` and `&X` are from different
instantiations of the function.
Various tests for this have been placed in value-simplify-instances.ll
and this commit fixes them all by avoiding to produce simplified values
that could be non-unique at runtime. Thus, the result of a simplify
value call will always be unique at runtime or the original value, both
do not allow to accidentally compare two instances of a value with each
other and conclude they are equal statically (pointer equivalence) while
they are unequal at runtime.
A call that is analyzed in an optimization needs to be verified against
the name and type of the runtime function to avoid that we look at
arguments that do not exist (anymore). This can happen if the signature
was rewritten. Since we will not set RFI.Declaration if the type doesn't
match we can use it (if it's not null) to determine if the signature is
as expected.
Differential Revision: https://reviews.llvm.org/D106341
This patch strips the NoInline attribute from known OpenMP runtime functions.
This is done so that we can denote certain runtime functions as NoInline to
ensure their call sites are intact so they can be checked by OpenMPOpt. We
don't wan't this noinline attribute to remain for any functions after OpenMPOpt
has been run however.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106482
This patch adds the ability to fold `__kmpc_is_generic_main_thread_id` if we
know for a fact that it is executed by the initial thread using
AAExecutionDomain. This combined with folding `__kmpc_is_spmd_exec_mode` will
allow us to fully fold `__kmpc_is_generic_main_thread`.
Depends on D106438 D106437
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106439
Function internalization can sometimes occur in situations where we want to
keep the call sites intact. This patch adds an option to disable function
internalization and prevents the device runtime from being internalized while
creating the bitcode library.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106438
Qualified kernels can be transformed from generic-mode to SPMD mode using an
optimization in OpenMPOpt. This patch introduces a new execution mode to
indicate kernels that have been transformed from generic-mode to SPMD-mode.
These kernels have SPMD-mode execution, but need generic-mode semantics for
scheduling the blocks and threads. Without this far too few blocks will be
scheduled for a generic region as SPMD mode expects the trip count to be
divided by the number of threads.
Reviewed By: ggeorgakoudis
Differential Revision: https://reviews.llvm.org/D106460
Manifesting AbstractAttributes may add new BBs in the IR. This patch provides an interface to register those BBs in the Attributor so that those BBs and containing instructions are not deleted as dead.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106383
There is no need for a non-const argument interface and the const argument modification covers existing and upcoming use cases.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106418
In weird cases, the inliner will inline internal recursive functions,
sometimes causing them to have no more uses, in which case the
inliner will mark the function to be deleted. The function is
actually deleted after the call to
updateCGAndAnalysisManagerForCGSCCPass(). In
updateCGAndAnalysisManagerForCGSCCPass(), UR.UpdatedC may be set to
the SCC containing the function to be deleted. Then the inliner calls
CG.removeDeadFunction() which can cause that SCC to be deleted, even
though it's still stored in UR.UpdatedC.
We could potentially check in the wrappers/pass managers if UR.UpdatedC
is in UR.InvalidatedSCCs before doing anything with it, but it's safer
to do this as close to possible to the call to CG.removeDeadFunction()
to avoid issues with allocating a new SCC in the same address as
the deleted one.
It's hard to find a small test case since we need to have recursive
internal functions be reachable from non-internal functions, yet they
need to become non-recursive and not referenced by other functions when
inlined.
Similar to https://reviews.llvm.org/D106306.
Fixes PR50788.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D106405
In the textual format, `noduplicates` means no COMDAT/section group
deduplication is performed. Therefore, if both sets of sections are retained, and
they happen to define strong external symbols with the same names,
there will be a duplicate definition linker error.
In PE/COFF, the selection kind lowers to `IMAGE_COMDAT_SELECT_NODUPLICATES`.
The name describes the corollary instead of the immediate semantics. The name
can cause confusion to other binary formats (ELF, wasm) which have implemented/
want to implement the "no deduplication" selection kind. Rename it to be clearer.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D106319
Create an internal alias with the original name for static functions
that are renamed in promoteInternals to avoid breaking inline
assembly references to them. This version uses module inline assembly
to avoid issues with LowerTypeTestsModule.
Relands commmit 8e3b5cb39e with arch
specific tests fixed.
Link: https://github.com/ClangBuiltLinux/linux/issues/1354
Reviewed By: nickdesaulniers, pcc
Differential Revision: https://reviews.llvm.org/D104058
SPMDization in D102307 does not change the RequiresFullRuntime argument of kmpc_target_init/deinit calls. However, the constraints of SPMDization detection for converting a target region to SPMD mode should guarantee that the region does not require full runtime support. Hence, this patch sets RequiresFullRuntime to false for improved execution performance.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D105556
AAMemoryBehaviorFloating used a custom use tracking mechanism even
though checkForAllUses exists and is already more powerful. Further,
AAMemoryBehaviorFloating uses AANoCapture to guarantee that there are no
aliases and following the uses is sufficient. This is an OK assumption
if checkForAllUses is used but custom tracking is easily out of sync
with AANoCapture and problems follow.
As with other patches before, the simplification callback interface
requires us to go through the Attributor::getAssumedSimplified API first
before we recurs.
It is unclear if the problem can be explicitly tested with our current
infrastructure.
We first simplify the operands of a compare and then reason on the
simplified versions, e.g., with AANonNull.
This does improve the simplification capabilities but also fixes a
potential problem that has not yet been observed by simplifying the
operands first.
A byval argument is a different value in the caller and callee, we
cannot propagate the information as part of AAValueSimplify. Users that
want to deal with byval arguments need to specifically perform the
argument -> call site step. We do not do this for now.
This patch introduces AAPointerInfo which tracks the uses of a pointer
and places them in "bins" based on their offset from the base and access
size.
As with other AAs, any pointer can be tracked but it is up to the user
to make sense of the results. The user in this patch is AAValueSimplify
and AAPotentialValues which both utilize AAPointerInfo to determine the
value of a load. For now, this is restricted to loads of allocas and
internal globals. Through the use of AAPointerInfo and the "bins" we can
track struct members separately. The users also know that storing only
zeros (at unknown indices) will result in loading only 0 (from unknown
indices). Other than that, the users are flow and context insensitive
(for now).
To deal with the "bins" more easily, AAPointerInfo provides a
forallInterfearingAccesses that applies a callback on all accesses
that might interfere with a given load or store.
Differential Revision: https://reviews.llvm.org/D104432
As a first step to simplify loads we only handle `null` and `undef`
underlying objects, as well as objects that have the load as a single user.
Loads of those values can be replaced by the initializer, if any.
Proper reasoning is introduced in a follow up patch
Differential Revision: https://reviews.llvm.org/D103862
We did not properly use SPMDCompatibilityTracker in various places.
This patch makes sure we look at the validity properly and also fix
the state if we can.
Differential Revision: https://reviews.llvm.org/D106085
This patch fixed two issues found when folding `__kmpc_is_spmd_exec_mode`:
1. When the reaching kernels are empty, it should not fold to generic mode.
2. When creating AA for the caller when updating information, the dependency
should be required.
Reviewed By: ye-luo
Differential Revision: https://reviews.llvm.org/D106209
Iterative-BFI produces better count quality and performance when evaluated on internal benchmarks. Turning it on by default now for CSSPGO. We can consider turn it on by default for AutoFDO as well in the future.
Differential Revision: https://reviews.llvm.org/D106202
Create an internal alias with the original name for static functions
that are renamed in promoteInternals to avoid breaking inline
assembly references to them. This version uses module inline assembly
to avoid issues with LowerTypeTestsModule.
Link: https://github.com/ClangBuiltLinux/linux/issues/1354
Reviewed By: nickdesaulniers, pcc
Differential Revision: https://reviews.llvm.org/D104058
This patch adds unique idenfitiers to the existing OpenMP remarks. This makes
it easier to identify the corresponding documentation for each remark that will
be hosted in the OpenMP webpage.
Depends on D105898
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D105939
Dávid Bolvanský