This is a simple version without the possibility to define distribute
points or followup-transformations. However, it is the first
transformation that has to check whether the transformation is correct.
It interprets the same metadata as the LoopDistribute pass.
This metadata was intended to mark all accesses within an iteration to be pairwise non-aliasing, in this case because every memory of a base pointer is touched (read or write) at most once. This is typical for 'sweeps' over all data. The stated motivation from D30606 is to ensure that unrolled iterations are considered non-aliasing.
Rhe implemention had multiple issues:
* The structure of the noalias metadata was malformed. D110026 added check in the verifier for this metadata, and the tests were failing since then.
* This is not true for the outer loops of the BLIS matrix multiplication, where it was being inserted. Each element of A, B, C is accessed multiple times, as often as the loop not used as an index is iterating.
* Scopes were added to SecondLevelOtherAliasScopeList (used for the !noalias scop list) on-the-fly when another SCEV was seen. This meant that previously visited instructions would not be updated with alias scopes that are only seen later, missing out those SCEVs they should not be aliasing with.
* Since the !noalias scope list would ideally consists of all other SCEV for this base pointer, we might run quickly into scalability issues. Especially after unrolling there would probably at least once SCEV per instruction and unroll instance.
* The inter-iteration noalias base pointer was not removed after leaving the loop marked with it, effectively marking everything after it to noalias as well.
A solution I considered was to mark each instruction as non-aliasing with its own scope. The instruction itself would obviously alias itself, but such construction might also be considered invalid. Duplicating the instruction (e.g. due to speculation) would mark the instruction non-aliasing with its clone. I don't want to go into this territory, especially since the original motivation of determining unrolled instances as noalias based on SCEV is the what scev-aa does as well.
This effectively reverts D30606 and D35761.
Polly does not use the count program itself, but somewhere in lit it is
expected to exists. Otherwise, the following error occurs:
llvm-lit: llvm-project/llvm/utils/lit/lit/llvm/subst.py:133: fatal: Did not find count in ./bin
Code outside the SCoP will be executed recardless of the code versioning
runtime check introduced by CodeGeneration. Assumption made based on
that these are never executed in Polly-optimized code does not hold.
This fixes the miscompilation of MultiSource/Applications/lambda-0.1.3
The new pass manager does not allow adding module passes at the
-polly-position=before-vectorizer extension point. Introduce a
DumpFunctionPass that dumps only current function. In contrast to the
legacy pass manager's -polly-dump-before, each function will be dumped
into its own file. -polly-dump-before-file is still not supported.
The DumpFunctionPass uses llvm::CloneModule to copy the current function
into a new module and then write it into a file.
This is needed for having the functions isl_{set,map}_n_basic_{set,map}
exported to the C++ interface.
Some tests have been modified to reflect the isl changes.
Commit 4c7f820b2b changed the llvm.powi intrinsic to support
different 'int' sizes for the exponent. That happened to break
the IntrinsicToLibdeviceFunc mapping in PPCGCodeGeneration, which
obviously should have been updated as part of commit 4c7f820b2b
(https://reviews.llvm.org/D99439).
The shortcoming was found by buildbots that use
-DPOLLY_ENABLE_GPGPU_CODEGEN=ON
This patch should fixup the problem.
When we're remapping an AddRec, the AddRec constructed by a partial
rewrite might not make sense. This triggers an assertion complaining
it's not loop-invariant.
Instead of constructing the partially rewritten AddRec, just skip
straight to calling evaluateAtIteration.
Testcase was automatically reduced using llvm-reduce, so it's a little
messy, but hopefully makes sense.
Differential Revision: https://reviews.llvm.org/D102959
Only supported with -polly-position=early. Unfortunately, the
extension point callpack for VectorizerStart only passes a
FunctionPassManager, making it impossible to add a module pass.
This required support for the canonicalization passes, inlcuding
porting RewriteByReferenceParams to the NPM.
For some reason, the legacy pass pipeline with -polly-position=early did
not run the CodePreparation pass. This was fixed as well.
We previously had a different interpretation of unroll transformation
attributes than how LoopUnroll interpreted it. In particular,
llvm.loop.unroll.enable was needed explicitly to enable it and disabling
metadata was ignored.
Additionally, it required that either full unrolling or an unroll factor
to be specified or fail otherwise. An unroll factor is still required,
but the transformation is ignored with the hope that LoopUnroll is going
to apply the unrolling, since Polly currently does not implement an
heuristic.
Fixes llvm.org/PR50109
We enumerated the cross product Domain x Scatter, but sorted only be the
scatter key. In case there are are multiple statement instances per
scatter value, the order between statement instances of the same loop
iteration was undefined.
Propertly enumerate and sort only by the scatter value, and group the
domains using the scatter dimension again.
Thanks to Leonard Chan for the report.
Make Polly look for unrolling metadata (https://llvm.org/docs/TransformMetadata.html#loop-unrolling) that is usually only interpreted by the LoopUnroll pass and apply it to the SCoP's schedule.
While not that useful by itself (there already is an unroll pass), it introduces mechanism to apply arbitrary loop transformation directives in arbitrary order to the schedule. Transformations are applied until no more directives are found. Since ISL's rescheduling would discard the manual transformations and it is assumed that when the user specifies the sequence of transformations, they do not want any other transformations to apply. Applying user-directed transformations can be controlled using the `-polly-pragma-based-opts` switch and is enabled by default.
This does not influence the SCoP detection heuristic. As a consequence, loop that do not fulfill SCoP requirements or the initial profitability heuristic will be ignored. `-polly-process-unprofitable` can be used to disable the latter.
Other than manually editing the IR, there is currently no way for the user to add loop transformations in an order other than the order in the default pipeline, or transformations other than the one supported by clang's LoopHint. See the `unroll_double.ll` test as example that clang currently is unable to emit. My own extension of `#pragma clang loop` allowing an arbitrary order and additional transformations is available here: https://github.com/meinersbur/llvm-project/tree/pragma-clang-loop. An effort to upstream this functionality as `#pragma clang transform` (because `#pragma clang loop` has an implicit transformation order defined by the loop pipeline) is D69088.
Additional transformations from my downstream pragma-clang-loop branch are tiling, interchange, reversal, unroll-and-jam, thread-parallelization and array packing. Unroll was chosen because it uses already-defined metadata and does not require correctness checks.
Reviewed By: sebastiankreutzer
Differential Revision: https://reviews.llvm.org/D97977
This reverts commit 329aeb5db4,
and relands commit 61f006ac65.
This is a continuation of D89456.
As it was suggested there, now that SCEV models `PtrToInt`,
we can try to improve SCEV's pointer handling.
In particular, i believe, i will need this in the future
to further fix `SCEVAddExpr`operation type handling.
This removes special handling of `ConstantPointerNull`
from `ScalarEvolution::createSCEV()`, and add constant folding
into `ScalarEvolution::getPtrToIntExpr()`.
This way, `null` constants stay as such in SCEV's,
but gracefully become zero integers when asked.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D98147
This is a continuation of D89456.
As it was suggested there, now that SCEV models `PtrToInt`,
we can try to improve SCEV's pointer handling.
In particular, i believe, i will need this in the future
to further fix `SCEVAddExpr`operation type handling.
This removes special handling of `ConstantPointerNull`
from `ScalarEvolution::createSCEV()`, and add constant folding
into `ScalarEvolution::getPtrToIntExpr()`.
This way, `null` constants stay as such in SCEV's,
but gracefully become zero integers when asked.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D98147
Emit llvm.loop.parallel_accesses metadata instead of
llvm.mem.parallel_loop_access. The latter is deprecated because it
assumes that LoopIDs are persistent, which they are not.
We also emit parallel access metadata for all surrounding parallel
loops, not just the innermost parallel.
DetectionContext objects are stored as values in a DenseMap. When the
DenseMap reaches its maximum load factor, it is resized and all its
objects moved to a new memory allocation. Unfortunately Scop object have
a reference to its DetectionContext. When the DenseMap resizes, all the
DetectionContexts reference now point to invalid memory, even if caused
by an unrelated DetectionContext.
Even worse, NewPM's ScopPassManager called isMaxRegionInScop with the
Verify=true parameter before each pass. This caused the old
DetectionContext to be removed an a new on created and re-verified.
Of course, the Scop object was already created pointing to the old
DetectionContext. Because the new DetectionContext would
usually be stored at the same position in the DenseMap, the reference
would usually reference the new DetectionContext of the same Region.
Usually.
If not, the old position still points to memory in the DenseMap
allocation (unless also a resizing occurs) such that tools like Valgrind
and AddressSanitizer would not be able to diagnose this.
Instead of storing the DetectionContext inside the DenseMap, use a
std::unique_ptr to a DetectionContext allocation, i.e. it will not move
around anymore. This also allows use to remove the very strange
DetectionContext(const DetectionContext &&)
copy/move(?) constructor. DetectionContext objects now are neither
copied nor moved.
As a result, every re-verification of a DetectionContext will use a new
allocation. Therefore, once a Scop object has been created using a
DetectionContext, it must not be re-verified (the Scop data structure
requires its underlying Region to not change before code generation
anyway). The NewPM may call isMaxRegionInScop only with
Validate=false parameter.
In addition to that regression tests should not test the intire pass
pipeline (unless they are testing the pipeline itself), the Polly-ACC
currently does not support the new pass manager. If enabled by default,
such tests will therefore fail.
Use the -polly-gpu-runtime and -polly-gpu-arch options also as default
values for the PPCGCodeGeneration pass. This requires to move the option
to be moved from the pipeline-building Register passes to the
PPCGCodeGeneration implementation.
Fixes the spir-typesize.ll buildbot fail.
ZoneAlgorithms's computePHI relies on being provided with consistent a
schedule to compute the statement prodecessors of a statement containing
PHINodes. Otherwise unexpected results such as PHI nodes with multiple
predecessors can occur which would result in problems in the
algorithms expecting consistent data.
In the added test case, statement instances are scrubbed from the
SCoP their execution would result in undefined behavior (Due to a nsw
overflow). As already being undefined behavior in LLVM-IR, neither
AssumedContext nor InvalidContext are updated, giving computePHI no
means to avoid these cases.
Intoduce a new SCoP property, the DefinedBehaviorContext, that among
the runtime-checked conditions, also tracks the assumptions not needing
a runtime check, in particular those affecting the assumed control flow.
This replaces the manual combination of the 3 other contexts that was
already done in computePHI and setNewAccessRelation. Currently, the only
additional assumption is that loop induction variables will nsw flag for
not wrap, but potentially more can be added. Use in
hasFeasibleRuntimeContext, isl::ast_build and gisting are other
potential uses.
To limit computational complexity, the DefinedBehaviorContext is not
availabe if it grows too large (atm hardcoded to 8 disjuncts).
Possible other fixes include bailing out in computePHI when
inconsistencies are detected, choose an arbitrary value for inconsistent
cases (since it is undefined behavior anyways), or make the code
receiving the result from ComputePHI handle inconsistent data. All of
them reduce the quality of implementation having to bail out more often
and disabling the ability to assert on actually wrong results.
This fixes llvm.org/PR48783.
In preparation for turning on opt's -enable-new-pm by default, this pins
uses of passes via the legacy "opt -passname" with pass names beginning
with "polly-" and "polyhedral-info" to the legacy PM. Many of these
tests use -analyze, which isn't supported in the new PM.
(This doesn't affect uses of "opt -passes=passname").
rL240766 accidentally removed `-polly-prepare` in
phi_not_grouped_at_top.ll, and it also doesn't use the output of
-analyze.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D94266
This fixes llvm.org/PR48554
Some test cases had to be updated because the hash function for
union_maps have been changed which affects the output order.
This patch updates IRBuilder to create insertelement/shufflevector using poison as a placeholder.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D93793
MemoryAccess::setNewAccessRelation() in assert-builds checks whether the
access relation for a READ has a memory location for every instance of
the domain. Otherwise, we would not have value to load from. That check
already considered that instances outside the Scop's context do not
matter since they are never executed (or would be undefined behavior).
In this patch also take instances of the InvalidContext into account,
as these can also be assumed to never occur. InvalidContext was
introduced to avoid the computational complexity of subtracting
restrictions from the AssumedContext. However, this additional check in
setNewAccessRelation is only done in assert-builds.
The assertion case with an InvalidContext may occur with DeLICM on a
conditionally infinite loops, as it is the case in the following code:
for (int i = 0; i < n; i+=b)
vreg = ...;
*Dest = vreg;
The loop is infinite when b=0, and [b] -> { : b = 0 } is part of the
InvalidContext. When DeLICM tries to map the memory for %vreg to *Dest,
there is no store instance that uses the value of vreg when b = 0, hence
no location to map it to. However, the case is irrelevant since Polly's
runtime condition check ensures that this is never case.
Fixes llvm.org/PR48445
ScalarEvolution::getSCEV cannot be used during codegen. ScalarEvolution
assumes a stable IR and control flow which is under construction during
Polly's CodeGen. In particular, it uses DominatorTree for compute the
backedge taken count. However the DominatorTree is not updated during
codegen.
In this case, SCEV was used to determine the base pointer of an array
access. Replace it by our own function. Polly generates only GEP and
BitCasts for array acceses, i.e. it is sufficient to handle these to to
find the base pointer.
Fixes llvm.org/PR48422
Operand tree forwarding can cause the change of an access kind; in
particular change from a scalar kind to an array kind if the scalar
dependency is not necessary. Such an access cannot and doesn't need to
be forwarded anymore.
Fixes llvm.org/PR48034
Michael Kruse