Summary:
Both `canUseISLTripCount()` and `addOverApproximatedRegion()` contained checks
to reject endless loops which are now removed and replaced by a single check
in `isValidLoop()`.
For reporting such loops the `ReportLoopOverlapWithNonAffineSubRegion` is
renamed to `ReportLoopHasNoExit`. The test case
`ReportLoopOverlapWithNonAffineSubRegion.ll` is adapted and renamed as well.
The schedule generation in `buildSchedule()` is based on the following
assumption:
Given some block B that is contained in a loop L and a SESE region R,
we assume that L is contained in R or the other way around.
However, this assumption is broken in the presence of endless loops that are
nested inside other loops. Therefore, in order to prevent erroneous behavior
in `buildSchedule()`, r265280 introduced a corresponding check in
`canUseISLTripCount()` to reject endless loops. Unfortunately, it was possible
to bypass this check with -polly-allow-nonaffine-loops which was fixed by adding
another check to reject endless loops in `allowOverApproximatedRegion()` in
r273905. Hence there existed two separate locations that handled this case.
Thank you Johannes Doerfert for helping to provide the above background
information.
Reviewers: Meinersbur, grosser
Subscribers: _jdoerfert, pollydev
Differential Revision: https://reviews.llvm.org/D24560
Contributed-by: Matthias Reisinger <d412vv1n@gmail.com>
llvm-svn: 281987
In case sequential kernels are found deeper in the loop tree than any parallel
kernel, the overall scop is probably mostly sequential. Hence, run it on the
CPU.
llvm-svn: 281849
Offloading to a GPU is only beneficial if there is a sufficient amount of
compute that can be accelerated. Many kernels just have a very small number
of dynamic compute, which means GPU acceleration is not beneficial. We
compute at run-time an approximation of how many dynamic instructions will be
executed and fall back to CPU code in case this number is not sufficiently
large. To keep the run-time checking code simple, we over-approximate the
number of instructions executed in each statement by computing the volume of
the rectangular hull of its iteration space.
llvm-svn: 281848
We may generate GPU kernels that store into scalars in case we run some
sequential code on the GPU because the remaining data is expected to already be
on the GPU. For these kernels it is important to not keep the scalar values
in thread-local registers, but to store them back to the corresponding device
memory objects that backs them up.
We currently only store scalars back at the end of a kernel. This is only
correct if precisely one thread is executed. In case more than one thread may
be run, we currently invalidate the scop. To support such cases correctly,
we would need to always load and store back from a corresponding global
memory slot instead of a thread-local alloca slot.
llvm-svn: 281838
Our alias checks precisely check that the minimal and maximal accessed elements
do not overlap in a kernel. Hence, we must ensure that our host <-> device
transfers do not touch additional memory locations that are not covered in
the alias check. To ensure this, we make sure that the data we copy for a
given array is only the data from the smallest element accessed to the largest
element accessed.
We also adjust the size of the array according to the offset at which the array
is actually accessed.
An interesting result of this is: In case array are accessed with negative
subscripts ,e.g., A[-100], we automatically allocate and transfer _more_ data to
cover the full array. This is important as such code indeed exists in the wild.
llvm-svn: 281611
This is the fourth patch to apply the BLIS matmul optimization pattern on matmul
kernels (http://www.cs.utexas.edu/users/flame/pubs/TOMS-BLIS-Analytical.pdf).
BLIS implements gemm as three nested loops around a macro-kernel, plus two
packing routines. The macro-kernel is implemented in terms of two additional
loops around a micro-kernel. The micro-kernel is a loop around a rank-1
(i.e., outer product) update. In this change we perform copying to created
arrays, which is the last step to implement the packing transformation.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D23260
llvm-svn: 281441
This line makes BUILD_SHARED_LIBS=ON work for Polly-ACC. Without it, ld
complains about missing isl symbols when constructing the shared library.
llvm-svn: 281396
The alias to the array element is read-only and a primitive type (pointer),
therefore use the value directly instead of a reference to it.
llvm-svn: 281311
The flag -fvisibility=hidden flag was used for the integrated Integer
Set Library (and PPCG) to keep their definitions local to Polly. The
motivation was the be loaded into a DragonEgg-powered GCC, where GCC
might itself use ISL for its Graphite extension. The symbols of Polly's
ISL and GCC's ISL would clash.
The DragonEgg project is not actively developed anymore, but Polly's
unittests need to call ISL functions to set up a testing environment.
Unfortunately, the -fvisibility=hidden flag means that the ISL symbols
are not available to the gtest executable as it resides outside of
libPolly when linked dynamically. Currently, CMake links a second copy
of ISL into the unittests which leads to subtle bugs. What got observed
is that two isl_ids for isl_id_none exist, one for each library
instance. Because isl_id's are compared by address, isl_id_none could
happen to be different from isl_id_none, depending on which library
instance set the address and does the comparison.
Also remove the FORCE_STATIC flag which was introduced to keep the ISL
symbols visible inside the same libPolly shared object, even when build
with BUILD_SHARED_LIBS.
Differential Revision: https://reviews.llvm.org/D24460
llvm-svn: 281242
We do not need the size of the outermost dimension in most cases, but if we
allocate memory for newly created arrays, that size is needed.
Reviewed-by: Michael Kruse <llvm@meinersbur.de>
Differential Revision: https://reviews.llvm.org/D23991
llvm-svn: 281234
Instead of aborting, we now bail out gracefully in case the kernel IR we
generate is invalid. This can currently happen in case the SCoP stores
pointer values, which we model as arrays, as data values into other arrays. In
this case, the original pointer value is not available on the device and can
consequently not be stored. As detecting this ahead of time is not so easy, we
detect these situations after the invalid IR has been generated and bail out.
llvm-svn: 281193
If these arrays have never been accessed we failed to derive an upper bound
of the accesses and consequently a size for the outermost dimension. We
now explicitly check for empty access sets and then just use zero as size
for the outermost dimension.
llvm-svn: 281165
The -polly-flatten-schedule pass reduces the number of scattering
dimensions in its isl_union_map form to make them easier to understand.
It is not meant to be used in production, only for debugging and
regression tests.
To illustrate, how it can make sets simpler, here is a lifetime set
used computed by the porposed DeLICM pass without flattening:
{ Stmt_reduction_for[0, 4] -> [0, 2, o2, o3] : o2 < 0;
Stmt_reduction_for[0, 4] -> [0, 1, o2, o3] : o2 >= 5;
Stmt_reduction_for[0, 4] -> [0, 1, 4, o3] : o3 > 0;
Stmt_reduction_for[0, i1] -> [0, 1, i1, 1] : 0 <= i1 <= 3;
Stmt_reduction_for[0, 4] -> [0, 2, 0, o3] : o3 <= 0 }
And here the same lifetime for a semantically identical one-dimensional
schedule:
{ Stmt_reduction_for[0, i1] -> [2 + 3i1] : 0 <= i1 <= 4 }
Differential Revision: https://reviews.llvm.org/D24310
llvm-svn: 280948
... to preserve reference counting logic.
In practice the missing assignment would not have caused any issues. We still
fix it as the code is wrong and it also causes noise in the clang static
analysis runs.
llvm-svn: 280946
When running the clang static analyser to check for memory issues, this code
originally showed a double free, as the analyser was unable to understand that
isl_set_free always returns NULL and consequently later uses of the isl object
we just freed will never be reached. Without this knowledge, the analyser has
to issue a warning.
We refactor the code to make it clear that for empty maps the current loop
iteration is aborted.
llvm-svn: 280940
When running the clang static analyser to check for memory issues, this code
originally showed a double free, as the analyser was unable to understand that
isl_union_map_free always returns NULL and consequently later uses of the isl
object we just freed will never be reached. Without this knowledge, the analyser
has to issue a warning.
We refactor the code to make it clear that for empty maps the current loop
iteration is aborted.
llvm-svn: 280938
Disable some Visual C++ warnings on ISL. These are not reported by GCC/Clang in
the ISL build system. We do not intend to fix them in the Polly in-tree copy,
hence disable these warnings.
llvm-svn: 280811
The check-polly-tests target runs regression/unit tests but without checking
formatting. This is useful to not having to reload a file in an open editor
(which eg. clears the undo buffer, moves cursor/window position) when running
polly-update-format.
After this change, the following test targets exist:
- check-polly-unittests to run unittests only
- check-polly-tests to run unit and regression tests
- polly-check-format to check formatting using clang-format
- check-polly to run them all
As a side-effect, when running check-polly, polly-check-format and run in
parallel (instead of polly-check-format first).
Differential Revision: https://reviews.llvm.org/D24191
llvm-svn: 280654
... but instead rely on the assumptions that we derive for load/store
instructions.
Before we were able to delinearize arrays, we used GEP pointer instructions
to derive information about the likely range of induction variables, which
gave us more freedom during loop scheduling. Today, this is not needed
any more as we delinearize multi-dimensional memory accesses and as part
of this process also "assume" that all accesses to these arrays remain
inbounds. The old derive-assumptions-from-GEP code has consequently become
mostly redundant. We drop it both to clean up our code, but also to improve
compile time. This change reduces the scop construction time for 3mm in
no-asserts mode on my machine from 48 to 37 ms.
llvm-svn: 280601
Without reductions we do not need a flat union_map schedule describing
the computation we want to perform, but can work purely on the schedule
tree. This reduces the dependence computation and scheduling time from 33ms
to 25ms. Another 30% reduction.
llvm-svn: 280558
In case we do not compute reduction dependences or dependences that are more
fine-grained than statement level dependences, we can avoid the corresponding
part of the dependence analysis all together. For the 3mm benchmark, this
reduces scheduling + dependence analysis time from 62ms to 33ms for a no-asserts
build. The majority of the compile time is anyhow spent in the LLVM backends,
when doing code generation. Nevertheless, there is no need to waste compile time
either.
llvm-svn: 280557
We replace the options
-polly-code-generator=none
=isl
with the options
-polly-code-generation=none
=ast
=full
This allows us to measure the overhead of Polly itself, versus the compile
time increases due to us generating more IR and consequently the LLVM backends
spending more time on this IR.
We also use this opportunity to rename the option. The original name was
introduced at a point where we still had two code generators. CLooG and the
isl AST generator. Since we only have one AST generator left, there is no need
to distinguish between 'isl' and something else. However, being able to disable
code generation all together has been shown useful for debugging. Hence, we
rename and extend this option to make it a good fit for its new use case.
llvm-svn: 280554
LLVM's coding guideline suggests to not use @brief for one-sentence doxygen
comments to improve readability. Switch this once and for all to ensure people
do not copy @brief comments from other parts of Polly, when writing new code.
llvm-svn: 280468
Change the code around setNewAccessRelation to allow to use a an existing array
element for memory instead of an ad-hoc alloca. This facility will be used for
DeLICM/DeGVN to convert scalar dependencies into regular ones.
The changes necessary include:
- Make the code generator use the implicit locations instead of the alloca ones.
- A test case
- Make the JScop importer accept changes of scalar accesses for that test case.
- Adapt the MemoryAccess interface to the fact that the MemoryKind can change.
They are named (get|is)OriginalXXX() to get the status of the memory access
before any change by setNewAccessRelation() (some properties such as
getIncoming() do not change even if the kind is changed and are still
required). To get the modified properties, there is (get|is)LatestXXX(). The
old accessors without Original|Latest become synonyms of the
(get|is)OriginalXXX() to not make functional changes in unrelated code.
Differential Revision: https://reviews.llvm.org/D23962
llvm-svn: 280408
There are some constraints on maps that can be access relations. In builds with assertions enabled, verify
- The access domain is the same space as the statement's domain (modulo parameters).
- Whether an access is defined for every instance of the statement. (codegen does not yet support partial access relations)
- Whether the access range links to an array, represented by a ScopArrayInfo.
- The number of access dimensions equals the dimensions of the array.
- The array is not an indirect access. (also not supported by codegen)
Differential Revision: https://reviews.llvm.org/D23916
llvm-svn: 280404
isl_val_int_from_ui takes an 'unsigned long' which has on 32-bit and LLP64
windows systems only 32 bit. Hence, make sure we do not use it with constants
that are larger than 32 bit.
Reported-by: Michael Kruse <llvm@meinersbur.de>
llvm-svn: 279824
This improves the readability of failing test results, as gtest prints always
the first argument as the 'expected value'.
In the previous commit we already changed the tests for isl_valFromAPInt. In
this commit, the tests for IslValToAPInt follow.
Suggested-by: Michael Kruse <llvm@meinersbur.de>
llvm-svn: 279817
The recent unit tests we gained made clear that the semantics of
isl_valFromAPInt are not clear, due to missing documentation. In this change we
document both the calling interface as well as the implementation of
isl_valFromAPInt.
We also make the implementation easier to read by removing integer wrappig in
abs() when passing in the minimal integer value for a given bitwidth. Even
though wrapping and subsequently interpreting the result as unsigned value gives
the correct result, this is far from obvious. Instead, we explicitly add one
more bit to the input type to ensure that abs will never wrap. This change did
not uncover a bug in the old implementation, but was introduced to increase
readability.
We update the tests to add a test case for this special case and use this
opportunity to also test a number larger than 64 bit. Finally, we order the
arguments of the test cases to make sure the expected output is first. This
helps readability in case of failing test cases as gtest assumes the first value
to be the exected value.
Reviewed-by: Michael Kruse <llvm@meinersbur.de>
Differential Revision: https://reviews.llvm.org/D23917
llvm-svn: 279815
The recent unit tests we gained made clear that the semantics of APIntFromVal
are not clear, due to missing documentation. In this change we document both
the calling interface as well as the implementation of APIntFromVal. We also
make the implementation easier to read by removing the use of magic numbers.
Finally, we add tests to check the bitwidth of the created values as well as
the correct modeling of very large numbers.
Reviewed-by: Michael Kruse <llvm@meinersbur.de>
Differential Revision: https://reviews.llvm.org/D23910
llvm-svn: 279813
Remove the unused function get_system_libs. Instead, run
'llvm-config --system-libs' to determine which libraries are required in
addition LLVM's for linking an executable. At the moment these are the unittests
that link to gtest and transitively depend on these system libs.
llvm-svn: 279743
Only build them for check-polly and check-polly-unittests in out-of-tree builds.
In LLVM, this behaviour is controlled with LLVM_BUILD_TESTS. Polly out-of-tree
does not have such a flag.
llvm-svn: 279740
We cannot built ISL as shared object because we build it with
-fvisibility=hidden; The created shared object would have no accessible symbols.
The reason it is built with -fvisibility=hidden is because opt/clang might load
other libraries that have ISL embedded and whose' symbols would conflict with
Polly's private ISL. This could happend with Draggonegg.
In the future we might instead statically link PollyISL into the Polly shared
object to avoid installing the static library.
Requested-by: Vedran Miletic <vedran@miletic.net>
See-also: llvm.org/PR27306
llvm-svn: 279737
Add the infrastructure for unittests to Polly and two simple tests for
conversion between isl_val and APInt. In addition, a build target
check-polly-unittests is added to run only the unittests but not the regression
tests.
Clang's unittest mechanism served as as a blueprint which then was adapted to
Polly.
Differential Revision: https://reviews.llvm.org/D23833
llvm-svn: 279734
configure_lit_site_cfg defines some more parameters that are used in
lit.site.cfg.in. configure_file would leave those empty. These additional
definitions seem to be unimportant for regression tests, but unittests do not
work without them.
In case of out-of-tree builds, define the additional parameters with default
values. These may not take all configuration parameters into account, as
configure_lit_site_cfg would.
llvm-svn: 279733
This previously was not required because in an out-of-tree build Polly would
only build libraries (LLVMPolly, libPolly, libPollyISL, libPollyPPCG), but no
executables where the libraries would be linked to. This will change when adding
unittests in a follow-up commit.
llvm-svn: 279730
The program 'llvm-lit', like 'not' and 'FileCheck' are necessary for running
check-polly. Warn of any of the three is not in LLVM_INSTALL_ROOT/bin directory.
llvm-svn: 279728
Dump polyhedral descriptions of Scops optimized with the isl scheduling
optimizer and the set of post-scheduling transformations applied
on the schedule tree to be able to check the work of the IslScheduleOptimizer
pass at the polyhedral level.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D23740
llvm-svn: 279395
getAccessFunctions() is dead code and the 'BB' argument
of getOrCreateAccessFunctions() is not used. This patch deletes
getAccessFunctions and transforms AccFuncMap into
a std::vector<std::unique_ptr<MemoryAccess>> AccessFunctions.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D23759
llvm-svn: 279394
The existing code would add the operands in the wrong order, and eventually
crash because the SCEV expression doesn't exactly match the parameter SCEV
expression in SCEVAffinator::visit. (SCEV doesn't sort the operands to
getMulExpr in general.)
Differential Revision: https://reviews.llvm.org/D23592
llvm-svn: 279087
We already invalidated a couple of critical values earlier on, but we now
invalidate all instructions contained in a scop after the scop has been code
generated. This is necessary as later scops may otherwise obtain SCEV
expressions that reference values in the earlier scop that before dominated
the later scop, but which had been moved into the conditional branch and
consequently do not dominate the later scop any more. If these very values are
then used during code generation of the later scop, we generate used that are
dominated by the values they use.
This fixes: http://llvm.org/PR28984
llvm-svn: 279047
Normally this is ensured when adding PHI nodes, but as PHI node dependences
do not need to be added in case all incoming blocks are within the same
non-affine region, this was missed.
This corrects an issue visible in LNT's sqlite3, in case invariant load hoisting
was disabled.
llvm-svn: 278792
With invariant load hoisting enabled the LLVM buildbots currently show some
miscompiles, which are possibly caused by invariant load hosting itself.
Confirming and fixing this requires a more in-depth analysis. To meanwhile get
back green buildbots that allow us to observe other regressions, we disable
invariant code hoisting temporarily. The relevant bug is tracked at:
http://llvm.org/PR28985
llvm-svn: 278681
This will make it easier to switch the default of Polly's invariant load
hoisting strategy and also makes it very clear that these test cases
indeed require invariant code hoisting to work.
llvm-svn: 278667
This is the third patch to apply the BLIS matmul optimization pattern on matmul
kernels (http://www.cs.utexas.edu/users/flame/pubs/TOMS-BLIS-Analytical.pdf).
BLIS implements gemm as three nested loops around a macro-kernel, plus two
packing routines. The macro-kernel is implemented in terms of two additional
loops around a micro-kernel. The micro-kernel is a loop around a rank-1
(i.e., outer product) update. In this change we perform replacement of
the access relations and create empty arrays, which are steps to implement
the packing transformation. In subsequent changes we will implement copying
to created arrays.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: http://reviews.llvm.org/D22187
llvm-svn: 278666
To do so we change the way array exents are computed. Instead of the precise
set of memory locations accessed, we now compute the extent as the range between
minimal and maximal address in the first dimension and the full extent defined
by the sizes of the inner array dimensions.
We also move the computation of the may_persist region after the construction
of the arrays, as it relies on array information. Without arrays being
constructed no useful information is computed at all.
llvm-svn: 278212
Ensure the right scalar allocations are used as the host location of data
transfers. For the device code, we clear the allocation cache before device
code generation to be able to generate new device-specific allocation and
we need to make sure to add back the old host allocations as soon as the
device code generation is finished.
llvm-svn: 278126
This increases the readability of the IR and also clarifies that the GPU
inititialization is executed _after_ the scalar initialization which needs
to before the code of the transformed scop is executed.
Besides increased readability, the IR should not change. Specifically, I
do not expect any changes in program semantics due to this patch.
llvm-svn: 278125
In case some code -- not guarded by control flow -- would be emitted directly in
the start block, it may happen that this code would use uninitalized scalar
values if the scalar initialization is only emitted at the end of the start
block. This is not a problem today in normal Polly, as all statements are
emitted in their own basic blocks, but Polly-ACC emits host-to-device copy
statements into the start block.
Additional Polly-ACC test coverage will be added in subsequent changes that
improve the handling of PHI nodes in Polly-ACC.
llvm-svn: 278124
After having generated the code for a ScopStmt, we run a simple dead-code
elimination that drops all instructions that are known to be and remain unused.
Until this change, we only considered instructions for dead-code elimination, if
they have a corresponding instruction in the original BB that belongs to
ScopStmt. However, when generating code we do not only copy code from the BB
belonging to a ScopStmt, but also generate code for operands referenced from BB.
After this change, we now also considers code for dead code elimination, which
does not have a corresponding instruction in BB.
This fixes a bug in Polly-ACC where such dead-code referenced CPU code from
within a GPU kernel, which is possible as we do not guarantee that all variables
that are used in known-dead-code are moved to the GPU.
llvm-svn: 278103
The function expandRegion() frees Region* objects again when it determines that
these are not valid SCoPs. However, the DetectionContext added to the
DetectionContextMap still holds a reference. The validity is checked using the
ValidRegions lookup table. When a new Region is added to that list, it might
share the same address, such that the DetectionContext contains two
Region* associations that are in ValidRegions, but that are unrelated and of
which one has already been free.
Also remove the DetectionContext when not a valid expansion.
llvm-svn: 278062
When adding code that avoids to pass values used in isl expressions and
LLVM instructions twice, we forgot to make single variable passed to the
kernel available in the ValueMap that makes it usable for instructions that
are not replaced with isl ast expressions. This change adds the variable
that is passed to the kernel to the ValueMap to ensure it is available
for such use cases as well.
llvm-svn: 278039
There is no need to reset the position of the builder, as we can just continue
to insert code at the current position of the IRBuilder, which happens to
be precisely the location we reset the builder to.
llvm-svn: 278014
... instead of adding instructions at the end of the basic block the builder
is currently at. This makes it easier to reason about where IR is generated,
as with the IRBuilder there is just a single location that specificies where
IR is generated.
llvm-svn: 278013
The map is iterated over when generating the values escaping the SCoP. The
indeterministic iteration order of DenseMap causes the output IR to change at
every compilation, adding noise to comparisons.
Replace DenseMap by a MapVector to ensure the same iteration order at every
compilation.
llvm-svn: 277832
Tobias Grosser