Only when load-hoisted we can be sure the base pointer is invariant
during the SCoP's execution. Most of the time it would be added to
the required hoists for the alias checks anyway, except with
-polly-ignore-aliasing, -polly-use-runtime-alias-checks=0 or if
AliasAnalysis is already sure it doesn't alias with anything
(for instance if there is no other pointer to alias with).
Two more parts in Polly assume that this load-hoisting took place:
- setNewAccessRelation() which contains an assert which tests this.
- BlockGenerator which would use to the base ptr from the original
code if not load-hoisted (if the access expression is regenerated)
Differential Revision: https://reviews.llvm.org/D30694
llvm-svn: 297195
There is no point in optimizing unreachable code, hence our test cases should
always return.
This commit is part of a series that makes Polly more robust on the presence of
unreachables.
llvm-svn: 297158
These test cases should work in combination with
https://reviews.llvm.org/D12676, but became outdated over time. Update them
in preparation of discussions with Daniel Berlin on how to represent unreachable
in the post-dominator tree.
llvm-svn: 297157
Scops that exit with an unreachable are today still permitted, but make little
sense to optimize. We therefore can already skip them during scop detection.
This speeds up scop detection in certain cases and also ensures that bugpoint
does not introduce unreachables when reducing test cases.
In practice this change should have little impact, as the performance of
unreachable code is unlikely to matter.
This commit is part of a series that makes Polly more robust in the presence
of unreachables.
llvm-svn: 297151
There is no point in optimizing unreachable code, hence our test cases should
always return.
This commit is part of a series that makes Polly more robust on the presence of
unreachables.
llvm-svn: 297150
There is no point in optimizing unreachable code, hence our test cases should
always return.
This commit is part of a series that makes Polly more robust on the presence of
unreachables.
llvm-svn: 297147
r296992 made ScalarEvolution's CompareValueComplexity less aggressive,
and that broke the polly test being fixed in this change. This change
explicitly bumps CompareValueComplexity in said test case to make it
pass.
Can someone from the polly team please can give me an idea on if this
case is important enough to have
scalar-evolution-max-value-compare-depth be 3 by default?
llvm-svn: 296994
Some Polly ACC test cases fail without a working NVPTX backend. We explicitly
specify this dependence in REQUIRES. Alternatively, we could have only marked
polly-acc as supported in case the NVPTX backend is available, but as we might
use other backends in the future, this does not seem to be the best choice.
For this to work, we also need to make the 'targets_to_build' information
available.
Suggested-by: Michael Kruse <llvm@meinersbur.de>
llvm-svn: 296853
These loads cannot be savely hoisted as the condition guarding the
non-affine region cannot be duplicated to also protect the hoisted load
later on. Today they are dropped in ScopInfo. By checking for this early, we
do not even try to model them and possibly can still optimize smaller regions
not containing this specific required-invariant load.
llvm-svn: 296744
Multi-disjunct access maps can easily result in inbound assumptions which
explode in case of many memory accesses and many parameters. This change reduces
compilation time of some larger kernel from over 15 minutes to less than 16
seconds.
Interesting is the test case test/ScopInfo/multidim_param_in_subscript.ll
which has a memory access
[n] -> { Stmt_for_body3[i0, i1] -> MemRef_A[i0, -1 + n - i1] }
which requires folding, but where only a single disjunct remains. We can still
model this test case even when only using limited memory folding.
For people only reading commit messages, here the comment that explains what
memory folding is:
To recover memory accesses with array size parameters in the subscript
expression we post-process the delinearization results.
We would normally recover from an access A[exp0(i) * N + exp1(i)] into an
array A[][N] the 2D access A[exp0(i)][exp1(i)]. However, another valid
delinearization is A[exp0(i) - 1][exp1(i) + N] which - depending on the
range of exp1(i) - may be preferrable. Specifically, for cases where we
know exp1(i) is negative, we want to choose the latter expression.
As we commonly do not have any information about the range of exp1(i),
we do not choose one of the two options, but instead create a piecewise
access function that adds the (-1, N) offsets as soon as exp1(i) becomes
negative. For a 2D array such an access function is created by applying
the piecewise map:
[i,j] -> [i, j] : j >= 0
[i,j] -> [i-1, j+N] : j < 0
After this patch we generate only the first case, except for situations where
we can proove the first case to be invalid and can consequently select the
second without introducing disjuncts.
llvm-svn: 296679
Without this simplification for a loop nest:
void foo(long n1_a, long n1_b, long n1_c, long n1_d,
long p1_b, long p1_c, long p1_d,
float A_1[][p1_b][p1_c][p1_d]) {
for (long i = 0; i < n1_a; i++)
for (long j = 0; j < n1_b; j++)
for (long k = 0; k < n1_c; k++)
for (long l = 0; l < n1_d; l++)
A_1[i][j][k][l] += i + j + k + l;
}
the assumption:
n1_a <= 0 or (n1_a > 0 and n1_b <= 0) or
(n1_a > 0 and n1_b > 0 and n1_c <= 0) or
(n1_a > 0 and n1_b > 0 and n1_c > 0 and n1_d <= 0) or
(n1_a > 0 and n1_b > 0 and n1_c > 0 and n1_d > 0 and
p1_b >= n1_b and p1_c >= n1_c and p1_d >= n1_d)
is taken rather than the simpler assumption:
p9_b >= n9_b and p9_c >= n9_c and p9_d >= n9_d.
The former is less strict, as it allows arbitrary values of p1_* in case, the
loop is not executed at all. However, in practice these precise constraints
explode when combined across different accesses and loops. For now it seems
to make more sense to take less precise, but more scalable constraints by
default. In case we find a practical example where more precise constraints
are needed, we can think about allowing such precise constraints in specific
situations where they help.
This change speeds up the new test case from taking very long (waited at least
a minute, but it probably takes a lot more) to below a second.
llvm-svn: 296456
This patch adds an option to build against a version of libisl already
installed on the system. The installation is autodetected using the
pkg-config file shipped with isl.
The detection of the library is in the FindISL.cmake module that creates
an imported target.
Contributed-by: Philip Pfaffe <philip.pfaffe@gmail.com>
Differential Revision: https://reviews.llvm.org/D30043
llvm-svn: 296361
These verify that some scalars are not mapped because it would be
incorrect to do so.
For these check we verify that no transformation has been executed from
output of the pass's '-analyze'. Adding optimization remarks is not useful
as it would result in too many messages, even repeated ones. I avoided
checking the '-debug-only=polly-delicm' output which is an antipattern.
llvm-svn: 296348
Currently, pattern based optimizations of Polly can identify matrix
multiplication and optimize it according to BLIS matmul optimization pattern
(see ScheduleTreeOptimizer for details). This patch makes optimizations
based on pattern matching be enabled by default.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D30293
llvm-svn: 295958
These tests were not included in the main DeLICM commit. These check the
cases where zone analysis cannot be successful because of assumption
violations.
We use the LLVM optimization remark infrastructure as it seems to be the
best fit for this kind of messages. I tried to make use if the
OptimizationRemarkEmitter. However, it would insert additional function
passes into the pass manager to get the hotness information. The pass
manager would insert them between the flatten pass and delicm, causing
the ScopInfo with the flattened schedule being thrown away.
Differential Revision: https://reviews.llvm.org/D30253
llvm-svn: 295846
Implement the -polly-delicm pass. The pass intends to undo the
effects of LoopInvariantCodeMotion (LICM) which adds additional scalar
dependencies into SCoPs. DeLICM will try to map those scalars back to
the array elements they were promoted from, as long as the array
element is unused.
The is the main patch from the DeLICM/DePRE patch series. It does not
yet undo GVN PRE for which additional information about known values
is needed and does not handle PHI write accesses that have have no
target. As such its usefulness is limited. Patches for these issues
including regression tests for error situatons will follow.
Reviewers: grosser
Differential Revision: https://reviews.llvm.org/D24716
llvm-svn: 295713
Instead of counting the number of read-only accesses, we now count the number of
distinct read-only array references when checking if a run-time alias check
may be too complex. The run-time alias check is quadratic in the number of
base pointers, not the number of accesses.
Before this change we accidentally skipped SPEC's lbm test case.
llvm-svn: 295567
This test case is a mini performance test case that shows the time needed for a
couple of simple reductions. It takes today about 325ms on my machine to run
this test case through 'opt' with scop construction and reduction detection. It
can be used as mini-proxy for further tuning of the reduction code.
Generally we do not commit performance test cases, but as this is very
small and also very fast it seems OK to keep it in the lit test suite.
This test case will also help to verify that future changes to the reduction
code will not affect the ordering of the reduction sets and will consequently
not cause spurious performance changes that only result from reordering of
dependences in the reduction set.
llvm-svn: 295549
Trying to fold such kind of dimensions will result in a division by zero,
which crashes the compiler. As such arrays are likely to invalidate the
scop anyhow (but are not illegal in LLVM-IR), there is no point in trying
to optimize the array layout. Hence, we just avoid the folding of
constant dimensions of size zero.
llvm-svn: 295415
Before this change wrapping range metadata resulted in exponential growth of
the context, which made context construction of large scops very slow. Instead,
we now just do not model the range information precisely, in case the number
of disjuncts in the context has already reached a certain limit.
llvm-svn: 295360
Commit r230230 introduced the use of range metadata to derive bounds for
parameters, instead of just looking at the type of the parameter. As part of
this commit support for wrapping ranges was added, where the lower bound of a
parameter is larger than the upper bound:
{ 255 < p || p < 0 }
However, at the same time, for wrapping ranges support for adding bounds given
by the size of the containing type has acidentally been dropped. As a result,
the range of the parameters was not guaranteed to be bounded any more. This
change makes sure we always add the bounds given by the size of the type and
then additionally add bounds based on signed wrapping, if available. For a
parameter p with a type size of 32 bit, the valid range is then:
{ -2147483648 <= p <= 2147483647 and (255 < p or p < 0) }
llvm-svn: 295349
When deriving the range of valid values of a scalar evolution expression might
be a range [12, 8), where the upper bound is smaller than the lower bound and
where the range is expected to possibly wrap around. We theoretically could
model such a range as a union of two non-wrapping ranges, but do not do this
as of yet. Instead, we just do not derive any bounds. Before this change,
we could have obtained bounds where the maximal possible value is strictly
smaller than the minimal possible value, which is incorrect and also caused
assertions during scop modeling.
llvm-svn: 294891
To determine parameters of the matrix multiplication, we check RAW dependencies
that can be expressed using only reduction dependencies. Consequently, we
should check the reduction dependencies, if this is the case.
Reviewed-by: Tobias Grosser <tobias@grosser.es>,
Sven Verdoolaege <skimo-polly@kotnet.org>
Michael Kruse <llvm@meinersbur.de>
Differential Revision: https://reviews.llvm.org/D29814
llvm-svn: 294836
The size of the operands type is the one of the parameters required
to determine the BLIS micro-kernel. We get the size of the widest type
of the matrix multiplication operands in case there are several
different types.
Reviewed-by: Michael Kruse <llvm@meinersbur.de>
Differential Revision: https://reviews.llvm.org/D29269
llvm-svn: 294828
Instead of iterating over statements and their memory accesses to extract the
set of available base pointers, just directly iterate over all ScopArray
objects. This reflects more the actual intend of the code: collect all arrays
(and their base pointers) to emit alias information that specifies that accesses
to different arrays cannot alias.
This change removes unnecessary uses of MemoryAddress::getBaseAddr() in
preparation for https://reviews.llvm.org/D28518.
llvm-svn: 294574
There are problems with using the machine information to derive the precise
vector size on polly-amd64-linux and polly-arm-linux. We temporarily disable
the problematic run lines.
llvm-svn: 294571
Before this change we used the name of the base pointer to mark reductions. This
is imprecise as the canonical reference is the ScopArray itself and not the
basepointer of a reduction. Using the base pointer of reductions is problematic
in cases where a single ScopArray is referenced through two different base
pointers.
This change removes unnecessary uses of MemoryAddress::getBaseAddr() in
preparation for https://reviews.llvm.org/D28518.
llvm-svn: 294568
optimization
Isolate a set of partial tile prefixes to allow hoisting and sinking out of
the unrolled innermost loops produced by the optimization of the matrix
multiplication.
In case it cannot be proved that the number of loop iterations can be evenly
divided by tile sizes and we tile and unroll the point loop, the isl generates
conditional expressions. Subsequently, the conditional expressions can prevent
stores and loads of the unrolled loops from being sunk and hoisted.
The patch isolates a set of partial tile prefixes, which have exactly Mr x Nr
iterations of the two innermost loops, the result of the loop tiling performed
by the matrix multiplication optimization, where Mr and Mr are parameters of
the micro-kernel. This helps to get rid of the conditional expressions of
the unrolled innermost loops. Probably this approach can be replaced with
padding in future.
In case of, for example, the gemm from Polybench/C 3.2 and parametric loop
bounds, it helps to increase the performance from 7.98 GFlops (27.71% of
theoretical peak) to 21.47 GFlops (74.57% of theoretical peak). Hence, we
get the same performance as in case of scalar loops bounds.
It also cause compile time regression. The compile-time is increased from
0.795 seconds to 0.837 seconds in case of scalar loops bounds and from 1.222
seconds to 1.490 seconds in case of parametric loops bounds.
Reviewed-by: Michael Kruse <llvm@meinersbur.de>
Differential Revision: https://reviews.llvm.org/D29244
llvm-svn: 294564
with optimizeMatMulPattern
This patch makes ScheduleTreeOptimizer::optimizeBand return a schedule node
optimized with optimizeMatMulPattern. Otherwise, it could not use the isolate
option, because standardBandOpts could try to tile a band node with anchored
subtree and get the error, since the use of the isolate option causes any tree
containing the node to be considered anchored. Furthermore, it is not intended
to apply standard optimizations, when the matrix multiplication has been
detected.
llvm-svn: 294444
multiplication
The current identification of a SCoP statement that implement a matrix
multiplication does not help to identify different permutations of loops that
contain it and check for dependencies, which can prevent it from being
optimized. It also requires external determination of the operands of
the matrix multiplication. This patch contains the implementation of a new
algorithm that helps to avoid these issues. It also modifies the test cases
that generate matrix multiplications with linearized accesses, because
the new algorithm does not support them.
Reviewed-by: Michael Kruse <llvm@meinersbur.de>,
Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D28357
llvm-svn: 293890
Before this change we created an additional reload in the copy of the incoming
block of a PHI node to reload the incoming value, even though the necessary
value has already been made available by the normally generated scalar loads.
In this change, we drop the code that generates this redundant reload and
instead just reuse the scalar value already available.
Besides making the generated code slightly cleaner, this change also makes sure
that scalar loads go through the normal logic, which means they can be remapped
(e.g. to array slots) and corresponding code is generated to load from the
remapped location. Without this change, the original scalar load at the
beginning of the non-affine region would have been remapped, but the redundant
scalar load would continue to load from the old PHI slot location.
It might be possible to further simplify the code in addOperandToPHI,
but this would not only mean to pull out getNewValue, but to also change the
insertion point update logic. As this did not work when trying it the first
time, this change is likely not trivial. To not introduce bugs last minute, we
postpone further simplications to a subsequent commit.
We also document the current behavior a little bit better.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D28892
llvm-svn: 292486
We rename the test case with -metarenamer to make the variable names easier to
read and add additional check lines that verify the code we currently generate
for PHI nodes. This code is interesting as it contains a PHI node in a
non-affine sub-region, where some incoming blocks are within the non-affine
sub-region and others are outside of the non-affine subregion.
As can be seen in the check lines we currently load the PHI-node value twice.
This commit documents this behavior. In a subsequent patch we will try to
improve this.
llvm-svn: 292470
Summary:
Instead of forbidding such access functions completely, we verify that their
base pointer has been hoisted and only assert in case the base pointer was
not hoisted.
I was trying for a little while to get a test case that ensures the assert is
correctly fired in case of invariant load hoisting being disabled, but I could
not find a good way to do so, as llvm-lit immediately aborts if a command
yields a non-zero return value. As we do not generally test our asserts,
not having a test case here seems OK.
This resolves http://llvm.org/PR31494
Suggested-by: Michael Kruse <llvm@meinersbur.de>
Reviewers: efriedma, jdoerfert, Meinersbur, gareevroman, sebpop, zinob, huihuiz, pollydev
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D28798
llvm-svn: 292213
This feature is currently not supported and an explicit assert to prevent the
introduction of such accesses has been added in r282893. This test case allows
to reproduce the assert (and without the assert the miscompile) added in
r282893. It will help when adding such support at some point.
llvm-svn: 292147
If the parameters of the target cache (i.e., cache level sizes, cache level
associativities) are not specified or have wrong values, we use ones for
parameters of the macro-kernel and do not perform data-layout optimizations of
the matrix multiplication. In this patch we specify the default values of the
cache parameters to be able to apply the pattern matching optimizations even in
this case. Since there is no typical values of this parameters, we use the
parameters of Intel Core i7-3820 SandyBridge that also help to attain the
high-performance on IBM POWER System S822 and IBM Power 730 Express server.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D28090
llvm-svn: 290518
Typically processor architectures do not include an L3 cache, which means that
Nc, the parameter of the micro-kernel, is, for all practical purposes,
redundant ([1]). However, its small values can cause the redundant packing of
the same elements of the matrix A, the first operand of the matrix
multiplication. At the same time, big values of the parameter Nc can cause
segmentation faults in case the available stack is exceeded.
This patch adds an option to specify the parameter Nc as a multiple of
the parameter of the micro-kernel Nr.
In case of Intel Core i7-3820 SandyBridge and the following options,
clang -O3 gemm.c -I utilities/ utilities/polybench.c -DPOLYBENCH_TIME
-march=native -mllvm -polly -mllvm -polly-pattern-matching-based-opts=true
-DPOLYBENCH_USE_SCALAR_LB -mllvm -polly-target-cache-level-associativity=8,8
-mllvm -polly-target-cache-level-sizes=32768,262144 -mllvm
-polly-target-latency-vector-fma=8
it helps to improve the performance from 11.303 GFlops/sec (39,247% of
theoretical peak) to 17.896 GFlops/sec (62,14% of theoretical peak).
Refs.:
[1] - http://www.cs.utexas.edu/users/flame/pubs/TOMS-BLIS-Analytical.pdf
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D28019
llvm-svn: 290256
multiplication
Previously we had two-dimensional accesses to store packed operands of
the matrix multiplication for the sake of simplicity of the packed arrays.
However, addition of the third dimension helps to simplify the corresponding
memory access, reduce the execution time of isl operations applied to it, and
consequently reduce the compile-time of Polly. For example, in case of
Intel Core i7-3820 SandyBridge and the following options,
clang -O3 gemm.c -I utilities/ utilities/polybench.c -DPOLYBENCH_TIME
-march=native -mllvm -polly -mllvm -polly-pattern-matching-based-opts=true
-DPOLYBENCH_USE_SCALAR_LB -mllvm -polly-target-cache-level-associativity=8,8
-mllvm -polly-target-cache-level-sizes=32768,262144 -mllvm
-polly-target-latency-vector-fma=7
it helps to reduce the compile-time from about 361.456 seconds to about 0.816
seconds.
Reviewed-by: Michael Kruse <llvm@meinersbur.de>,
Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D27878
llvm-svn: 290251
To prevent copy statements from accessing arrays out of bounds, ranges of their
extension maps are restricted, according to the constraints of domains.
Reviewed-by: Michael Kruse <llvm@meinersbur.de>
Differential Revision: https://reviews.llvm.org/D25655
llvm-svn: 289815
gemm ([1]). In particular, elements of the matrix B, the second operand of
matrix multiplication, are reused between iterations of the innermost loop.
To keep the reused data in cache, only elements of matrix A, the first operand
of matrix multiplication, should be evicted during an iteration of the
innermost loop. To provide such a cache replacement policy, elements of the
matrix A can, in particular, be loaded first and, consequently, be
least-recently-used.
In our case matrices are stored in row-major order instead of column-major
order used in the BLIS implementation ([1]). One of the ways to address it is
to accordingly change the order of the loops of the loop nest. However, it
makes elements of the matrix A to be reused in the innermost loop and,
consequently, requires to load elements of the matrix B first. Since the LLVM
vectorizer always generates loads from the matrix A before loads from the
matrix B and we can not provide it. Consequently, we only change the BLIS micro
kernel and the computation of its parameters instead. In particular, reused
elements of the matrix B are successively multiplied by specific elements of
the matrix A .
Refs.:
[1] - http://www.cs.utexas.edu/users/flame/pubs/TOMS-BLIS-Analytical.pdf
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D25653
llvm-svn: 289806
This allows us to delinearize code such as the one below, where the array
sizes are A[][2 * n] as there are n times two elements in the innermost
dimension. Alternatively, we could try to generate another dimension for the
struct in the innermost dimension, but as the struct has constant size,
recovering this dimension is easy.
struct com {
double Real;
double Img;
};
void foo(long n, struct com A[][n]) {
for (long i = 0; i < 100; i++)
for (long j = 0; j < 1000; j++)
A[i][j].Real += A[i][j].Img;
}
int main() {
struct com A[100][1000];
foo(1000, A);
llvm-svn: 288489
Add an empty DeLICM pass, without any functional parts.
Extracting the boilerplate from the the functional part reduces the size of the
code to review (https://reviews.llvm.org/D24716)
Suggested-by: Tobias Grosser <tobias@grosser.es>
llvm-svn: 288160
We now collect:
Number of total loops
Number of loops in scops
Number of scops
Number of scops with maximal loop depth 1
Number of scops with maximal loop depth 2
Number of scops with maximal loop depth 3
Number of scops with maximal loop depth 4
Number of scops with maximal loop depth 5
Number of scops with maximal loop depth 6 and larger
Number of loops in scops (profitable scops only)
Number of scops (profitable scops only)
Number of scops with maximal loop depth 1 (profitable scops only)
Number of scops with maximal loop depth 2 (profitable scops only)
Number of scops with maximal loop depth 3 (profitable scops only)
Number of scops with maximal loop depth 4 (profitable scops only)
Number of scops with maximal loop depth 5 (profitable scops only)
Number of scops with maximal loop depth 6 and larger (profitable scops only)
These statistics are certainly completely accurate as we might drop scops
when building up their polyhedral representation, but they should give a good
indication of the number of scops we detect.
llvm-svn: 287973
Our original statistics were added before we introduced a more fine-grained
diagnostic system, but the granularity of our statistics has never been
increased accordingly. This change introduces now one statistic counter per
diagnostic to enable us to collect fine-grained statistics about who certain
scops are not detected. In case coarser grained statistics are needed, the
user is expected to combine counters manually.
llvm-svn: 287968
Introduce the new flag -polly-codegen-generate-expressions which forces Polly
to code generate AST expressions instead of using our SCEV based access
expression generation even for cases where the original memory access relation
was not changed and the SCEV based access expression could be code generated
without any issue.
This is an experimental option for better testing the isl ast expression
generation. The default behavior of Polly remains unchanged. We also exclude
a couple of cases for which the AST expression is not yet working.
llvm-svn: 287694
Drop instructions that do not influence the memory impact of a basic block.
They are not needed to reproduce the original bug (verified) and will cause
random test noise if we would decide to only model the instructions that
have visible side-effects.
llvm-svn: 287626
Add two store instructions at the end of basic blocks that are required to
reproduce the original bug to ensure we always process and model these basic
blocks. This makes this test case stable even in case we would decide to bail
out early of basic blocks which do not modify the global state. Also add
additional check lines to verify how we model the basic block.
llvm-svn: 287625
We add CHECK lines to this test case to make it easier to see the difference
between affine and non-affine memory accesses. We also change the test case to
use a parameteric index expression as otherwise our range analysis will
understand that the non-affine memory access can only access input[1],
which makes it difficult to see that the memory access is in-fact modeled as
non-affine access.
llvm-svn: 287623
Do not assume a load to be hoistable/invariant if the pointer is used by
another instruction in the SCoP that might write to memory and that is
always executed.
llvm-svn: 287272
The validity of a branch condition must be verified at the location of the
branch (the branch instruction), not the location of the icmp that is
used in the branch instruction. When verifying at the wrong location, we
may accept an icmp that is defined within a loop which itself dominates, but
does not contain the branch instruction. Such loops cannot be modeled as
we only introduce domain dimensions for surrounding loops. To address this
problem we change the scop detection to evaluate and verify SCEV expressions at
the right location.
This issue has been around since at least r179148 "scop detection: properly
instantiate SCEVs to the place where they are used", where we explicitly
set the scope to the wrong location. Before this commit the scope
was not explicitly set, which probably also resulted in the scope around the
ICmp to be choosen.
This resolves http://llvm.org/PR30989
Reported-by: Eli Friedman <efriedma@codeaurora.org>
llvm-svn: 286769
Assumptions can either be added for a given basic block, in which case the set
describing the assumptions is expected to match the dimensions of its domain.
In case no basic block is provided a parameter-only set is expected to describe
the assumption.
The piecewise expressions that are generated by the SCEVAffinator sometimes
have a zero-dimensional domain (e.g., [p] -> { [] : p <= -129 or p >= 128 }),
which looks similar to a parameter-only domain, but is still a set domain.
This change adds an assert that checks that we always pass parameter domains to
addAssumptions if BB is empty to make mismatches here fail early.
We also change visitTruncExpr to always convert to parameter sets, if BB is
null. This change resolves http://llvm.org/PR30941
Another alternative to this change would have been to inspect all code to make
sure we directly generate in the SCEV affinator parameter sets in case of empty
domains. However, this would likely complicate the code which combines parameter
and non-parameter domains when constructing a statement domain. We might still
consider doing this at some point, but as this likely requires several non-local
changes this should probably be done as a separate refactoring.
Reported-by: Eli Friedman <efriedma@codeaurora.org>
llvm-svn: 286444
Providing the context to the ast generator allows for additional simplifcations
and -- more importantly -- allows to generate loops with only partially bounded
domains, assuming the domains are bounded for all parameter configurations
that are valid as defined by the context.
This change fixes the crash reported in http://llvm.org/PR30956
The original reason why we did not include the context when generating an
AST was that CLooG and later isl used to sometimes transfer some of the
constraints that bound the size of parameters from the context into the
generated AST. This resulted in operations with very large constants, which
sometimes introduced problematic integer overflows. The latest versions of
the isl AST generator are careful to not introduce such constants.
Reported-by: Eli Friedman <efriedma@codeaurora.org>
llvm-svn: 286442
When extracting constant expressions out of SCEVs, new parameters may be
introduced, which have not been registered before. This change scans
SCEV expressions after constant extraction again to make sure newly
introduced parameters are registered.
We may for example extract the constant '8' from the expression '((8 * ((%a *
%b) + %c)) + (-8 * %a))' and obtain the expression '(((-1 + %b) * %a) + %c)'.
The new expression has a new parameter '(-1 + %b) * %a)', which was not
registered before, but must be registered to not crash.
This closes http://llvm.org/PR30953
Reported-by: Eli Friedman <efriedma@codeaurora.org>
llvm-svn: 286430
In r248701 "Allow switch instructions in SCoPs" support for switch statements
has been introduced, but support for switch statements in loop latches was
incomplete. This change completely disables switch statements in loop latches.
The original commit changed addLoopBoundsToHeaderDomain to support non-branch
terminator instructions, but this change was incorrect: it added a check for
BI != null to the if-branch of a condition, but BI was used in the else branch
es well. As a result, when a non-branch terminator instruction is encounted a
nullptr dereference is triggered. Due to missing test coverage, this bug was
overlooked.
r249273 "[FIX] Approximate non-affine loops correctly" added code to disallow
switch statements for non-affine loops, if they appear in either a loop latch
or a loop exit. We adapt this code to now prohibit switch statements in
loop latches even if the control condition is affine.
We could possibly add support for switch statements in loop latches, but such
support should be evaluated and tested separately.
This fixes llvm.org/PR30952
Reported-by: Eli Friedman <efriedma@codeaurora.org>
llvm-svn: 286426
We don't actually check whether a MemoryAccess is affine in very many
places, but one important one is in checks for aliasing.
Differential Revision: https://reviews.llvm.org/D25706
llvm-svn: 285746
When adding an llvm.memcpy instruction to AliasSetTracker, it uses the raw
source and target pointers which preserve bitcasts.
MemAccInst::getPointerOperand() also returns the raw target pointers, but
Scop::buildAliasGroups() did not for the source pointer. This lead to mismatches
between AliasSetTracker and ScopInfo on which pointer to use.
Fixed by also using raw pointers in Scop::buildAliasGroups().
llvm-svn: 285071
Integer math in LLVM IR is modular. Integer math in isl is
arbitrary-precision. Modeling LLVM IR math correctly in isl requires
either adding assumptions that math doesn't actually overflow, or
explicitly wrapping the math. However, expressions with the "nsw" flag
are special; we can pretend they're arbitrary-precision because it's
undefined behavior if the result wraps. SCEV expressions based on IR
instructions with an nsw flag also carry an nsw flag (roughly; actually,
the real rule is a bit more complicated, but the details don't matter
here).
Before this patch, SCEV flags were also overloaded with an additional
function: the ZExt code was mutating SCEV expressions as a hack to
indicate to checkForWrapping that we don't need to add assumptions to
the operand of a ZExt; it'll add explicit wrapping itself. This kind of
works... the problem is that if anything else ever touches that SCEV
expression, it'll get confused by the incorrect flags.
Instead, with this patch, we make the decision about whether to
explicitly wrap the math a bit earlier, basing the decision purely on
the SCEV expression itself, and not its users.
Differential Revision: https://reviews.llvm.org/D25287
llvm-svn: 284848
Update test after commit r284501:
[SCEV] Make CompareValueComplexity a little bit smarter
Contributed-by: Sanjoy Das <sanjoy@playingwithpointers.com>
llvm-svn: 284543
lit recursively iterates through the test subdirectories and finds the ISL
unittest. For this test to work, the polly-isl-test executable needs to be
compiled.
Add the polly-isl-test dependency to POLLY_TEST_DEPS. This makes check-polly and
check-polly-tests work from a fresh build directory.
llvm-svn: 284339
The test non_affine_loop_used_later.ll also tests the profability heuristic. Add
the option -polly-unprofitable-scalar-accs explicitely to ensure that the test
succeeds if the default value is changed.
llvm-svn: 284338
Under some conditions MK_Value read accessed where converted to MK_ExitPHI read
accessed. This is unexpected because MK_ExitPHI read accesses are implicit after
the scop execution. This behaviour was introduced in r265261, which fixed a
failed assertion/crash in CodeGen.
Instead, we fix this failure in CodeGen itself. createExitPHINodeMerges(),
despite its name, also handles accesses of kind MK_Value, only to skip them
because they access values that are usually not PHI nodes in the SCoP region's
exit block. Except in the situation observed in r265261.
Do not convert value accessed to ExitPHI accesses and do not handle
value accesses like ExitPHI accessed in CodeGen anymore.
llvm-svn: 284023
Folders in Visual Studio solutions help organize the build artifacts from all
LLVM projects. There is a folder to keep Polly-built files in.
llvm-svn: 283546
Running isl tests is important to gain confidence that the isl build we created
works as expected. Besides the actual isl tests, there are also isl AST
generation tests shipped with isl. This change only adds support for the isl
unit tests. AST generation test support is left for a later commit.
There is a choice to run tests directly through the build system or in the
context of lit. We choose to run tests as part of lit to as this allows us to
easily set environment variables, print output only on error and generally run
the tests directly from the lit command.
Reviewers: brad.king, Meinersbur
Subscribers: modocache, brad.king, pollydev, beanz, llvm-commits, mgorny
Differential Revision: https://reviews.llvm.org/D25155
llvm-svn: 283245
With this option one can disable the heuristic that assumes that statements with
a scalar write access cannot be profitably optimized. Such a statement instances
necessarily have WAW-dependences to itself. With DeLICM scalar accesses can be
changed to array accesses, which can avoid these WAW-dependence.
llvm-svn: 283233
ScopArrayInfo used to determine base pointer origins by looking up whether the
base pointer is a load. The "base pointer" for scalar accesses is the
llvm::Value being accessed. This is only a symbolic base pointer, it
represents the alloca variable (.s2a or .phiops) generated for it at code
generation.
This patch disables determining base pointer origin for scalars.
A test case where this caused a crash will be added in the next commit. In that
test SAI tried to get the origin base pointer that was only declared later,
therefore not existing. This is probably only possible for scalars used in
PHINode incoming blocks.
llvm-svn: 283232
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
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
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
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
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
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
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
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
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
Before this commit we generated the array type in reverse order and we also
added the outermost dimension size to the new array declaration, which is
incorrect as Polly additionally assumed an additional unsized outermost
dimension, such that we had an off-by-one error in the linearization of access
expressions.
llvm-svn: 277802
These annotations ensure that the NVIDIA PTX assembler limits the number of
registers used such that we can be certain the resulting kernel can be executed
for the number of threads in a thread block that we are planning to use.
llvm-svn: 277799
Pass the content of scalar array references to the alloca on the kernel side
and do not pass them additional as normal LLVM scalar value.
llvm-svn: 277699
Otherwise, we would try to re-optimize them with Polly-ACC and possibly even
generate kernels that try to offload themselves, which does not work as the
GPURuntime is not available on the accelerator and also does not make any
sense.
llvm-svn: 277589
Extend the jscop interface to allow the user to export arrays. It is required
that already existing arrays of the list of arrays correspond to arrays
of the SCoP. Each array that is appended to the list will be newly created.
Furthermore, we allow the user to modify access expressions to reference
any array in case it has the same element type.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D22828
llvm-svn: 277263
Before this change we used the array index, which would result in us accessing
the parameter array out-of-bounds. This bug was visible for test cases where not
all arrays in a scop are passed to a given kernel.
llvm-svn: 276961
Adding a new pass PolyhedralInfo. This pass will be the interface to Polly.
Initially, we will provide the following interface:
- #IsParallel(Loop *L) - return a bool depending on whether the loop is
parallel or not for the given program order.
Patch by Utpal Bora <cs14mtech11017@iith.ac.in>
Differential Revision: https://reviews.llvm.org/D21486
llvm-svn: 276637
Also factor out getArraySize() to avoid code dupliciation and reorder some
function arguments to indicate the direction into which data is transferred.
llvm-svn: 276636
At the beginning of each SCoP, we allocate device arrays for all arrays
used on the GPU and we free such arrays after the SCoP has been executed.
llvm-svn: 276635
Do not process SCoPs with infeasible runtime context in the new
ScopInfoWrapperPass. Do not compute dependences for such SCoPs in the new
DependenceInfoWrapperPass.
Patch by Utpal Bora <cs14mtech11017@iith.ac.in>
Differential Revision: https://reviews.llvm.org/D22402
llvm-svn: 276631
This is the second 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 create the BLIS macro-kernel by applying a combination of tiling
and interchanging. In subsequent changes we will implement the packing
transformation.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: http://reviews.llvm.org/D21491
llvm-svn: 276627
There is no need to expose the selected device at the moment. We also pass back
pointers as return values, as this simplifies the interface.
llvm-svn: 276623
Run the NVPTX backend over the GPUModule IR and write the resulting assembly
code in a string.
To work correctly, it is important to invalidate analysis results that still
reference the IR in the kernel module. Hence, this change clears all references
to dominators, loop info, and scalar evolution.
Finally, the NVPTX backend has troubles to generate code for various special
floating point types (not surprising), but also for uncommon integer types. This
commit does not resolve these issues, but pulls out problematic test cases into
separate files to XFAIL them individually and resolve them in future (not
immediate) changes one by one.
llvm-svn: 276396
This change introduces the actual compute code in the GPU kernels. To ensure
all values referenced from the statements in the GPU kernel are indeed available
we scan all ScopStmts in the GPU kernel for references to llvm::Values that
are not yet covered by already modeled outer loop iterators, parameters, or
array base pointers and also pass these additional llvm::Values to the
GPU kernel.
For arrays used in the GPU kernel we introduce a new ScopArrayInfo object, which
is referenced by the newly generated access functions within the GPU kernel and
which is used to help with code generation.
llvm-svn: 276270
This ensures that no trivially dead code is generated. This is not only cleaner,
but also avoids troubles in case code is generated in a separate function and
some of this dead code contains references to values that are not available.
This issue may happen, in case the memory access functions have been updated
and old getelementptr instructions remain in the code. With normal Polly,
a test case is difficult to draft, but the upcoming GPU code generation can
possibly trigger such problems. We will later extend this dead-code elimination
to region and vector statements.
llvm-svn: 276263
It seems the order in which we generated memory accesses changed such that
the import of these updated memory accesses failed for the 'loop3' statement
in this test case. Unfortunately, the existing CHECK lines were not strict
enough to catch this. Hence, besides fixing the order of the memory access
lines we also ensure that the memory access changes are both clearly visibly
and well checked.
llvm-svn: 276247
This is currently not supported and will only be added later. Also update the
test cases to ensure no invariant code hoisting is applied.
llvm-svn: 275987
We use this opportunity to further classify the different user statements that
can arise and add TODOs for the ones not yet implemented.
llvm-svn: 275957
Create for each kernel a separate LLVM-IR module containing a single function
marked as kernel function and taking one pointer for each array referenced
by this kernel. Add debugging output to verify the kernels are generated
correctly.
llvm-svn: 275952
Initialize the list of references to a GPU array to ensure that the arrays that
need to be passed to kernel calls are computed correctly. Furthermore, the very
same information is also necessary to compute synchronization correctly. As the
functionality to compute these references is already available, what is left for
us to do is only to connect the necessary functionality to compute array
reference information.
llvm-svn: 275798
Create LLVM-IR for all host-side control flow of a given GPU AST. We implement
this by introducing a new GPUNodeBuilder class derived from IslNodeBuilder. The
IslNodeBuilder will take care of generating all general-purpose ast nodes, but
we provide our own createUser implementation to handle the different GPU
specific user statements. For now, we just skip any user statement and only
generate a host-code sceleton, but in subsequent commits we will add handling of
normal ScopStmt's performing computations, kernel calls, as well as host-device
data transfers. We will also introduce run-time check generation and LICM in
subsequent commits.
llvm-svn: 275783
Otherwise ppcg would try to call into pet functionality that this not available,
which obviously will cause trouble. As we can easily print these statements
ourselves, we just do so.
llvm-svn: 275579
This option increases the scalability of the scheduler and allows us to remove
the 'gisting' workaround we introduced in r275565 to handle a more complicated
test case. Another benefit of using this option is also that the generated
code looks a lot more streamlined.
Thanks to Sven Verdoolaege for reminding me of this option.
llvm-svn: 275573
This works around a shortcoming of the isl scheduler, which even for some
smaller test cases does not terminate in case domain constraints are part
of the flow dependences.
llvm-svn: 275565
Arrays with integer base type are similar to arrays with floating point types,
with the exception that LLVM's integer types can take some odd values. We
add a selection of different values to make sure we correctly round these
types when necessary.
References to scalar integer types are special, as we currently do not model
these types as array accesses as they are considered 'synthesizable' by Polly.
As a result, we do not generate explicit data-transfers for them, but instead
will need to keep track of all references to 'synthesizable' values separately.
At the current stage, this is only visible by missing host-to-device
data-transfer calls. In the future, we will also require special code generation
strategies.
llvm-svn: 275551
We currently only test that the code structure we generate for these scalar
parameters is correct and we add these types to make sure later code generation
additions have sufficient test coverage.
In case some of these types cannot be mapped due to missing hardware support
on the GPU some of these test cases may need to be updated later on.
llvm-svn: 275548
A sequence of CHECK lines allows additional statements to appear in the
output of the tested program without any test failures appearing. As we do
not want this to happen, switch this test case to use CHECK-NEXT.
llvm-svn: 275534
For this we need to provide an explicit list of statements as they occur in
the polly::Scop to ppcg.
We also setup basic AST printing facilities to facilitate debugging. To allow
code reuse some (minor) changes in ppcg are have been necessary.
llvm-svn: 275436
The tile size was previously uninitialized. As a result, it was often zero (aka.
no tiling), which is not what we want in general. More importantly, there was
the risk for arbitrary tile sizes to be choosen, which we did not observe, but
which still is highly problematic.
llvm-svn: 275418
This change now applies ppcg's GPU mapping on our initial schedule. For this
to work, we need to also initialize the set of all names (isl_ids) used in
the scop as well as the program context.
llvm-svn: 275396
To do so we copy the necessary information to compute an initial schedule from
polly::Scop to ppcg's scop. Most of the necessary information is directly
available and only needs to be passed on to ppcg, with the exception of 'tagged'
access relations, access relations that additionally carry information about
which memory access an access relation originates from.
We could possibly perform the construction of tagged accesses as part of
ScopInfo, but as this format is currently specific to ppcg we do not do this
yet, but keep this functionality local to our GPU code generation.
After the scop has been initialized, we compute data dependences and ask ppcg to
compute an initial schedule. Some of this functionality is already available in
polly::DependenceInfo and polly::ScheduleOptimizer, but to keep differences
to ppcg small we use ppcg's functionality here. We may later investiage if
a closer integration of these tools makes sense.
llvm-svn: 275390
At this stage, we do not yet modify the IR but just generate a default
initialized ppcg_scop and gpu_prog and free both immediately. Both will later be
filled with data from the polly::Scop and are needed to use PPCG for GPU
schedule generation. This commit does not yet perform any GPU code generation,
but ensures that the basic infrastructure has been put in place.
We also add a simple test case to ensure the new code is run and use this
opportunity to verify that GPU_CODEGEN tests are only run if GPU code generation
has been enabled in cmake.
llvm-svn: 275389
Check not only that the compiler is not crashing, but also whether the
probablematic part (The sequence of instructions simplified to '4') is reflected
in the output.
Thanks to Tobias for the hint.
llvm-svn: 275189
An assertion in visitSDivInstruction() checked whether the divisor is constant
by checking whether the argument is a ConstantInt. However, SCEVValidator allows
the divisor to be simplified to a constant by ScalarEvolution.
We synchronize the implementation of SCEVValidator and SCEVAffinator to both
accept simplified SCEV expressions.
llvm-svn: 275174
For llvm the memory accesses from nonaffine loops should be visible,
however for polly those nonaffine loops should be invisible/boxed.
This fixes llvm.org/PR28245
Cointributed-by: Huihui Zhang <huihuiz@codeaurora.org>
Differential Revision: http://reviews.llvm.org/D21591
llvm-svn: 274842
Reject and report regions that contains loops overlapping nonaffine region.
This situation typically happens in the presence of inifinite loops.
This addresses bug llvm.org/PR28071.
Differential Revision: http://reviews.llvm.org/D21312
Contributed-by: Huihui Zhang <huihuiz@codeaurora.org>
llvm-svn: 273905
This patch addresses:
- A new function pass to compute polyhedral dependences. This is
required to avoid the region pass manager.
- Stores a map of Scop to Dependence object for all the scops present
in a function. By default, access wise dependences are stored.
Patch by Utpal Bora <cs14mtech11017@iith.ac.in>
Differential Revision: http://reviews.llvm.org/D21105
llvm-svn: 273881
This patch adds a new function pass ScopInfoWrapperPass so that the
polyhedral description of a region, the SCoP, can be constructed and
used in a function pass.
Patch by Utpal Bora <cs14mtech11017@iith.ac.in>
Differential Revision: http://reviews.llvm.org/D20962
llvm-svn: 273856
This is the first 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 create the BLIS micro-kernel by applying
a combination of tiling and unrolling. In subsequent changes
we will add the extraction of the BLIS macro-kernel
and implement the packing transformation.
Contributed-by: Roman Gareev <gareevroman@gmail.com>
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: http://reviews.llvm.org/D21140
llvm-svn: 273397
With this update the isl AST generation extracts disjunctive constraints early
on. As a result, code that previously resulted in two branches with (close-to)
identical code within them:
if (P <= -1) {
for (int c0 = 0; c0 < N; c0 += 1)
Stmt_store(c0);
} else if (P >= 1)
for (int c0 = 0; c0 < N; c0 += 1)
Stmt_store(c0);
results now in only a single branch body:
if (P <= -1 || P >= 1)
for (int c0 = 0; c0 < N; c0 += 1)
Stmt_store(c0);
This resolves http://llvm.org/PR27559
Besides the above change, this isl update brings better simplification of
sets/maps containing existentially quantified dimensions and fixes a bug in
isl's coalescing.
llvm-svn: 272500
As these test cases will be changed in a subsequent commit, we expand and
tighten them to make the subsequent changes to them more obvious. As part of
this we add more context to some test cases and add CHECK-NEXT lines to ensure
no intermediate lines are missed by accident.
llvm-svn: 272499
IntToPtr and PtrToInt instructions are basically no-ops that we can handle as
such. In order to generate them properly as parameters we had to improve the
ScopExpander, though the change is the first in the direction of a more
aggressive scalar synthetization.
This patch was originally contributed by Johannes Doerfert in r271888, but was
in conflict with the revert in r272483. This is a recommit with some minor
adjustment to the test cases to take care of differing instruction names.
llvm-svn: 272485
The recent expression type changes still need more discussion, which will happen
on phabricator or on the mailing list. The precise list of commits reverted are:
- "Refactor division generation code"
- "[NFC] Generate runtime checks after the SCoP"
- "[FIX] Determine insertion point during SCEV expansion"
- "Look through IntToPtr & PtrToInt instructions"
- "Use minimal types for generated expressions"
- "Temporarily promote values to i64 again"
- "[NFC] Avoid unnecessary comparison for min/max expressions"
- "[Polly] Fix -Wunused-variable warnings (NFC)"
- "[NFC] Simplify min/max expression generation"
- "Simplify the type adjustment in the IslExprBuilder"
Some of them are just reverted as we would otherwise get conflicts. I will try
to re-commit them if possible.
llvm-svn: 272483
This patch refactors the code generation for divisions. This allows to
always generate a shift for a power-of-two division and to utilize
information about constant divisors in order to truncate the result
type.
llvm-svn: 271898
We now generate runtime checks __after__ the SCoP code generation and
not before, though they are still inserted at the same position int
the code. This allows to modify the runtime check during SCoP code
generation.
llvm-svn: 271894
IntToPtr and PtrToInt instructions are basically no-ops that we can handle as
such. In order to generate them properly as parameters we had to improve the
ScopExpander, though the change is the first in the direction of a more
aggressive scalar synthetization.
llvm-svn: 271888
We now use the minimal necessary bit width for the generated code. If
operations might overflow (add/sub/mul) we will try to adjust the types in
order to ensure a non-wrapping computation. If the type adjustment is not
possible, thus the necessary type is bigger than the type value of
--polly-max-expr-bit-width, we will use assumptions to verify the computation
will not wrap. However, for run-time checks we cannot build assumptions but
instead utilize overflow tracking intrinsics.
llvm-svn: 271878
In case of modulo compared to zero, we need to do signed modulo
operation as unsigned can give different results based on whether the
dividend is negative or not.
This addresses llvm.org/PR27707
Contributed-by: Chris Jenneisch <chrisj@codeaurora.org>
Reviewers: _jdoerfert, grosser, Meinersbur
Differential Revision: http://reviews.llvm.org/D20145
llvm-svn: 271707
Operands of binary operations that might overflow will be temporarily
promoted to i64 again, though that is not a sound solution for the problem.
llvm-svn: 271538
Summary:
After rL271151 (SCEV change) SCEV no longer unconditionally transfers
nuw/nsw from the increment operation to the post-inc value; this
transfer only happens if there is undefined behavior in the program if
the increment overflowed (as opposed to just generating poison).
The loops in `wraping_signed_expr_1.ll` are in non-canonical
form (they're not rotated), and that defeats LLVM's poison-is-UB
analysis. IMO the easiest fix here is to run `wraping_signed_expr_1.ll`
through `-loop-rotate` to canonicalize the loops, which is what this
patch does.
Reviewers: jdoerfert, Meinersbur, grosser
Subscribers: grosser, mcrosier, pollydev
Differential Revision: http://reviews.llvm.org/D20778
llvm-svn: 271536
We now have a simple function to adjust/unify the types of two (or three)
operands before an operation that requieres the same type for all operands.
Due to this change we will not promote parameters that are added to i64
anymore if that is not needed.
llvm-svn: 271513
multiplication
Fix small issues related to characters, operators and descriptions of tests.
Differential Revision: http://reviews.llvm.org/D20806
llvm-svn: 271264
Add determination of statements that contain, in particular,
matrix multiplications and can be optimized with [1] to try to
get close-to-peak performance. It can be enabled
via polly-pm-based-opts, which is false by default.
Refs:
[1] - http://www.cs.utexas.edu/users/flame/pubs/TOMS-BLIS-Analytical.pdf
Contributed-by: Roman Gareev <gareevroman@gmail.com>
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: http://reviews.llvm.org/D20575
llvm-svn: 271128
Before this patch we bailed if a required invariant load was potentially
overwritten. However, now we will optimistically assume it is actually
invariant and, to this end, restrict the valid parameter space as well as the
execution context with regards to potential overwrites of the location.
llvm-svn: 270416
Since the base pointer of a possibly aliasing pointer might not alias
with any other pointer it (the base pointer) might not be tagged as
"required invariant". However, we need it do be in order to compare
the accessed addresses of the derived (possibly aliasing) pointer.
This patch also tries to clean up the load hoisting a little bit.
llvm-svn: 270412
So far we bailed if a required invariant load was potentially overwritten in
the SCoP. From now on we will optimistically assume it is actually invariant
and, to this end, restrict the valid parameter space.
llvm-svn: 270060
This patch cleans up the rejection log handling during the
ScopDetection. It consists of two interconnected parts:
- We keep all detection contexts for a function in order to provide
more information to the user, e.g., about the rejection of
extended/intermediate regions.
- We remove the mutable "RejectLogs" member as the information is
available through the detection contexts.
llvm-svn: 269323
Truncate operations are basically modulo operations, thus we can model
them that way. However, for large types we assume the operand to fit
in the new type size instead of introducing a modulo with a very large
constant.
llvm-svn: 269300
We utilize assumptions on the input to model IR in polyhedral world.
To verify these assumptions we version the code and guard it with a
runtime-check (RTC). However, since the RTCs are themselves generated
from the polyhedral representation we generate them under the same
assumptions that they should verify. In other words, the guarantees
that we try to provide with the RTCs do not hold for the RTCs
themselves. To this end it is necessary to employ a different check
for the RTCs that will verify the assumptions did hold for them too.
Differential Revision: http://reviews.llvm.org/D20165
llvm-svn: 269299
If a profitable run is performed we will check if the SCoP seems to be
profitable after creation but before e.g., dependence are computed. This is
needed as SCoP detection only approximates the actual SCoP representation.
In the end this should allow us to be less conservative during the SCoP
detection while keeping the compile time in check.
llvm-svn: 269074
Regions with one affine loop can be profitable if the loop is
distributable. To this end we will allow them to be treated as
profitable if they contain at least two non-trivial basic blocks.
llvm-svn: 269064
The assumption attached to an llvm.assume in the SCoP needs to be
combined with the domain of the surrounding statement but can
nevertheless be used to refine the context.
This fixes the problems mentioned in PR27067.
llvm-svn: 269060
This patches makes the propagation of complexity problems during
domain generation consistent. Additionally, it makes it less likely to
encounter ill-formed domains later, e.g., during schedule generation.
llvm-svn: 269055
Before this patch we generated error-restrictions only for
error-blocks, thus blocks (or regions) containing a not represented
function call. However, the same reasoning is needed if the invalid
domain of a statement subsumes its actual domain. To this end we move
the generation of error-restrictions after the propagation of the
invalid domains. Consequently, error-statements are now defined more
general as statements that are assumed to be not executed.
Additionally, we do not record an empty domain for such statements but
a nullptr instead. This allows to distinguish between error-statements
and dead-statements.
llvm-svn: 269053
We now use context information to simplify the domains and access
functions of the SCoP instead of just aligning them with the parameter
space.
llvm-svn: 269048
Min/max expressions are easier to read and can in some cases also result in
more concise IR that is generated as the min/max --- when lowered to a
cmp+select pattern -- commonly has a simpler condition then the ternary
condition isl would normally generate.
llvm-svn: 268855
This release includes sevaral improvments compared to the previous
version isl-0.16.1-145-g243bf7c (from the ISL 0.17 announcement):
- optionally combine SCCs incrementally in scheduler
- optionally maximize coincidence in scheduler
- optionally avoid loop coalescing in scheduler
- minor AST generator improvements
- improve support for expansions in schedule trees
llvm-svn: 268500
Add a command line switch to set the
isl_options_set_schedule_outer_coincidence option. ISL then tries to
build schedules where the outer member of a band satisfies the
coincidence constraints.
In practice this allows loop skewing for more parallelism in inner
loops.
llvm-svn: 268222
After zero-extend operations and unsigned comparisons we now allow
unsigned divisions. The handling is basically the same as for signed
division, except the interpretation of the operands. As the divisor
has to be constant in both cases we can simply interpret it as an
unsigned value without additional complexity in the representation.
For the dividend we could choose from the different representation
schemes introduced for zero-extend operations but for now we will
simply use an assumption.
llvm-svn: 268032
It does not suffice to take a global assumptions for unsigned comparisons but
we also need to adjust the invalid domain of the statements guarded by such
an assumption. To this end we allow to specialize the getPwAff call now in
order to indicate unsigned interpretation.
llvm-svn: 268025
Assumptions and restrictions can both be simplified with the domain of a
statement but not the same way. After this patch we will correctly
distinguish them.
llvm-svn: 267885
Instead of matching for %6, we use a regexp to match for the result strings.
This test case caused unrelated noise in http://reviews.llvm.org/D15722.
llvm-svn: 267875
If the base pointer of an invariant load is is loaded conditionally, that
condition needs to hold for the invariant load too. The structure of the
program will imply this for domain constraints but not for imprecisions in
the modeling. To this end we will propagate the execution context of base
pointers during code generation and thus ensure the derived pointer does
not access an invalid base pointer.
llvm-svn: 267707
With this patch we will optimistically assume that the result of an unsigned
comparison is the same as the result of the same comparison interpreted as
signed.
llvm-svn: 267559
Additive expressions can have constant factors too that we can extract
and thereby simplify the internal representation. For now we do
compute the gcd of all constant factors but only extract the same
(possibly negated) factor if there is one.
llvm-svn: 267445
Before, we checked all GEPs in a statement in order to derive
out-of-bound assumptions. However, this can not only introduce new
parameters but it is also not clear what we can learn from GEPs that
are not immediately used in a memory accesses inside the SCoP. As this
case is very rare, no actual change in the behaviour is expected.
llvm-svn: 267442
Before, assumptions derived from llvm.assume could reference new
parameters that were not known to the SCoP before. These were neither
beneficial to the representation nor to the user that reads the
emitted remark. Now we project them out and keep only user assumptions
on known parameters. Nevertheless, the new parameters are still part
of the SCoPs parameter space as the SCEVAffinator currently adds them
on demand.
llvm-svn: 267441
A zero-extended value can be interpreted as a piecewise defined signed
value. If the value was non-negative it stays the same, otherwise it
is the sum of the original value and 2^n where n is the bit-width of
the original (or operand) type. Examples:
zext i8 127 to i32 -> { [127] }
zext i8 -1 to i32 -> { [256 + (-1)] } = { [255] }
zext i8 %v to i32 -> [v] -> { [v] | v >= 0; [256 + v] | v < 0 }
However, LLVM/Scalar Evolution uses zero-extend (potentially lead by a
truncate) to represent some forms of modulo computation. The left-hand side
of the condition in the code below would result in the SCEV
"zext i1 <false, +, true>for.body" which is just another description
of the C expression "i & 1 != 0" or, equivalently, "i % 2 != 0".
for (i = 0; i < N; i++)
if (i & 1 != 0 /* == i % 2 */)
/* do something */
If we do not make the modulo explicit but only use the mechanism described
above we will get the very restrictive assumption "N < 3", because for all
values of N >= 3 the SCEVAddRecExpr operand of the zero-extend would wrap.
Alternatively, we can make the modulo in the operand explicit in the
resulting piecewise function and thereby avoid the assumption on N. For the
example this would result in the following piecewise affine function:
{ [i0] -> [(1)] : 2*floor((-1 + i0)/2) = -1 + i0;
[i0] -> [(0)] : 2*floor((i0)/2) = i0 }
To this end we can first determine if the (immediate) operand of the
zero-extend can wrap and, in case it might, we will use explicit modulo
semantic to compute the result instead of emitting non-wrapping assumptions.
Note that operands with large bit-widths are less likely to be negative
because it would result in a very large access offset or loop bound after the
zero-extend. To this end one can optimistically assume the operand to be
positive and avoid the piecewise definition if the bit-width is bigger than
some threshold (here MaxZextSmallBitWidth).
We choose to go with a hybrid solution of all modeling techniques described
above. For small bit-widths (up to MaxZextSmallBitWidth) we will model the
wrapping explicitly and use a piecewise defined function. However, if the
bit-width is bigger than MaxZextSmallBitWidth we will employ overflow
assumptions and assume the "former negative" piece will not exist.
llvm-svn: 267408
Memory accesses can have non-precisely modeled access functions that
would cause us to build incorrect execution context for hoisted loads.
This is the same issue that occurred during the domain construction for
statements and it is dealt with the same way.
llvm-svn: 267289
The SCEVAffinator will now produce not only the isl representaiton of
a SCEV but also the domain under which it is invalid. This is used to
record possible overflows that can happen in the statement domains in
the statements invalid domain. The result is that invalid loads have
an accurate execution contexts with regards to the validity of their
statements domain. While the SCEVAffinator currently is only taking
"no-wrapping" assumptions, we can add more withouth worrying about the
execution context of loads that are optimistically hoisted.
llvm-svn: 267288
As discussed in the Polly weekly phone call and reviews.llvm.org/D18878,
the assumed contexts changed (widen) due to D18878/r265942. Also check
these contexts in the tests affected by that change.
llvm-svn: 266323
We used checks to minimize the number of remarks we present to a user
but these checks can become expensive, especially since all wrapping
assumptions are emitted separately. Because there is not benefit for a
"headless" run we put these checks under a command line flag. Thus, if
the flag is not given we will emit "non-effective" remarks, e.g.,
duplicates and revert to the old behaviour if it is given. As this
also changes the internal representation of some sets we set the flag
by default for our unit tests.
llvm-svn: 266087
Utilizing the record option for assumptions we can simplify the wrapping
assumption generation a lot. Additionally, we can now report locations
together with wrapping assumptions, though they might not be accurate yet.
llvm-svn: 266069
There are three reasons why we want to record assumptions first before we
add them to the assumed/invalid context:
1) If the SCoP is not profitable or otherwise invalid without the
assumed/invalid context we do not have to compute it.
2) Information about the context are gathered rather late in the SCoP
construction (basically after we know all parameters), thus the user
might see overly complicated assumptions to be taken while they would
have been simplified later on.
3) Currently we cannot take assumptions at any point but have to wait,
e.g., for the domain generation to finish. This makes wrapping
assumptions much more complicated as they need to be and it will
have a similar effect on "signed-unsigned" assumptions later.
llvm-svn: 266068
Collect the error domain contexts (formerly in the ErrorDomainCtxMap)
for each statement in the new InvalidContext member variable. While
this commit is basically a [NFC] it is a first step to make hoisting
sound by allowing a more fine grained record of invalid contexts,
e.g., here on statement level.
llvm-svn: 266053
Allow overflow of indices into the next higher dimension if it has
constant size. E.g.
float A[32][2];
((float*)A)[5];
is effectively the same as
A[2][1];
This can happen since r265379 as a side effect if ScopDetection
recognizes an access as affine, but ScopInfo rejects the GetElementPtr.
Differential Revision: http://reviews.llvm.org/D18878
llvm-svn: 265942
In r247147 we disabled pointer expressions because the IslExprBuilder did not
fully support them. This patch reintroduces them by simply treating them as
integers. The only special handling for pointers that is left detects the
comparison of two address_of operands and uses an unsigned compare.
llvm-svn: 265894
This reverts commit 2879c53e80e05497f408f21ce470d122e9f90f94.
Additionally, it adds SDiv and SRem instructions to the set of values
discovered by the findValues function even if we add the operands to
be able to recompute the SCEVs. In subfunctions we do not want to
recompute SDiv and SRem instructions but pass them instead as they
might have been created through the IslExprBuilder and are more
complicated than simple SDiv/SRem instructions in the code.
llvm-svn: 265873
The way to get the elements size with getPrimitiveSizeInBits() is not
the same as used in other parts of Polly which should use
DataLayout::getTypeAllocSize(). Its use only queries the size of the
pointer and getPrimitiveSizeInBits returns 0 for types that require a
DataLayout object such as pointers.
Together with r265379, this should fix PR27195.
llvm-svn: 265795
If we build the domains for error blocks and later remove them we lose
the information that they are not executed. Thus, in the SCoP it looks
like the control will always reach the statement S:
for (i = 0 ... N)
if (*valid == 0)
doSth(&ptr);
S: A[i] = *ptr;
Consequently, we would have assumed "ptr" to be always accessed and
preloaded it unconditionally. However, only if "*valid != 0" we would
execute the optimized version of the SCoP. Nevertheless, we would have
hoisted and accessed "ptr"regardless of "*valid". This changes the
semantic of the program as the value of "*valid" can cause a change of
"ptr" and control if it is executed or not.
To fix this problem we adjust the execution context of hoisted loads
wrt. error domains. To this end we introduce an ErrorDomainCtxMap that
maps each basic block to the error context under which it might be
executed. Thus, to the context under which it is executed but an error
block would have been executed to. To fill this map one traversal of
the blocks in the SCoP suffices. During this traversal we do also
"remove" error statements and those that are only reachable via error
statements. This was previously done by the removeErrorBlockDomains
function which is therefor not needed anymore.
This fixes bug PR26683 and thereby several SPEC miscompiles.
Differential Revision: http://reviews.llvm.org/D18822
llvm-svn: 265778
If ScalarEvolution cannot look through some expression but we do, it
might happen that a multiplication will arrive at the
SCEVAffinator::visitMulExpr. While we could always try to improve the
extractConstantFactor function we might still miss something, thus we
reintroduce the code to generate multiplicative piecewise-affine
functions as a fall-back.
llvm-svn: 265777
The findValues() function did not look through div & srem instructions
that were part of the argument SCEV. However, in different other
places we already look through it. This mismatch caused us to preload
values in the wrong order.
llvm-svn: 265775
If all exiting blocks of a SCoP are error blocks and therefor not
represented we will not generate accesses and consequently no SAI
objects for exit PHIs. However, they are needed in the code generation
to generate the merge PHIs between the original and optimized region.
With this patch we enusre that the SAI objects for exit PHIs exist
even if all exiting blocks turn out to be eror blocks.
This fixes the crash reported in PR27207.
llvm-svn: 265393
Even before we build the domain the branch condition can become very
complex, especially if we have to build the complement of a lot of
equality constraints. With this patch we bail if the branch condition
has a lot of basic sets and parameters.
After this patch we now successfully compile
External/SPEC/CINT2000/186_crafty/186_crafty
with "-polly-process-unprofitable -polly-position=before-vectorizer".
llvm-svn: 265286
As a CFG is often structured we can simplify the steps performed during
domain generation. When we push domain information we can utilize the
information from a block A to build the domain of a block B, if A dominates B
and there is no loop backede on a path from A to B. When we pull domain
information we can use information from a block A to build the domain of a
block B if B post-dominates A. This patch implements both ideas and thereby
simplifies domains that were not simplified by isl. For the FINAL basic block
in test/ScopInfo/complex-successor-structure-3.ll we used to build a universe
set with 81 basic sets. Now it actually is represented as universe set.
While the initial idea to utilize the graph structure depended on the
dominator and post-dominator tree we can use the available region
information as a coarse grained replacement. To this end we push the
region entry domain to the region exit and pull it from the region
entry for the region exit if applicable.
With this patch we now successfully compile
External/SPEC/CINT2006/400_perlbench/400_perlbench
and
SingleSource/Benchmarks/Adobe-C++/loop_unroll.
Differential Revision: http://reviews.llvm.org/D18450
llvm-svn: 265285
If a loop has no exiting blocks the region covering we use during
schedule genertion might not cover that loop properly. For now we bail
out as we would not optimize these loops anyway.
llvm-svn: 265280
If an exit PHI is written and also read in the SCoP we should not create two
SAI objects but only one. As the read is only modeled to ensure OpenMP code
generation knows about it we can simply use the EXIT_PHI MemoryKind for both
accesses.
llvm-svn: 265261
If a loop has no exiting blocks the region covering we use during
schedule genertion might not cover that loop properly. For now we bail
out as we would not optimize these loops anyway.
llvm-svn: 265260
If a non-affine region PHI is generated we should not move the insert
point prior to the synthezised value in the same block as we might
split that block at the insert point later on. Only if the incoming
value should be placed in a different block we should change the
insertion point.
llvm-svn: 265132
These caused LNT failures due to new assertions when running with
-polly-position=before-vectorizer -polly-process-unprofitable for:
FAIL: clamscan.compile_time
FAIL: cjpeg.compile_time
FAIL: consumer-jpeg.compile_time
FAIL: shapes.compile_time
FAIL: clamscan.execution_time
FAIL: cjpeg.execution_time
FAIL: consumer-jpeg.execution_time
FAIL: shapes.execution_time
The failures have been introduced by r264782, but r264789 had to be reverted
as it depended on the earlier patch.
llvm-svn: 264885
As a CFG is often structured we can simplify the steps performed
during domain generation. When we push domain information we can
utilize the information from a block A to build the domain of a
block B, if A dominates B. When we pull domain information we can
use information from a block A to build the domain of a block B
if B post-dominates A. This patch implements both ideas and thereby
simplifies domains that were not simplified by isl. For the FINAL
basic block in
test/ScopInfo/complex-successor-structure-3.ll .
we used to build a universe set with 81 basic sets. Now it actually is
represented as universe set.
While the initial idea to utilize the graph structure depended on the
dominator and post-dominator tree we can use the available region
information as a coarse grained replacement. To this end we push the
region entry domain to the region exit and pull it from the region
entry for the region exit.
Differential Revision: http://reviews.llvm.org/D18450
llvm-svn: 264789
This patch applies the restrictions on the number of domain conjuncts
also to the domain parts of piecewise affine expressions we generate.
To this end the wording is change slightly. It was needed to support
complex additions featuring zext-instructions but it also fixes PR27045.
lnt profitable runs reports only little changes that might be noise:
Compile Time:
Polybench/[...]/2mm +4.34%
SingleSource/[...]/stepanov_container -2.43%
Execution Time:
External/[...]/186_crafty -2.32%
External/[...]/188_ammp -1.89%
External/[...]/473_astar -1.87%
llvm-svn: 264514
This fixes PR27035. While we now exclude MemIntrinsics from the
polyhedral model if they would access "null" we could exploit this
even more, e.g., remove all parameter combinations that would lead to
the execution of this statement from the context.
llvm-svn: 264284
Similar to r262612 we need to check not only the pointer SCEV and the
type of an alias group but also the actual access instruction. The
reason is again the same: The pointer SCEV is not flow sensitive but the
access function is. In r262612 we avoided consolidating alias groups
even though the pointer SCEV and the type were the same but the access
function was not. Here it is simpler as we can simply check all members
of an alias group against the given access instruction.
llvm-svn: 264274
When codegenerating invariant loads in some rare cases we cannot generate code
and bail out. This change ensures that we maintain a valid dominator tree
in these situations. This fixes llvm.org/PR26736
Contributed-by: Matthias Reisinger <d412vv1n@gmail.com>
llvm-svn: 264142
This might be useful to evaluate the benefit of us handling modref funciton
calls. Also, a new bug that was triggered by modref function calls was
recently reported http://llvm.org/PR27035. To ensure the same issue does not
cause troubles for other people, we temporarily disable this until the bug
is resolved.
llvm-svn: 264140
ISL can conclude additional conditions on parameters from restrictions
on loop variables. Such conditions persist when leaving the loop and the
loop variable is projected out. This results in a narrower domain for
exiting the loop than entering it and is logically impossible for
non-infinite loops.
We fix this by not adding a lower bound i>=0 when constructing BB
domains, but defer it to when also the upper bound it computed, which
was done redundantly even before this patch.
This reduces the number of LNT fails with -polly-process-unprofitable
-polly-position=before-vectorizer from 8 to 6.
llvm-svn: 264118
We bail out if current scop has a complex control flow as this could lead to
building of large domain conditions. This is to reduce compile time. This
addresses r26382.
Contributed-by: Chris Jenneisch <chrisj@codeaurora.org>
Differential Revision: http://reviews.llvm.org/D18362
llvm-svn: 264105
Affine branches are fully modeled and regenerated from the polyhedral domain and
consequently do not require any input conditions to be propagated.
llvm-svn: 263678
Index calculations can use the last value that come out of a loop.
Ideally, ScalarEvolution can compute that exit value directly without
depending on the loop induction variable, but not in all cases.
This changes isAffine to not consider such loop exit values as affine to
avoid that SCEVExpander adds uses of the original loop induction
variable.
This fix is analogous to r262404 that applies to general uses of loop
exit values instead of index expressions and loop bouds as in this
patch.
This reduces the number of LNT test-suite fails with
-polly-position=before-vectorizer -polly-unprofitable
from 10 to 8.
llvm-svn: 262665
Value merging is only necessary for scalars when they are used outside
of the scop. While an array's base pointer can be used after the scop,
it gets an extra ScopArrayInfo of type MK_Value. We used to generate
phi's for both of them, where one was assuming the reault of the other
phi would be the original value, because it has already been replaced by
the previous phi. This resulted in IR that the current IR verifier
allows, but is probably illegal.
This reduces the number of LNT test-suite fails with
-polly-position=before-vectorizer -polly-process-unprofitable
from 16 to 10.
Also see llvm.org/PR26718.
llvm-svn: 262629
This should fix PR19422.
Thanks to Jeremy Huddleston Sequoia for reporting this.
Thanks to Roman Gareev for his investigation and the reduced test case.
llvm-svn: 262612
Polly recognizes affine loops that ScalarEvolution does not, in
particular those with loop conditions that depend on hoisted invariant
loads. Check for SCEVAddRec dependencies on such loops and do not
consider their exit values as synthesizable because SCEVExpander would
generate them as expressions that depend on the original induction
variables. These are not available in generated code.
llvm-svn: 262404
In order to speed up compile time and to avoid random timeouts we now
separately track assumptions and restrictions. In this context
assumptions describe parameter valuations we need and restrictions
describe parameter valuations we do not allow. During AST generation
we create a runtime check for both, whereas the one for the
restrictions is negated before a conjunction is build.
Except the In-Bounds assumptions we currently only track restrictions.
Differential Revision: http://reviews.llvm.org/D17247
llvm-svn: 262328
Originally committed in r261899 and reverted in r262202 due to failing
in out-of-LLVM tree builds.
Replace the use of LLVM_TOOLS_BINARY_DIR by LLVM_TOOLS_DIR which exists
in both, in-tree and out-of-tree builds.
Original commit message:
The script updates a lit test case that uses FileCheck using the actual
output of the 'RUN:'-lines program. Useful when updating test cases due
to expected output changes and diff'ing expected and actual output.
llvm-svn: 262227
This reverts commit r261899. Even though I am not yet 100% certain, this is
commit is the only one that has some relation to the recent cmake failures
in Polly.
llvm-svn: 262202
The script updates a lit test case that uses FileCheck using the actual
output of the 'RUN:'-lines program. Useful when updating test cases due
expected output changes and diff'ing expected and actual output.
llvm-svn: 261899
Check the ModRefBehaviour of functions in order to decide whether or
not a call instruction might be acceptable.
Differential Revision: http://reviews.llvm.org/D5227
llvm-svn: 261866
The generated dedicated subregion exit block was assumed to have the same
dominance relation as the original exit block. This is incorrect if the exit
block receives other edges than only from the subregion, which results in that
e.g. the subregion's entry block does not dominate the exit block.
llvm-svn: 261865
From now on we bail only if a non-trivial alias group contains a non-affine
access, not when we discover aliasing and non-affine accesses are allowed.
llvm-svn: 261863
The test style guide defines that opt should get its input from stdin.
(instead by file argument to avoid that the file name appears in its
output)
CHECK-FORCED is not recognized by FileCheck; remove it.
llvm-svn: 261786
This patch adds support for memcpy, memset and memmove intrinsics. They are
represented as one (memset) or two (memcpy, memmove) memory accesses in the
polyhedral model. These accesses have an access range that describes the
summarized effect of the intrinsic, i.e.,
memset(&A[i], '$', N);
is represented as a write access from A[i] to A[i+N].
Differential Revision: http://reviews.llvm.org/D5226
llvm-svn: 261489
To support non-aligned accesses we introduce a virtual element size
for arrays that divides each access function used for this array. The
adjustment of the access function based on the element size of the
array was therefore moved after this virtual element size was
determined, thus after all accesses have been created.
Differential Revision: http://reviews.llvm.org/D17246
llvm-svn: 261226
A load can only be invariant if its base pointer is invariant too. To
this end, we check if the base pointer is defined inside the region or
outside. In the former case we recursively check if we can (and
therefore will) hoist the base pointer too. Only if that happends we
can hoist the load.
llvm-svn: 260886
This reverts commit 98efa006c96ac981c00d2e386ec1102bce9f549a.
The fix was broken since we do not use AA in the ScopDetection anymore to
check for invariant accesses.
llvm-svn: 260884
Before this patch it could happen that we did not hoist a load that
was a base pointer of another load even though AA already declared the
first one as invariant (during ScopDetection). If this case arises we
will now skipt the "can be overwriten" check because in this case the
over-approximating nature causes us to generate broken code.
llvm-svn: 260862
So far we separated constant factors from multiplications, however,
only when they are at the outermost level of a parameter SCEV. Now,
we also separate constant factors from the parameter SCEV if the
outermost expression is a SCEVAddRecExpr. With the changes to the
SCEVAffinator we can now improve the extractConstantFactor(...)
function at will without worrying about any other code part. Thus,
if needed we can implement a more comprehensive
extractConstantFactor(...) function that will traverse the SCEV
instead of looking only at the outermost level.
Four test cases were affected. One did not change much and the other
three were simplified.
llvm-svn: 260859
We now distinguish invariant loads to the same memory location if they
have different types. This will cause us to pre-load an invariant
location once for each type that is used to access it. However, we can
thereby avoid invalid casting, especially if an array is accessed
though different typed/sized invariant loads.
This basically reverts the changes in r260023 but keeps the test
cases.
llvm-svn: 260045
We also disable this feature by default, as there are still some issues in
combination with invariant load hoisting that slipped through my initial
testing.
llvm-svn: 260025
Invariant load hoisting of memory accesses with non-canonical element
types lacks support for equivalence classes that contain elements of
different width/size. This support should be added, but to get our buildbots
back to green, we disable load hoisting for memory accesses with non-canonical
element size for now.
llvm-svn: 260023
Always use access-instruction pointer type to load the invariant values.
Otherwise mismatches between ScopArrayInfo element type and memory access
element type will result in invalid casts. These type mismatches are after
r259784 a lot more common and also arise with types of different size, which
have not been handled before.
Interestingly, this change actually simplifies the code, as we now have only
one code path that is always taken, rather then a standard code path for the
common case and a "fixup" code path that replaces the standard code path in
case of mismatching types.
llvm-svn: 260009
The previously implemented approach is to follow value definitions and
create write accesses ("push defs") while searching for uses. This
requires the same relatively validity- and requirement conditions to be
replicated at multiple locations (PHI instructions, other instructions,
uses by PHIs).
We replace this by iterating over the uses in a SCoP ("pull in
requirements"), and add writes only when at least one read has been
added. It turns out to be simpler code because each use is only iterated
over once and writes are added for the first access that reads it. We
need another iteration to identify escaping values (uses not in the
SCoP), which also makes the difference between such accesses more
obvious. As a side-effect, the order of scalar MemoryAccess can change.
Differential Revision: http://reviews.llvm.org/D15706
llvm-svn: 259987
This allows code such as:
void multiple_types(char *Short, char *Float, char *Double) {
for (long i = 0; i < 100; i++) {
Short[i] = *(short *)&Short[2 * i];
Float[i] = *(float *)&Float[4 * i];
Double[i] = *(double *)&Double[8 * i];
}
}
To model such code we use as canonical element type of the modeled array the
smallest element type of all original array accesses, if type allocation sizes
are multiples of each other. Otherwise, we use a newly created iN type, where N
is the gcd of the allocation size of the types used in the accesses to this
array. Accesses with types larger as the canonical element type are modeled as
multiple accesses with the smaller type.
For example the second load access is modeled as:
{ Stmt_bb2[i0] -> MemRef_Float[o0] : 4i0 <= o0 <= 3 + 4i0 }
To support code-generating these memory accesses, we introduce a new method
getAccessAddressFunction that assigns each statement instance a single memory
location, the address we load from/store to. Currently we obtain this address by
taking the lexmin of the access function. We may consider keeping track of the
memory location more explicitly in the future.
We currently do _not_ handle multi-dimensional arrays and also keep the
restriction of not supporting accesses where the offset expression is not a
multiple of the access element type size. This patch adds tests that ensure
we correctly invalidate a scop in case these accesses are found. Both types of
accesses can be handled using the very same model, but are left to be added in
the future.
We also move the initialization of the scop-context into the constructor to
ensure it is already available when invalidating the scop.
Finally, we add this as a new item to the 2.9 release notes
Reviewers: jdoerfert, Meinersbur
Differential Revision: http://reviews.llvm.org/D16878
llvm-svn: 259784
We support now code such as:
void multiple_types(char *Short, char *Float, char *Double) {
for (long i = 0; i < 100; i++) {
Short[i] = *(short *)&Short[2 * i];
Float[i] = *(float *)&Float[4 * i];
Double[i] = *(double *)&Double[8 * i];
}
}
To support such code we use as element type of the modeled array the smallest
element type of all original array accesses. Accesses with larger types are
modeled as multiple accesses with the smaller type.
For example the second load access is modeled as:
{ Stmt_bb2[i0] -> MemRef_Float[o0] : 4i0 <= o0 <= 3 + 4i0 }
To support jscop-rewritable memory accesses we need each statement instance to
only be assigned a single memory location, which will be the address at which
we load the value. Currently we obtain this address by taking the lexmin of
the access function. We may consider keeping track of the memory location more
explicitly in the future.
llvm-svn: 259587
For schedule generation we assumed that the reverse post order traversal used by
the domain generation is sufficient, however it is not. Once a loop is
discovered, we have to completely traverse it, before we can generate the
schedule for any block/region that is only reachable through a loop exiting
block.
To this end, we add a "loop stack" that will keep track of loops we
discovered during the traversal but have not yet traversed completely.
We will never visit a basic block (or region) outside the most recent
(thus smallest) loop in the loop stack but instead queue such blocks
(or regions) in a waiting list. If the waiting list is not empty and
(might) contain blocks from the most recent loop in the loop stack the
next block/region to visit is drawn from there, otherwise from the
reverse post order iterator.
We exploit the new property of loops being always completed before additional
loops are processed, by removing the LoopSchedules map and instead keep all
information in LoopStack. This clarifies that we indeed always only keep a
stack of in-process loops, but will never keep incomplete schedules for an
arbitrary set of loops. As a result, we can simplify some of the existing code.
This patch also adds some more documentation about how our schedule construction
works.
This fixes http://llvm.org/PR25879
This patch is an modified version of Johannes Doerfert's initial fix.
Differential Revision: http://reviews.llvm.org/D15679
llvm-svn: 259354
The autotools build system is based on and requires LLVM's autotools
build system to work, which has been depricated and finally removed in
r258861. Consequently we also remove the autotools build system from
Polly.
Differential Revision: http://reviews.llvm.org/D16655
llvm-svn: 259041
Before adding a MK_Value READ MemoryAccess, check whether the read is
necessary or synthesizable. Synthesizable values are later generated by
the SCEVExpander and therefore do not need to be transferred
explicitly. This can happen because the check for synthesizability has
presumbly been forgotten in the case where a phi's incoming value has
been defined in a different statement.
Differential Revision: http://reviews.llvm.org/D15687
llvm-svn: 258998
Michael Kruse