Aligning data to cache lines boundaries helps to avoid overheads related to
an access to it ([1]). This patch aligns newly created arrays and adds an
option to specify the first level cache line size. By default we use 64 bytes,
which is a typical cache-line size ([2]).
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 12.63 GFlops/sec (43,8542% of theoretical peak).
Refs.:
[1] - http://www.alexonlinux.com/aligned-vs-unaligned-memory-access
[2] - http://igoro.com/archive/gallery-of-processor-cache-effects/
Differential Revision: https://reviews.llvm.org/D28020
Reviewed-by: Tobias Grosser <tobias@grosser.es>
llvm-svn: 290253
In '[DBG] Allow to emit the RTC value at runtime' the diagnostics were printed
without a newline at the end of each diagnostic. We add such a newline to
improve readability.
llvm-svn: 288323
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
The new command line flag "polly-codegen-emit-rtc-print" can be used to
place a "printf" in the generated code that will print the RTC value and
the overflow state.
llvm-svn: 287265
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
This makes polly generate a CFG which is closer to what we want
in LLVM IR, with a loop preheader for the original loop. This is
just a cleanup, but it exposes some fragile assumptions.
I'm not completely happy with the changes related to expandCodeFor;
RTCBB->getTerminator() is basically a random insertion point which
happens to work due to the way we generate runtime checks. I'm not
sure what the right answer looks like, though.
Differential Revision: https://reviews.llvm.org/D26053
llvm-svn: 285864
Summary: Iterating over SeenBlocks which is a SmallPtrSet results in non-determinism in codegen
Reviewers: jdoerfert, zinob, grosser
Tags: #polly
Differential Revision: https://reviews.llvm.org/D25778
llvm-svn: 284622
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
The core of the change is supposed to be NFC, however it also fixes
what I believe was an undefined behavior when calling:
va_start(ValueArgs, Desc);
with Desc being a StringRef.
Differential Revision: https://reviews.llvm.org/D25342
llvm-svn: 283671
Currently Polly cannot generate code for index expressions if the base pointer
is computed within the scop. The base pointer must be generated as well, but
there is no code that triggers that.
Add an assertion to detect when this would occur and miscompile. The IR verifier
should catch it as well.
llvm-svn: 282893
generateScalarLoad() and generateScalarStore() are used for explicit (MK_Array)
memory accesses, therefore the method names were misleading. The names also
were similar to generateScalarLoads() and generateScalarStores() (plural forms)
which indeed handle scalar accesses. Presumbly, they were originally named to
contrast VectorBlockGenerator::generateLoad().
Rename the two methods to generateArrayLoad(),
respectively generateArrayStore().
llvm-svn: 282861
The code generator always adds unconditional LoadInst and StoreInst, hence the
MemoryAccess must be defined over all statement instances.
llvm-svn: 282853
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
LLVM's coding guideline suggests to not use @brief for one-sentence doxygen
comments to improve readability. Switch this once and for all to ensure people
do not copy @brief comments from other parts of Polly, when writing new code.
llvm-svn: 280468
Change the code around setNewAccessRelation to allow to use a an existing array
element for memory instead of an ad-hoc alloca. This facility will be used for
DeLICM/DeGVN to convert scalar dependencies into regular ones.
The changes necessary include:
- Make the code generator use the implicit locations instead of the alloca ones.
- A test case
- Make the JScop importer accept changes of scalar accesses for that test case.
- Adapt the MemoryAccess interface to the fact that the MemoryKind can change.
They are named (get|is)OriginalXXX() to get the status of the memory access
before any change by setNewAccessRelation() (some properties such as
getIncoming() do not change even if the kind is changed and are still
required). To get the modified properties, there is (get|is)LatestXXX(). The
old accessors without Original|Latest become synonyms of the
(get|is)OriginalXXX() to not make functional changes in unrelated code.
Differential Revision: https://reviews.llvm.org/D23962
llvm-svn: 280408
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
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
There is no need to reset the position of the builder, as we can just continue
to insert code at the current position of the IRBuilder, which happens to
be precisely the location we reset the builder to.
llvm-svn: 278014
... instead of adding instructions at the end of the basic block the builder
is currently at. This makes it easier to reason about where IR is generated,
as with the IRBuilder there is just a single location that specificies where
IR is generated.
llvm-svn: 278013
The map is iterated over when generating the values escaping the SCoP. The
indeterministic iteration order of DenseMap causes the output IR to change at
every compilation, adding noise to comparisons.
Replace DenseMap by a MapVector to ensure the same iteration order at every
compilation.
llvm-svn: 277832
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
Roman Gareev