This is a stylistic change to make the function a little more readable.
Also add a debug print to show what instruction contains a use of a
function we don't understand in the kernel.
Differential Revision: https://reviews.llvm.org/D37058
llvm-svn: 311648
Add statistics about
- Which optimizations are applied
- Number of loops in Scops at various stages
- Number of scalar/singleton writes at various stages representative
for scalar false dependencies
- Number of parallel loops
These will be useful to find regressions due to moving Polly further
down of LLVM's pass pipeline.
Differential Revision: https://reviews.llvm.org/D37049
llvm-svn: 311553
MSVC warns about comparison between a signed and unsigned integer.
The rules of C(++) define that an unsigned comparison has to be
carried-out in this case. This is unlikely to be intended.
Fix by assigning the loop's upper bound to a signed integer first.
This also avoids repeated evaluation of the invariant upper bound.
llvm-svn: 311548
Summary:
There is no need to emit alias metadata for scalars, as basicaa will easily
distinguish them from arrays. This reduces the size of the metadata we generate.
This is especially useful after we moved to -polly-position=before-vectorizer,
where a lot more scalar dependences are introduced, which increased the size of
the alias analysis metadata and made us commonly reach the limits after which
we do not emit alias metadata that have been introduced to prevent quadratic
growth of this alias metadata.
This improves 2mm performance from 1.5 seconds to 0.17 seconds.
Reviewers: Meinersbur, bollu, singam-sanjay
Reviewed By: Meinersbur
Subscribers: pollydev, llvm-commits
Tags: #polly
Differential Revision: https://reviews.llvm.org/D37028
llvm-svn: 311498
Currently, in case of GEMM and the pattern matching based optimizations, we
use only the SLP Vectorizer out of two LLVM vectorizers. Since the Loop
Vectorizer can get in the way of optimal code generation, we disable the Loop
Vectorizer for the innermost loop using mark nodes and emitting the
corresponding metadata.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D36928
llvm-svn: 311473
This feature was not enabled for `PPCGCodeGeneration`. Now that this is
enabled, we can benchmark Scops that have been optimised with
`-polly-codegen-ppcg` with the `-polly-codegen-perf-monitoring` option.
Differential Revision: https://reviews.llvm.org/D36934
llvm-svn: 311328
We still see some issues with parameter space mismatches. Revert this to get
a clean baseline. We will recommit after these issues have been resolved.
This reverts commit 0e360a14194f722ded7aa2bc9d4be2ed2efeeb49.
llvm-svn: 311268
Instead of using Twines and temporary expressions, we do string manipulation
through a std::string. This resolves a memory corruption issue, which likely
was caused by twines loosing their underlying string too soon.
llvm-svn: 311264
- We should iterate over `I`, which is `Cur` expanded out to an
instruction, and not `Cur` itself.
- This is a bugfix.
Differential Revision: https://reviews.llvm.org/D36923
llvm-svn: 311261
Summary:
This information is necessary for PPCG to perform correct life range reordering.
With these changes applied we can live-range reorder some of the important
kernels in COSMO.
We also update and rename one test case, which previously could not be optimized
and now is optimized thanks to live-range reordering. To preserve test coverage
we add a new test case scalar-writes-in-scop-requires-abort.ll, which exercises
our automatic abort in case of scalar writes in the kernel.
Reviewers: Meinersbur, bollu, singam-sanjay
Subscribers: nemanjai, pollydev, llvm-commits, kbarton
Tags: #polly
Differential Revision: https://reviews.llvm.org/D36929
llvm-svn: 311259
Kernel argument sizes now only get appended to the kernel launch parameter list if the OpenCL runtime is selected, not if CUDA runtime is chosen.
Differential revision: D36925
llvm-svn: 311248
When using -polly-ignore-integer-wrapping and -polly-acc-codegen-managed-memory
we add parameter dimensions lazily to the domains, which results in PPCG not
including parameter dimensions that are only used in memory accesses in the
kernel space. To make sure these parameters are still passed to the kernel, we
collect these parameter dimensions and align the kernel's parameter space
before code-generating it.
llvm-svn: 311239
Summary:
Drop unused parameter dimensions to reduce the size of the sets we are working
with. Especially the computed dependences tend to accumulate a lot of parameters
that are present in the input memory accesses, but often not necessary to
express the actual dependences. As isl represents maps and sets with dense
matrices, reducing the dimensionality of isl sets commonly reduces code
generation performance.
This reduces compile time from 17 to 11 seconds for our test case. While this is
not impressive, this patch helped me to identify the previous two performance
improvements and additionally also increases readability of the isl data
structures we use.
Reviewers: Meinersbur, bollu, singam-sanjay
Reviewed By: bollu
Subscribers: nemanjai, pollydev, llvm-commits, kbarton
Tags: #polly
Differential Revision: https://reviews.llvm.org/D36869
llvm-svn: 311161
Summary:
They are not used and consequently do not even need to be computed. This reduces
the overall compile time for our kernel from 1m33s to 17s.
Reviewers: Meinersbur, bollu, singam-sanjay
Reviewed By: bollu
Subscribers: nemanjai, pollydev, llvm-commits, kbarton
Tags: #polly
Differential Revision: https://reviews.llvm.org/D36868
llvm-svn: 311157
Summary:
This change reduces the overall number of synchronize calls for kernels with
a lot of output data at the cost of additional synchronize calls for kernels
launched in sequence without any device to host transfers in between. As the
latter pattern is a lot less frequent, this seems a better tradeoff.
Even though the above motivation would be motivation enough, this is just
a step towards enabling ppcg to not compute to and from device copy calls
at all, which would be incorrect in case we still relied on these calls to
place our synchronization statements.
Reviewers: Meinersbur, bollu, singam-sanjay
Reviewed By: bollu
Subscribers: nemanjai, kbarton, pollydev, llvm-commits
Tags: #polly
Differential Revision: https://reviews.llvm.org/D36867
llvm-svn: 311155
This avoid the construction of very large sets and in many cases also keeps the
number of parameters low. As a result, we see a compile time reduction from 5
minutes to only slightly above 1 minute for one of our larger test cases.
llvm-svn: 311127
Reuse the machinery built for replacing global arrays to replace malloc/free as
well. Example replacement that was missed earlier:
```
call void \
bitcast (void (i8*)* @free to void (%custom_type*)*) (%custom_type* %13)
```
- Since the `bitcast` is a `ConstantExpr`, `replaceAllUsesWith` would miss
this. We don't miss this anymore.
Differential Revision: https://reviews.llvm.org/D36825
llvm-svn: 311121
- If we have global arrays, we would like to rewrite them to global
pointers which are allocated using `cudaMallocManaged`.
- If we have allocas in a function, we would like to rewrite them to
heap-allocations with `cudaMallocManaged` and `cudaFree`.
- With these rewrite mechanisms, we can offload _any_ function to the
GPU with no code rewrite whatsover.
Differential Revision: https://reviews.llvm.org/D36516
llvm-svn: 311080
Summary:
During code generation for a Scop we modify the IR of a function.
While this shouldn't affect a Scop in the formal sense, the implementation
caches various information about the IR such as SCEV expressions for bounds or
parameters. This cached information needs to be updated or invalidated. To this
end, SPMUpdater allows passes to report when they've invalidated a Scop to the
PassManager, which will then flush and recompute all Scops. This in turn
invalidates all iterators, so references to Scops shouldn't be held.
Reviewers: grosser, Meinersbur, bollu
Reviewed By: grosser
Subscribers: llvm-commits, pollydev
Differential Revision: https://reviews.llvm.org/D36524
llvm-svn: 310551
This pass is useful to automatically convert a codebase that uses malloc/free
to use their managed memory counterparts.
Currently, rewrite malloc and free to the `polly_{malloc,free}Managed` variants.
A future patch will teach ManagedMemoryRewrite to rewrite global arrays
as pointers to globally allocated managed memory.
Differential Revision: https://reviews.llvm.org/D36513
llvm-svn: 310471
Codegen with -polly-parallel queried the unmapped MemoryAccess, but only
the MemoryKind after mapping is relevant for codegen.
This should fix various fails of the
perf-x86_64-penryn-O3-polly-parallel-fast buildbot.
llvm-svn: 310466
Previously, we used to compute this with `elementSizeInBits / 8`. This
would yield an element size of 0 when the array had element size < 8 in
bits.
To fix this, ask data layout what the size in bytes should be.
Differential Revision: https://reviews.llvm.org/D36459
llvm-svn: 310448
We introduce another level of alias metadata to distinguish the individual
non-aliasing accesses that have inter iteration alias-free base pointers
marked with "Inter iteration alias-free" mark nodes. To distinguish two
accesses, the comparison of raw pointers representing base pointers is used.
In case of, for example, ublas's prod function that implements GEMM, and
DeLiCM we can get accesses to same location represented by different raw
pointers. Consequently, we create different alias sets that can prevent
accesses from, for example, being sinked or hoisted.
To avoid the issue, we compare the corresponding SCEV information instead
of the corresponding raw pointers.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D35761
llvm-svn: 310380
To do this, we replicate what `CodeGeneration` does. We expose
`markNodeUnreachable` from `CodeGeneration` to `PPCGCodeGeneration`.
Differential Revision: https://reviews.llvm.org/D36457
llvm-svn: 310350
Summary:
This resolves some "instruction does not dominate use" errors, as we used to
prepare the arrays at the location of the first kernel, which not necessarily
dominated all other kernel calls.
Reviewers: Meinersbur, bollu, singam-sanjay
Subscribers: nemanjai, pollydev, llvm-commits, kbarton
Differential Revision: https://reviews.llvm.org/D36372
llvm-svn: 310196
A Scop with a loop outside it is not handled currently by
PPCGCodeGeneration. The test case is such that the Scop has only one inner loop
that is detected. This currently breaks codegen.
The fix is to reuse the existing mechanism in `IslNodeBuilder` within
`GPUNodeBuilder.
Differential Revision: https://reviews.llvm.org/D36290
llvm-svn: 310193
This logic is duplicated, so we refactor it into a separate function.
This will be used in a later patch to teach PPCGCodeGen code generation
for loops that are outside the scop.
Differential Revision: https://reviews.llvm.org/D36310
llvm-svn: 310192
Summary:
In case the option -polly-ignore-parameter-bounds is set, not all parameters
will be added to context and domains. This is useful to keep the size of the
sets and maps we work with small. Unfortunately, for AST generation it is
necessary to ensure all parameters are part of the schedule tree. Hence,
we modify the GPGPU code generation to make sure this is the case.
To obtain the necessary information we expose a new function
Scop::getFullParamSpace(). We also make a couple of functions const to be
able to make SCoP::getFullParamSpace() const.
Reviewers: Meinersbur, bollu, gareevroman, efriedma, huihuiz, sebpop, simbuerg
Subscribers: nemanjai, kbarton, pollydev, llvm-commits
Tags: #polly
Differential Revision: https://reviews.llvm.org/D36243
llvm-svn: 309939
When we have `-polly-ignore-parameter-bounds`, `Scop::Context` does not contain
all the paramters present in the program.
The construction of the `isl_multi_pw_aff` requires all the indivisual `pw_aff`
to have the same parameter dimensions. To achieve this, we used to realign
every `pw_aff` with `Scop::Context`. However, in conjunction with
`-polly-ignore-parameter-bounds`, this is now incorrect, since `Scop::Context`
does not contain all parameters.
We set this up correctly by creating a space that has all the parameters
used by all the `isl_pw_aff`. Then, we realign all `isl_pw_aff` to this space.
llvm-svn: 309934
Summary:
**Remove debug metadata from instruction to be copied to prevent the source file's debug metadata being copied into GPUModule and eventually failing Module verification and ASM string codegeneration.**
When copying the instruction onto the Module meant for the GPU, debug metadata attached to an instruction causes all related metadata to be pulled into the Module, including the DICompileUnit, which is not listed in llvm.dbg.cu of the Module. This fails the verification of the Module and generation of the ASM string.
The only debug metadata of the instruction, the DebugLoc, is unset by this patch.
This patch reattempts https://reviews.llvm.org/D35630 by targeting only those instructions that are to end up in a Module meant for the GPU.
Reviewers: grosser, bollu
Reviewed By: grosser
Subscribers: pollydev
Tags: #polly
Differential Revision: https://reviews.llvm.org/D36161
llvm-svn: 309822
It is possible that the `HostPtr` that coresponds to an array could be
invariant load hoisted. Make sure we use the invariant load hoisted
value by using `IslNodeBuilder::getLatestValue`.
Differential Revision: https://reviews.llvm.org/D36001
llvm-svn: 309681
We populate `IslNodeBuilder::ValueMap` which contains replacements for
`llvm::Value`s. There was no simple method to pick up a replacement if
it exists, otherwise fall back to the original.
Create a method `IslNodeBuilder::getLatestValue` which provides this
functionality.
This will be used in a later patch to fix bugs in `PPCGCodeGeneration`
where the latest value is not being used.
Differential Revision: https://reviews.llvm.org/D36000
llvm-svn: 309674
Summary:
This allows us to map functions such as exp, expf, expl, for which no
LLVM intrinsics exist. Instead, we link to NVIDIA's libdevice which provides
high-performance implementations of a wide range of (math) functions. We
currently link only a small subset, the exp, cos and copysign functions. Other
functions will be enabled as needed.
Reviewers: bollu, singam-sanjay
Reviewed By: bollu
Subscribers: tstellar, tra, nemanjai, pollydev, mgorny, llvm-commits, kbarton
Tags: #polly
Differential Revision: https://reviews.llvm.org/D35703
llvm-svn: 309560
This reverts commit r309490 as it triggers on our AOSP buildbut error messages
of the form:
inlinable function call in a function with debug info must have a !dbg location
llvm-svn: 309556
Summary:
**Remove debug metadata from instruction to be copied to prevent the source file's debug metadata being copied into GPUModule and eventually failing Module verification and ASM string codegeneration.**
When copying the instruction onto the Module meant for the GPU, debug metadata attached to an instruction causes all related metadata to be pulled into the Module, including the DICompileUnit, which is not listed in llvm.dbg.cu of the Module. This fails the verification of the Module and generation of the ASM string.
The only debug metadata of the instruction, the DebugLoc, is unset by this patch.
Reviewers: grosser, bollu, Meinersbur
Reviewed By: grosser, bollu
Subscribers: pollydev
Tags: #polly
Differential Revision: https://reviews.llvm.org/D35630
llvm-svn: 309490
In future, there will be no more a 1:1 correspondence between statements
and basic blocks, the name `contains` does not correctly capture their
relationship. A BB may infact comprise of multiple statements; hence we
describe a statement 'representing' a basic block.
Differential Revision: https://reviews.llvm.org/D35838
llvm-svn: 308982
Invariant load hoisted scalars, and arrays whose size we can statically compute
to be 0 do not need to be allocated as arrays.
Invariant load hoisted scalars are sent to the kernel directly as parameters.
Earlier, we used to allocate `0` bytes of memory for these because our
computation of size from `PPCGCodeGeneration::getArraySize` would result in `0`.
Now, since we don't invariant loads as arrays in PPCGCodeGeneration, this
problem does not occur anymore.
Differential Revision: https://reviews.llvm.org/D35795
llvm-svn: 308971
Summary:
Added SPIR Code Generation to the PPCG Code Generator. This can be invoked using
the polly-gpu-arch flag value 'spir32' or 'spir64' for 32 and 64 bit code respectively.
In addition to that, runtime support has been added to execute said SPIR code on Intel
GPU's, where the system is equipped with Intel's open source driver Beignet (development
version). This requires the cmake flag 'USE_INTEL_OCL' to be turned on, and the polly-gpu-runtime
flag value to be 'libopencl'.
The transformation of LLVM IR to SPIR is currently quite a hack, consisting in part of regex
string transformations.
Has been tested (working) with Polybench 3.2 on an Intel i7-5500U (integrated graphics chip).
Reviewers: bollu, grosser, Meinersbur, singam-sanjay
Reviewed By: grosser, singam-sanjay
Subscribers: pollydev, nemanjai, mgorny, Anastasia, kbarton
Tags: #polly
Differential Revision: https://reviews.llvm.org/D35185
llvm-svn: 308751
Print a statement's instruction on dump() regardless of
-polly-print-instructions. dump() is supposed to be used in the debugger
only and never in regression tests. While debugging, get all the
information we have and we are not bound to break anything. For non-dump
purposes of print, forward the setting of -polly-print-instructions as
parameters.
Some calls to print() had to be changed because the
PollyPrintInstructions setting is only available in ScopInfo.cpp.
In ScheduleOptimizer.cpp, dump() was used in regression tests.
That's not what dump() is for.
The print parameter "PrintInstructions" will also be useful for an
explicit print SCoP pass in a future patch.
llvm-svn: 308746
When performing invariant load hoisting we check that invariant load expressions
are not too complex. Up to this commit, we performed this check by counting the
sum of dimensions in the access range as a very simple heuristic. This heuristic
is a little too conservative, as it prevents hoisting for any scops with a
very large number of parameters. Hence, we update the heuristic to only count
existentially quantified dimensions and set dimensions. We expect this to still
detect the problematic expressions in h264 because of which this check was
originally introduced.
For some unknown reason, this complexity check was originally committed in
IslNodeBuilder. It really belongs in ScopInfo, as there is no point in
optimizing a program which we could have known earlier cannot be code generated.
The benefit of running the check early is that we can avoid to even hoist checks
that are expensive to code generate as invariant loads. This can be seen in
the changed tests, where we now indeed detect the scop, but just not invariant
load hoist the complicated access.
We also improve the formatting of the code, document it, and use isl++ to
simplify expressions.
llvm-svn: 308659
This commit *WILL COMPILE*.
1. `PPCG` now uses `isl_multi_pw_aff` instead of an array of `pw_aff`.
This needs us to adjust how we index array bounds and how we construct
array bounds.
2. `PPCG` introduces two new kinds of nodes: `init_device` and `clear_device`.
We should investigate what the correct way to handle these are.
3. `PPCG` has gotten smarter with its use of live range reordering, so some of
the tests have a qualitative improvement.
4. `PPCG` changed its output style, so many test cases need to be updated to
fit the new style for `polly-acc-dump-code` checks.
Differential Revision: https://reviews.llvm.org/D35677
llvm-svn: 308625
We extended kills in Polly to handle both `phi` nodes and scalars that
are not used within the Scop. Update the comments and choice of
variable names to reflect this.
llvm-svn: 308279
- We should call `preloadInvariantLoads` to make sure that code is
generated for invariant loads in the kernel.
Differential Revision: https://reviews.llvm.org/D35410
llvm-svn: 308187
- There is a conditional branch that is used to switch between the old
and new versions of the code.
- If we detect that the build was unsuccessful, `PPCGCodeGeneration` will
change the runtime check to be always set to false.
- To actually *reach* this runtime check instruction, `PPCGCodeGeneration`
was using assumptions about the layout of the BBs.
- However, invariant load hoisting violates this assumption by inserting
an extra basic block in the middle.
- Fix the assumption on the layout by having `createScopConditionally`
return the conditional branch instruction.
- Use this reference to set to always-false.
llvm-svn: 308010
We need to relax constraints on invariant loads so that they do not
create fake RAW dependences. So, we do not consider invariant loads as
scalar dependences in a region.
During these changes, it turned out that we do not consider `llvm::Value`
replacements correctly within `PPCGCodeGeneration` and `ISLNodeBuilder`.
The replacements dictated by `ValueMap` were not being followed in all
places. This was fixed in this commit. There is no clean way to decouple
this change because this bug only seems to arise when the relaxed
version of invariant load hoisting was enabled.
Differential Revision: https://reviews.llvm.org/D35120
llvm-svn: 307907
Summary:
Add a sequence number that identifies a ptx_kernel's parent Scop within a function to it's name to differentiate it from other kernels produced from the same function, yet different Scops.
Kernels produced from different Scops can end up having the same name. Consider a function with 2 Scops and each Scop being able to produce just one kernel. Both of these kernels have the name "kernel_0". This can lead to the wrong kernel being launched when the runtime picks a kernel from its cache based on the name alone. This patch supplements D33985, by differentiating kernels across Scops as well.
Previously (even before D33985) while profiling kernels generated through JIT e.g. Julia, [[ https://groups.google.com/d/msg/polly-dev/J1j587H3-Qw/mR-jfL16BgAJ | kernels associated with different functions, and even different SCoPs within a function, would be grouped together due to the common name ]]. This patch prevents this grouping and the kernels are reported separately.
Reviewers: grosser, bollu
Reviewed By: grosser
Subscribers: mehdi_amini, nemanjai, pollydev, kbarton
Tags: #polly
Differential Revision: https://reviews.llvm.org/D35176
llvm-svn: 307814
When providing the option "-polly-ast-print-accesses" Polly also prints the
memory accesses that are generated:
#pragma known-parallel
for (int c0 = 0; c0 <= 1023; c0 += 4)
#pragma simd
for (int c1 = c0; c1 <= c0 + 3; c1 += 1)
Stmt_for_body(
/* read */ &MemRef_B[0]
/* write */ MemRef_A[c1]
);
This makes writing and debugging memory layout transformations easier.
Based on a patch contributed by Thomas Lang (ETH Zurich)
llvm-svn: 307579
- By definition, we can pass something as a `kill` to PPCG if we know
that no data can flow across a kill.
- This is useful for more complex examples where we have scalars that
are local to a scop.
- If the local is only used within a scop, we are free to kill it.
Differential Revision: https://reviews.llvm.org/D35045
llvm-svn: 307260
Summary:
Provide more context to the name of a GPU kernel by prefixing its name with the host function that calls it. E.g. The first kernel called by `gemm` would be `FUNC_gemm_KERNEL_0`.
Kernels currently follow the "kernel_#" (# = 0,1,2,3,...) nomenclature. This patch makes it easier to map host caller and device callee, especially when there are many kernels produced by Polly-ACC.
Reviewers: grosser, Meinersbur, bollu, philip.pfaffe, kbarton!
Reviewed By: grosser
Subscribers: nemanjai, pollydev
Tags: #polly
Differential Revision: https://reviews.llvm.org/D33985
llvm-svn: 307173
Polly did not use PPCG's live range reordering feature. Teach
PPCGCodeGeneration to use this.
Documentation on this is sparse, so much of the code is conservative.
We currently kill all phi nodes in a Scop by appending them to the
must_kill map we pass to PPCG. I do not have a proof of correctness,
but it seems to be intuitively correct.
We also do not handle `array_order`, which, quoting PPCG, is:
PPCG/gpu.h: "Order dependences on non-scalars."
It seems to consist of RAW dependences between arrays. We need to
pass this information for more complex privatization cases.
Differential Revision: https://reviews.llvm.org/D34941
llvm-svn: 307163
Summary:
Introduce a "hybrid" `-polly-target` option to optimise code for either the GPU or CPU.
When this target is selected, PPCGCodeGeneration will attempt first to optimise a Scop. If the Scop isn't modified, it is then sent to the passes that form the CPU pipeline, i.e. IslScheduleOptimizerPass, IslAstInfoWrapperPass and CodeGeneration.
In case the Scop is modified, it is marked to be skipped by the subsequent CPU optimisation passes.
Reviewers: grosser, Meinersbur, bollu
Reviewed By: grosser
Subscribers: kbarton, nemanjai, pollydev
Tags: #polly
Differential Revision: https://reviews.llvm.org/D34054
llvm-svn: 306863
This patch aims to implement the option of allocating new arrays created
by polly on heap instead of stack. To enable this option, a key named
'allocation' must be written in the imported json file with the value
'heap'.
We need such a feature because in a next iteration, we will implement a
mechanism of maximal static expansion which will need a way to allocate
arrays on heap. Indeed, the expansion is very costly in terms of memory
and doing the allocation on stack is not worth considering.
The malloc and the free are added respectively at polly.start and
polly.exiting such that there is no use-after-free (for instance in case
of Scop in a loop) and such that all memory cells allocated with a
malloc are free'd when we don't need them anymore.
We also add :
- In the class ScopArrayInfo, we add a boolean as member called IsOnHeap
which represents the fact that the array in allocated on heap or not.
- A new branch in the method allocateNewArrays in the ISLNodeBuilder for
the case of heap allocation. allocateNewArrays now takes a BBPair
containing polly.start and polly.exiting. allocateNewArrays takes this
two blocks and add the malloc and free calls respectively to
polly.start and polly.exiting.
- As IntPtrTy for the malloc call, we use the DataLayout one.
To do that, we have modified :
- createScopArrayInfo and getOrCreateScopArrayInfo such that it returns
a non-const SAI, in order to be able to call setIsOnHeap in the
JSONImporter.
- executeScopConditionnaly such that it return both start block and end
block of the scop, because we need this two blocs to be able to add
the malloc and the free calls at the right position.
Differential Revision: https://reviews.llvm.org/D33688
llvm-svn: 306540
Before we would 'guess' the correct location for the MergeBlock
that got introduced when executing a Scop conditionally. This
implicitly depends on the situation that at this point during
CodeGen there will be nothing between polly.start and polly.exiting.
With this commit we explicitly state that we want the block that
directly follows polly.exiting.
llvm-svn: 306398
- In D33414, if any function call was found within a kernel, we would bail out.
- This is an over-approximation. This patch changes this by allowing the
`llvm.sqrt.*` family of intrinsics.
- This introduces an additional step when creating a separate llvm::Module
for a kernel (GPUModule). We now copy function declarations from the
original module to new module.
- We also populate IslNodeBuilder::ValueMap so it replaces the function
references to the old module to the ones in the new module
(GPUModule).
Differential Revision: https://reviews.llvm.org/D34145
llvm-svn: 306284
This commit returns both the start and the exit block that are created
by executeScopConditionally.
In a future commit we will make use of the exit block. Before we would
have to use the implicit property that there won't be any code generated
between polly.start and polly.exiting at the time of use to find the
correct block ('polly.exiting').
All usage location are semantically unchanged.
llvm-svn: 306283
The condition that disallowed code generation in PPCGCodeGeneration with
invariant loads is not required. I haven't been able to construct a
counterexample where this generates invalid code.
Differential Revision: https://reviews.llvm.org/D34604
llvm-svn: 306245
Ensure that all array base pointers are assigned before generating
aliasing metadata by allocating new arrays beforehand.
Before this patch, getBasePtr() returned nullptr for new arrays because
the arrays were created at a later point. Nullptr did not match to any
array after the created array base pointers have been assigned and when
the loads/stores are generated.
llvm-svn: 305675
In `PPCGCodeGeneration`, we try to take the references of every `Value`
that is used within a Scop to offload to the kernel. This occurs in
`GPUNodeBuilder::createLaunchParameters`.
This breaks if one of the values is a function pointer, since one of
these cases will trigger:
1. We try to to take the references of an intrinsic function, and this
breaks at `verifyModule`, since it is illegal to take the reference of
an intrinsic.
2. We manage to take the reference to a function, but this fails at
`verifyModule` since the function will not be present in the module that
is created in the kernel.
3. Even if `verifyModule` succeeds (which should not occur), we would
then try to call a *host function* from the *device*, which is
illegal runtime behaviour.
So, we disable this entire range of possibilities by simply not allowing
function references within a `Scop` which corresponds to a kernel.
However, note that this is too conservative. We *can* allow intrinsics
within kernels if the backend can lower the intrinsic correctly. For
example, an intrinsic like `llvm.powi.*` can actually be lowered by the `NVPTX`
backend.
We will now gradually whitelist intrinsics which are known to be safe.
Differential Revision: https://reviews.llvm.org/D33414
llvm-svn: 305185
Summary:
The RegionGenerator traditionally kept a BlockMap that mapped from original
basic blocks to newly generated basic blocks. With the introduction of partial
writes such a 1:1 mapping is not possible any more, as a single basic block
can be code generated into multiple basic blocks. Hence, depending on the use
case we need to either use the first basic block or the last basic block.
This is intended to address the last four cases of incorrect code generation
in our AOSP buildbot and hopefully should turn it green.
Reviewers: Meinersbur, bollu, gareevroman, efriedma, huihuiz, sebpop, simbuerg
Reviewed By: Meinersbur
Subscribers: pollydev, llvm-commits
Tags: #polly
Differential Revision: https://reviews.llvm.org/D33767
llvm-svn: 304808
- Add a counter that is incremented once on exit from a scop.
- Test cases got split into two: one to test the cycles, and another one
to test trip counts.
- Sample output:
```name=sample-output.txt
scop function, entry block name, exit block name, total time, trip count
warmup, %entry.split, %polly.merge_new_and_old, 5180, 1
f, %entry.split, %polly.merge_new_and_old, 409944, 500
g, %entry.split, %polly.merge_new_and_old, 1226, 1
```
Differential Revision: https://reviews.llvm.org/D33822
llvm-svn: 304543
We should bail out if performance monitoring is not supported, since
we would have no information to print per-scop, and `FinalStartBB`,
`ReturnFromFinal` would be `nullptr`.
Assert that these are not `nullptr` if performance monitoring is supported.
llvm-svn: 304529
Previously, we would generate one performance counter for all scops.
Now, we generate both the old information, as well as a per-scop
performance counter to generate finer grained information.
This patch needed a way to generate a unique name for a `Scop`.
The start region, end region, and function name combined provides a
unique `Scop` name. So, `Scop` has a new public API to provide its start
and end region names.
Differential Revision: https://reviews.llvm.org/D33723
llvm-svn: 304528
For when statements do not contain all instructions of a BasicBlock
anymore, the block generator needs to go through the explicit list of
instructions it contains.
Contributed-by: Nandini Singhal <cs15mtech01004@iith.ac.in>
Differential Revision: https://reviews.llvm.org/D33653
llvm-svn: 304502
A partial write is a write where the domain of the values written is a subset of
the execution domain of the parent statement containing the write. Originally,
we directly checked this subset relation whereas it is indeed only important
that the subset relation holds for the parameter values that are known to be
valid in the execution context of the scop. We update our check to avoid the
unnecessary introduction of partial writes in situations where the write appears
to be partial without context information, but where context information allows
us to understand that a full write can be generated.
This change fixes (hides) a recent regression introduced in r303517, which broke
our AOSP builds. The part that is correctly fixed in this change is that we do
not any more unnecessarily generate a partial write. This is good performance
wise and, as we currently do not yet explicitly introduce partial writes in the
default configuration, this also hides possible bugs in the partial writes
implementation. The crashes that we have originally seen were caused by such
a bug, where partial writes were incorrectly generated in region statements. An
additional patch in a subsequent commit is needed to address this problem.
Reported-by: Reported-by: Eli Friedman <efriedma@codeaurora.org>
Differential Revision: https://reviews.llvm.org/D33759
llvm-svn: 304398
Summary: To move CG to the new PM I outlined the various helper that were previously members of the CG class into free static functions. The CG class itself I moved into a header, which is required because we need to include it in `RegisterPasses` eventually.
Reviewers: grosser, Meinersbur
Reviewed By: grosser
Subscribers: pollydev, llvm-commits, sanjoy
Tags: #polly
Differential Revision: https://reviews.llvm.org/D33423
llvm-svn: 303624
Summary: This patch ports IslAst to the new PM. The change is mostly straightforward. The only major modification required is making IslAst move-only, to correctly manage the isl resources it owns.
Reviewers: grosser, Meinersbur
Reviewed By: grosser
Subscribers: nemanjai, pollydev, llvm-commits
Tags: #polly
Differential Revision: https://reviews.llvm.org/D33422
llvm-svn: 303622
The SCEVs of loops surrounding the escape users of a merge blocks are
forgotten, so that loop trip counts based on old values can be revoked.
This fixes llvm.org//PR32536
Contributed-by: Baranidharan Mohan <mbdharan@gmail.com>
Differential Revision: https://reviews.llvm.org/D33195
llvm-svn: 303561
Allow the BlockGenerator to generate memory writes that are not defined
over the complete statement domain, but only over a subset of it. It
generates a condition that evaluates to 1 if executing the subdomain,
and only then execute the access.
Only write accesses are supported. Read accesses would require a PHINode
which has a value if the access is not executed.
Partial write makes DeLICM able to apply mappings that are not defined
over the entire domain (for instance, a branch that leaves a loop with
a PHINode in its header; a MemoryKind::PHI write when leaving is never
read by its PHI read).
Differential Revision: https://reviews.llvm.org/D33255
llvm-svn: 303517
- We use the outermost dimension of arrays since we need this
information to generate GPU transfers.
- In general, if we do not know the outermost dimension of the array
(because the indexing expression is non-affine, for example) then we
simply cannot generate transfer code.
- However, for Fortran arrays, we can use the Fortran array
representation which stores the dimensions of all arrays.
- This patch uses the Fortran array representation to generate code that
computes the outermost dimension size.
Differential Revision: https://reviews.llvm.org/D32967
llvm-svn: 303429
Summary:
Implements PR889
Removing the virtual table pointer from Value saves 1% of RSS when doing
LTO of llc on Linux. The impact on time was positive, but too noisy to
conclusively say that performance improved. Here is a link to the
spreadsheet with the original data:
https://docs.google.com/spreadsheets/d/1F4FHir0qYnV0MEp2sYYp_BuvnJgWlWPhWOwZ6LbW7W4/edit?usp=sharing
This change makes it invalid to directly delete a Value, User, or
Instruction pointer. Instead, such code can be rewritten to a null check
and a call Value::deleteValue(). Value objects tend to have their
lifetimes managed through iplist, so for the most part, this isn't a big
deal. However, there are some places where LLVM deletes values, and
those places had to be migrated to deleteValue. I have also created
llvm::unique_value, which has a custom deleter, so it can be used in
place of std::unique_ptr<Value>.
I had to add the "DerivedUser" Deleter escape hatch for MemorySSA, which
derives from User outside of lib/IR. Code in IR cannot include MemorySSA
headers or call the MemoryAccess object destructors without introducing
a circular dependency, so we need some level of indirection.
Unfortunately, no class derived from User may have any virtual methods,
because adding a virtual method would break User::getHungOffOperands(),
which assumes that it can find the use list immediately prior to the
User object. I've added a static_assert to the appropriate OperandTraits
templates to help people avoid this trap.
Reviewers: chandlerc, mehdi_amini, pete, dberlin, george.burgess.iv
Reviewed By: chandlerc
Subscribers: krytarowski, eraman, george.burgess.iv, mzolotukhin, Prazek, nlewycky, hans, inglorion, pcc, tejohnson, dberlin, llvm-commits
Differential Revision: https://reviews.llvm.org/D31261
llvm-svn: 303362
Summary: This is a proof of concept of how to port polly-passes to the new PassManager architecture. This approach works ootb for Function-Passes, but might not be directly applicable to Scop/Region-Passes. While we could just run the Analyses/Transforms over functions instead, we'd surrender the nice pipelining behaviour we have now.
Reviewers: Meinersbur, grosser
Reviewed By: grosser
Subscribers: pollydev, sanjoy, nemanjai, llvm-commits
Tags: #polly
Differential Revision: https://reviews.llvm.org/D31459
llvm-svn: 302902
Today Polly generates induction variable in this way:
polly.indvar = phi 0, polly.indvar.next
...
polly.indvar.next = polly.indvar + stide
polly.loop_cond = predicate polly.indvar, (UB - stride)
Instead of:
polly.indvar = phi 0, polly.indvar.next
...
polly.indvar.next = polly.indvar + stide
polly.loop_cond = predicate polly.indvar.next, UB
The way Polly generate induction variable cause some problem in the indvar simplify pass.
This patch make polly generate the later form, by assuming the induction variable never overflow
Differential Revision: https://reviews.llvm.org/D33089
llvm-svn: 302866
Summary:
In case two arrays share base pointers in the same invariant load equivalence
class, we canonicalize all memory accesses to the first of these arrays
(according to their order in the equivalence class).
This enables us to optimize kernels such as boost::ublas by ensuring that
different references to the C array are interpreted as accesses to the same
array. Before this change the runtime alias check for ublas would fail, as it
would assume models of the C array with differing (but identically valued) base
pointers would reference distinct regions of memory whereas the referenced
memory regions were indeed identical.
As part of this change we remove most of the MemoryAccess::get*BaseAddr
interface. We removed already all references to get*BaseAddr in previous
commits to ensure that no code relies on matching base pointers between
memory accesses and scop arrays -- except for three remaining uses where we
need the original base pointer. We document for these situations that
MemoryAccess::getOriginalBaseAddr may return a base pointer that is distinct
to the base pointer of the scop array referenced by this memory access.
Reviewers: sebpop, Meinersbur, zinob, gareevroman, pollydev, huihuiz, efriedma, jdoerfert
Reviewed By: Meinersbur
Subscribers: etherzhhb
Tags: #polly
Differential Revision: https://reviews.llvm.org/D28518
llvm-svn: 302636
Summary: PPCGCodeGeneration now attaches the size of the kernel launch parameters at the end of the parameter list. For the existing CUDA Runtime, this gets ignored, but the OpenCL Runtime knows to check for kernel-argument size at the end of the parameter list. (The resulting parameters list is twice as long. This has been accounted for in the corresponding test cases).
Reviewers: grosser, Meinersbur, bollu
Reviewed By: bollu
Subscribers: nemanjai, yaxunl, Anastasia, pollydev, llvm-commits
Tags: #polly
Differential Revision: https://reviews.llvm.org/D32961
llvm-svn: 302515
Summary:
When compiling for GPU, one can now choose to compile for OpenCL or CUDA,
with the corresponding polly-gpu-runtime flag (libopencl / libcudart). The
GPURuntime library (GPUJIT) has been extended with the OpenCL Runtime library
for that purpose, correctly choosing the corresponding library calls to the
option chosen when compiling (via different initialization calls).
Additionally, a specific GPU Target architecture can now be chosen with -polly-gpu-arch (only nvptx64 implemented thus far).
Reviewers: grosser, bollu, Meinersbur, etherzhhb, singam-sanjay
Reviewed By: grosser, Meinersbur
Subscribers: singam-sanjay, llvm-commits, pollydev, nemanjai, mgorny, yaxunl, Anastasia
Tags: #polly
Differential Revision: https://reviews.llvm.org/D32431
llvm-svn: 302379
This reverts commit 17a84e414adb51ee375d14836d4c2a817b191933.
Patches should have been submitted in the order of:
1. D32852
2. D32854
3. D32431
I mistakenly pushed D32431(3) first. Reverting to push in the correct
order.
llvm-svn: 302217
Summary:
When compiling for GPU, one can now choose to compile for OpenCL or CUDA,
with the corresponding polly-gpu-runtime flag (libopencl / libcudart). The
GPURuntime library (GPUJIT) has been extended with the OpenCL Runtime library
for that purpose, correctly choosing the corresponding library calls to the
option chosen when compiling (via different initialization calls).
Additionally, a specific GPU Target architecture can now be chosen with -polly-gpu-arch (only nvptx64 implemented thus far).
Reviewers: grosser, bollu, Meinersbur, etherzhhb, singam-sanjay
Reviewed By: grosser, Meinersbur
Subscribers: singam-sanjay, llvm-commits, pollydev, nemanjai, mgorny, yaxunl, Anastasia
Tags: #polly
Differential Revision: https://reviews.llvm.org/D32431
llvm-svn: 302215
If a ScopStmt references a (scalar) value, there are multiple
possibilities where this value can come. The decision about what kind of
use it is must be handled consistently at different places, which can be
error-prone. VirtualUse is meant to centralize the handling of the
different types of value uses.
This patch makes ScopBuilder and CodeGeneration use VirtualUse. This
already helps to show inconsistencies with the value handling. In order
to keep this patch NFC, exceptions to the general rules are added.
These might be fixed later if they turn to problems. Overall, this
should result in fewer post-codegen IR-verification errors, but instead
assertion failures in `getNewValue` that are closer to the actual error.
Differential Revision: https://reviews.llvm.org/D32667
llvm-svn: 302157
As has been reported in the previous commit, codegen verification can result in
quadratic compile time increases for large functions with many scops. This is
certainly not something we would like to have in the Polly default
configuration. Hence, we disable codegen verification by default -- also to see
if this resolves some of the compilation timeouts we currently see on the AOSP
buildbots. We still leave this feature in Polly as it has shown _very_ useful
for debugging. In fact, we may want to have a discussion if we can bring this
feature back in a way that does not impact compilation time so much.
Thanks to Eli Friedman <efriedma@codeaurora.org> for reporting this issue and
for providing the test case in the previous commit (where I forgot to
acknowledge him).
llvm-svn: 301670
Before this change, we always tried to verify the function and printed
verification errors, but just did not abort in case -polly-codegen-verify=false
was set and verification failed. As verification can become very cosly -- for
large functions with many scops we may verify the very same function very often
-- this can affect compile time very negatively. Hence, we respect the
-polly-codegen-verify flag with this check, ensuring that no verification is run
if -polly-codegen-verify=false.
This reduces code generation time from 26 seconds to 4 seconds on the test
case below with -polly-codegen-verify=false:
struct X { int x; };
void a();
#define SIG (int x, X **y, X **z)
typedef void (*fn)SIG;
#define FN { for (int i = 0; i < x; ++i) { (*y)[i].x += (*z)[i].x; } a(); }
#define FN5 FN FN FN FN FN
#define FN25 FN5 FN5 FN5 FN5
#define FN125 FN25 FN25 FN25 FN25 FN25
#define FN250 FN125 FN125
#define FN1250 FN250 FN250 FN250 FN250 FN250
void x SIG { FN1250 }
llvm-svn: 301669
generation.
This needs changes to GPURuntime to expose synchronization between host
and device.
1. Needs better function naming, I want a better name than
"getOrCreateManagedDeviceArray"
2. DeviceAllocations is used by both the managed memory and the
non-managed memory path. This exploits the fact that the two code paths
are never run together. I'm not sure if this is the best design decision
Reviewed by: PhilippSchaad
Tags: #polly
Differential Revision: https://reviews.llvm.org/D32215
llvm-svn: 301640
Added a small change to the way pointer arguments are set in the kernel
code generation. The way the pointer is retrieved now, specifically requests
global address space to be annotated. This is necessary, if the IR should be
run through NVPTX to generate OpenCL compatible PTX.
The changes do not affect the PTX Strings generated for the CUDA target
(nvptx64-nvidia-cuda), but are necessary for OpenCL (nvptx64-nvidia-nvcl).
Additionally, the data layout has been updated to what the NVPTX Backend requests/recommends.
Contributed-by: Philipp Schaad
Reviewers: Meinersbur, grosser, bollu
Reviewed By: grosser, bollu
Subscribers: jlebar, pollydev, llvm-commits, nemanjai, yaxunl, Anastasia
Tags: #polly
Differential Revision: https://reviews.llvm.org/D32215
llvm-svn: 301299
The current StackColoring algorithm does not correctly handle the
situation when some, but not all paths from a BB to the entry node
cross a llvm.lifetime.start. According to an interpretation of the
language reference at
http://llvm.org/docs/LangRef.html#llvm-lifetime-start-intrinsic
this might be correct, but it would cost too much effort to handle
in StackColoring.
To be on the safe side, remove all lifetime markers even in the original
code version (they have never been copied to the optimized version)
to ensure that no path to the entry block will cross a
llvm.lifetime.start.
The same principle applies to paths the a function return and the
llvm.lifetime.end marker, so we remove them as well.
This fixes llvm.org/PR32251.
Also see the discussion at
http://lists.llvm.org/pipermail/llvm-dev/2017-March/111551.html
llvm-svn: 299585
Summary:
A couple of the utilities used to analyze or build IR make explicit use of the legacy PM on their interface, to access analysis results. This patch removes the legacy PM from the interface, and just passes the required results directly.
This shouldn't introduce any function changes, although the API technically allowed to obtain two different analysis results before, one passed by reference and one through the PM. I don't believe that was ever intended, however.
Reviewers: grosser, Meinersbur
Reviewed By: grosser
Subscribers: nemanjai, pollydev, llvm-commits
Tags: #polly
Differential Revision: https://reviews.llvm.org/D31653
llvm-svn: 299423
Instead of creating the declaration ourselves, we obtain it directly from the
LLVM intrinsic definitions. This addresses a post-review comment for r299359.
Suggested-by: Hongzing Zheng <etherzhhb@gmail.com>
llvm-svn: 299360
Add support for -polly-codegen-perf-monitoring. When performance monitoring
is enabled, we emit performance monitoring code during code generation that
prints after program exit statistics about the total number of cycles executed
as well as the number of cycles spent in scops. This gives an estimate on how
useful polyhedral optimizations might be for a given program.
Example output:
Polly runtime information
-------------------------
Total: 783110081637
Scops: 663718949365
In the future, we might also add functionality to measure how much time is spent
in optimized scops and how many cycles are spent in the fallback code.
Reviewers: bollu,sebpop
Tags: #polly
Differential Revision: https://reviews.llvm.org/D31599
llvm-svn: 299359
No-alias metadata grows quadratic in the size of arrays involved, which can
become very costly for large programs. This commit bounds the number of arrays
for which we construct no-alias information to ten. This is conservatively
correct, as we just provide less information to LLVM and speeds up the compile
time of one of my internal test cases from 'does-not-terminate' to
'finishes-in-less-than-a-minute'. In the future we might try to be more clever
here, but this change should provide a good baseline.
llvm-svn: 299352
Provide an common way for testing if a statement contains something
for region and block statements. First user is
RegionGenerator::addOperandToPHI.
Suggested-by: Tobias Grosser <tobias@grosser.es>
llvm-svn: 298617
Introduce another level of alias metadata to distinguish the individual
non-aliasing accesses that have inter iteration alias-free base pointers
marked with "Inter iteration alias-free" mark nodes. It can be used to,
for example, distinguish different stores (loads) produced by unrolling of
the innermost loops and, subsequently, sink (hoist) them by LICM.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D30606
llvm-svn: 298510
Map the new load to the base pointer of the invariant load hoisted load
to be able to find the alias information for it.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D30605
llvm-svn: 298507
Siddharth Bhat