The pattern matching optimization of Polly detects and optimizes dense general
matrix-matrix multiplication. The generated code is close to high performance
implementations of matrix-matrix multiplications, which are contained in
manually tuned libraries. The described pattern matching optimization is
a particular case of tensor contraction optimization, which was
introduced in [1].
This patch generalizes the pattern matching to the case of tensor contractions
using the form of data dependencies and memory accesses produced by tensor
contractions [1].
Optimization of tensor contractions will be added in the next patch. Following
the ideas introduced in [2], it will logically represent tensor contraction
operands as matrix multiplication operands and use an approach for
optimization of matrix-matrix multiplications.
[1] - Gareev R., Grosser T., Kruse M. High-Performance Generalized Tensor
Operations: A Compiler-Oriented Approach // ACM Transactions on
Architecture and Code Optimization (TACO). 2018. Vol. 15, no. 3.
P. 34:1–34:27. DOI: 10.1145/3235029.
[2] - Matthews D. High-Performance Tensor Contraction without BLAS // SIAM
Journal on Scientific Computing. 2018. Vol. 40, no. 1. P. C 1—C 24.
DOI: 110.1137/16m108968x.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D114336
The copy statements inserted by the matrix-multiplication optimization
introduce new dependencies between the copy statements and other
statements. As a result, the DependenceInfo must be recomputed.
Not recomputing them caused IslAstInfo to deduce that some loops are
parallel but cause race conditions when accessing the packed arrays.
As a result, matrix-matrix multiplication currently cannot be
parallelized.
Also see discussion at https://reviews.llvm.org/D125202
This make is obivious that a class was not intended to be derived from.
NPM analysis pass can unfortunately not marked as final because they are
derived from a llvm::Checker<T> template internally by the NPM.
Also normalize the use of classes/structs
* NPM passes are structs
* Legacy passes are classes
* structs that have methods and are not a visitor pattern are classes
* structs have public inheritance by default, remove "public" keyword
* Use typedef'ed type instead of inline forward declaration
This is part of an effort to reduce the differences between the custom C++ bindings used right now by polly in lib/External/isl/include/isl/isl-noxceptions.h and the official isl C++ interface.
In the official interface the type `isl::size` cannot be casted to an unsigned without previously having checked if it contains a valid value with the function `isl::size::is_error()`.
For this reason two helping functions have been added:
- `IslAssert`: assert that no errors are present in debug builds and just disables the mandatory error check in non-debug builds
- `unisgnedFromIslSIze`: cast the `isl::size` object to `unsigned`
Changes made:
- Add the functions `IslAssert` and `unsignedFromIslSize`
- Add the utility function `rangeIslSize()`
- Retype `MaxDisjunctsInDomain` from `int` to `unsigned`
- Retype `RunTimeChecksMaxAccessDisjuncts` from `int` to `unsigned`
- Retype `MaxDimensionsInAccessRange` from `int` to `unsigned`
- Replaced some usages of `isl_size` to `unsigned` since we aim not to use `isl_size` anymore
- `isl-noexceptions.h` has been generated by e704f73c88
No functional change intended.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D113101
multiplication
The following code modifies elements of the array D.
for (i = 0; i < _PB_NI; i++)
for (j = 0; j < _PB_NJ; j++)
{
for (k = 0; k < _PB_NK; k++)
{
double Mul = A[i][k] * B[k][j];
D[i][j][k] += Mul;
C[i][j] += Mul;
}
}
Nevertheless, the code is recognised as a matrix-matrix multiplication, since
the second and third dimensions of D are accessed with non-zero strides.
This fixes the typo, which was made during the translation to C++ bindings
(https://reviews.llvm.org/D35845).
Reviewed By: Michael Kruse <llvm@meinersbur.de>
Differential Revision: https://reviews.llvm.org/D110491
This metadata was intended to mark all accesses within an iteration to be pairwise non-aliasing, in this case because every memory of a base pointer is touched (read or write) at most once. This is typical for 'sweeps' over all data. The stated motivation from D30606 is to ensure that unrolled iterations are considered non-aliasing.
Rhe implemention had multiple issues:
* The structure of the noalias metadata was malformed. D110026 added check in the verifier for this metadata, and the tests were failing since then.
* This is not true for the outer loops of the BLIS matrix multiplication, where it was being inserted. Each element of A, B, C is accessed multiple times, as often as the loop not used as an index is iterating.
* Scopes were added to SecondLevelOtherAliasScopeList (used for the !noalias scop list) on-the-fly when another SCEV was seen. This meant that previously visited instructions would not be updated with alias scopes that are only seen later, missing out those SCEVs they should not be aliasing with.
* Since the !noalias scope list would ideally consists of all other SCEV for this base pointer, we might run quickly into scalability issues. Especially after unrolling there would probably at least once SCEV per instruction and unroll instance.
* The inter-iteration noalias base pointer was not removed after leaving the loop marked with it, effectively marking everything after it to noalias as well.
A solution I considered was to mark each instruction as non-aliasing with its own scope. The instruction itself would obviously alias itself, but such construction might also be considered invalid. Duplicating the instruction (e.g. due to speculation) would mark the instruction non-aliasing with its clone. I don't want to go into this territory, especially since the original motivation of determining unrolled instances as noalias based on SCEV is the what scev-aa does as well.
This effectively reverts D30606 and D35761.
This is part of an effort to reduce the differences between the custom C++ bindings used right now by polly in `lib/External/isl/include/isl/isl-noxceptions.h` and the official isl C++ interface.
With this commit we are moving from the `polly-generator` branch to the `new-polly-generator` branch that is more mantainable and is based on the official C++ interface `cpp-checked.h`.
Changes made:
- There are now many sublcasses for `isl::ast_node` representing different isl types. Use `isl::ast_node_for`, `isl::ast_node_user`, `isl::ast_node_block` and `isl::ast_node_mark` where needed.
- There are now many sublcasses for `isl::schedule_node` representing different isl types. Use `isl::schedule_node_mark`, `isl::schedule_node_extension`, `isl::schedule_node_band` and `isl::schedule_node_filter` where needed.
- Replace the `isl::*::dump` with `dumpIslObj` since the isl dump method is not exposed in the C++ interface.
- `isl::schedule_node::get_child` has been renamed to `isl::schedule_node::child`
- `isl::pw_multi_aff::get_pw_aff` has been renamed to `isl::pw_multi_aff::at`
- The constructor `isl::union_map(isl::union_pw_multi_aff)` has been replaced with the static method `isl::union_map::from()`
- Replace usages of `isl::val::add_ui` with `isl::val::add`
- `isl::union_set_list::alloc` is now a constructor
- All the `isl_size` values are now wrapped inside the class `isl::size` use `isl::size::release` to get the internal `isl_size` value where needed.
- `isl-noexceptions.h` has been generated by 73f5ed1f4d
No functional change intended.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D107225
This is part of an effort to reduce the differences between the custom C++ bindings used right now by polly in `lib/External/isl/include/isl/isl-noxceptions.h` and the official isl C++ interface.
Changes made:
- Use `isl::*::ctx()` instead of `isl::*::get_ctx()` (for example `isl::space::ctx()` instead of `isl::space::get_ctx()`)
- Add `isl::` namespace in front of isl types to avoid confusion (for example `isl::space::ctx` and `isl::ctx`
- `isl-noexceptions.h` has been generated by this b64e33c62d
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D105691
This is part of an effort to reduce the differences between the custom C++ bindings used right now by polly in `lib/External/isl/include/isl/isl-noxceptions.h` and the official isl C++ interface.
Changes made:
- Use `isl::set::tuple_dim` instead of `isl::set::dim` and `isl::set::n_dim`
- Use `isl::map::domain_tuple_dim` instead of `isl::map::dim`
- Use `isl::map::range_tuple_dim` instead of `isl::map::dim`
- isl-noexceptions.h has been generated by this 45576e1b42
Note that not all the usage of `isl::{set,map}::dim` where replaced
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D104994
This is part of an effort to reduce the differences between the custom C++ bindings used right now by polly in `lib/External/isl/include/isl/isl-noxceptions.h` and the official isl C++ interface.
Changes made:
- Removing explicit operator bool() from all the classes in the isl C++ bindings.
- Replace each call to operator bool() to method `is_null()`.
- isl-noexceptions.h has been generated by this 27396daac5
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D103976
Functions shared between generalized matrix-multiplication optimization
and other post-reschedule optimizations (tiling, prevect) are moved into
the schedule tree transformation utility ScheduleTreeTransform.
Fangrui Song