The patch extends the runner utils by verification methods that compare two memrefs. The methods compare the content of the two memrefs and print success if the data is identical up to a small numerical error. The methods are meant to simplify the development of integration tests that compare the results against a reference implementation (cf. the updates to the linalg matmul integration tests).
Originally landed in 5fa893c (https://reviews.llvm.org/D96326) and reverted in dd719fd due to a Windows build failure.
Changes:
- Remove the max function that requires the "algorithm" header on Windows
- Eliminate the truncation warning in the float specialization of verifyElem by using a float constant
Reviewed By: Kayjukh
Differential Revision: https://reviews.llvm.org/D96593
The dimension order of a filter in tensorflow is
[filter_height, filter_width, in_channels, out_channels], which is different
from current definition. The current definition follows TOSA spec. Add TF
version conv ops to .tc, so we do not have to insert a transpose op around a
conv op.
Reviewed By: antiagainst
Differential Revision: https://reviews.llvm.org/D96038
The patch extends the runner utils by verification methods that compare two memrefs. The methods compare the content of the two memrefs and print success if the data is identical up to a small numerical error. The methods are meant to simplify the development of integration tests that for example compare optimized and unoptimized code paths (cf. the updates to the linalg matmul integration tests).
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D96326
Historically, Linalg To LLVM conversion subsumed numerous other conversions,
including (affine) loop lowerings to CFG and conversions from the Standard and
Vector dialects to the LLVM dialect. This was due to the insufficient support
for partial conversions in the infrastructure that essentially required
conversions that involve type change (in this case, !linalg.range to
!llvm.struct) to be performed in a single conversion sweep. This is no longer
the case so remove the subsumed conversions and run them as separate passes
when necessary.
Depends On D95317
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D96008
Historically, the Vector to LLVM dialect conversion subsumed the Standard to
LLVM dialect conversion patterns. This was necessary because the conversion
infrastructure did not have sufficient support for reconciling type
conversions. This support is now available. Only keep the patterns related to
the Vector dialect in the Vector to LLVM conversion and require type casts
operations to be inserted if necessary. These casts will be removed by
following conversions if possible. Update integration tests to also run the
Standard to LLVM conversion.
There is a significant amount of test churn, which is due to (a) unnecessarily
strict tests in VectorToLLVM and (b) many patterns actually targeting Standard
dialect ops instead of LLVM dialect ops leading to tests actually exercising a
Vector->Standard->LLVM conversion. This churn is a good illustration of the
reason to make the conversion partial: now the tests only check the code in the
Vector to LLVM conversion and will not be randomly broken by changes in
Standard to LLVM conversion.
Arguably, it may be possible to extract Vector to Standard patterns into a
separate pass, but given the ongoing splitting of the Standard dialect, such
pass will be short-lived and will require further refactoring.
Depends On D95626
Reviewed By: nicolasvasilache, aartbik
Differential Revision: https://reviews.llvm.org/D95685
This revision defines a Linalg contraction in general terms:
1. Has 2 input and 1 output shapes.
2. Has at least one reduction dimension.
3. Has only projected permutation indexing maps.
4. its body computes `u5(u1(c) + u2(u3(a) * u4(b)))` on some field
(AddOpType, MulOpType), where u1, u2, u3, u4 and u5 represent scalar unary
operations that may change the type (e.g. for mixed-precision).
As a consequence, when vectorization of such an op occurs, the only special
behavior is that the (unique) MulOpType is vectorized into a
`vector.contract`. All other ops are handled in a generic fashion.
In the future, we may wish to allow more input arguments and elementwise and
constant operations that do not involve the reduction dimension(s).
A test is added to demonstrate the proper vectorization of matmul_i8_i8_i32.
Differential revision: https://reviews.llvm.org/D95939
This revision adds a layer of SFINAE to the composable codegen strategy so it does
not have to require statically defined ops but instead can also be used with OpInterfaces, Operation* and an op name string.
A linalg.matmul_i8_i8_i32 is added to the .tc spec to demonstrate how all this works end to end.
Differential Revision: https://reviews.llvm.org/D95600
This revision improves the usage of the codegen strategy by adding a few flags that
make it easier to control for the CLI.
Usage of ModuleOp is replaced by FuncOp as this created issues in multi-threaded mode.
A simple benchmarking capability is added for linalg.matmul as well as linalg.matmul_column_major.
This latter op is also added to linalg.
Now obsolete linalg integration tests that also take too long are deleted.
Correctness checks are still missing at this point.
Differential revision: https://reviews.llvm.org/D95531
OffsetSizeAndStrideOpInterface now have the ability to specify only a leading subset of
offset, sizes, strides operands/attributes.
The size of that leading subset must be limited by the corresponding entry in `getArrayAttrMaxRanks` to avoid overflows.
Missing trailing dimensions are assumed to span the whole range (i.e. [0 .. dim)).
This brings more natural semantics to slice-like op on top of subview and is a simplifies to removing all uses of SliceOp in dependent projects.
Differential revision: https://reviews.llvm.org/D95441
Depends On D95000
Move async.execute outlining and async -> async.runtime lowering into the separate Async transformation pass
Reviewed By: mehdi_amini
Differential Revision: https://reviews.llvm.org/D95311
In the overwhelmingly common case, enum attribute case strings represent valid identifiers in MLIR syntax. This revision updates the format generator to format as a keyword in these cases, removing the need to wrap values in a string. The parser still retains the ability to parse the string form, but the printer will use the keyword form when applicable.
Differential Revision: https://reviews.llvm.org/D94575
Continue the convergence between LLVM dialect and built-in types by using the
built-in vector type whenever possible, that is for fixed vectors of built-in
integers and built-in floats. LLVM dialect vector type is still in use for
pointers, less frequent floating point types that do not have a built-in
equivalent, and scalable vectors. However, the top-level `LLVMVectorType` class
has been removed in favor of free functions capable of inspecting both built-in
and LLVM dialect vector types: `LLVM::getVectorElementType`,
`LLVM::getNumVectorElements` and `LLVM::getFixedVectorType`. Additional work is
necessary to design an implemented the extensions to built-in types so as to
remove the `LLVMFixedVectorType` entirely.
Note that the default output format for the built-in vectors does not have
whitespace around the `x` separator, e.g., `vector<4xf32>` as opposed to the
LLVM dialect vector type format that does, e.g., `!llvm.vec<4 x fp128>`. This
required changing the FileCheck patterns in several tests.
Reviewed By: mehdi_amini, silvas
Differential Revision: https://reviews.llvm.org/D94405
This change makes the scatter/gather syntax more consistent with
the syntax of all the other memory operations in the Vector dialect
(order of types, use of [] for index, etc.). This will make the MLIR
code easier to read. In addition, the pass_thru parameter of the
gather has been made mandatory (there is very little benefit in
using the implicit "undefined" values).
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D94352
Adding the ability to index the base address brings these operations closer
to the transfer read and write semantics (with lowering advantages), ensures
more consistent use in vector MLIR code (easier to read), and reduces the
amount of code duplication to lower memrefs into base addresses considerably
(making codegen less error-prone).
Reviewed By: ThomasRaoux
Differential Revision: https://reviews.llvm.org/D94278
Continue the convergence between LLVM dialect and built-in types by replacing
the bfloat, half, float and double LLVM dialect types with their built-in
counterparts. At the API level, this is a direct replacement. At the syntax
level, we change the keywords to `bf16`, `f16`, `f32` and `f64`, respectively,
to be compatible with the built-in type syntax. The old keywords can still be
parsed but produce a deprecation warning and will be eventually removed.
Depends On D94178
Reviewed By: mehdi_amini, silvas, antiagainst
Differential Revision: https://reviews.llvm.org/D94179
The LLVM dialect type system has been closed until now, i.e. did not support
types from other dialects inside containers. While this has had obvious
benefits of deriving from a common base class, it has led to some simple types
being almost identical with the built-in types, namely integer and floating
point types. This in turn has led to a lot of larger-scale complexity: simple
types must still be converted, numerous operations that correspond to LLVM IR
intrinsics are replicated to produce versions operating on either LLVM dialect
or built-in types leading to quasi-duplicate dialects, lowering to the LLVM
dialect is essentially required to be one-shot because of type conversion, etc.
In this light, it is reasonable to trade off some local complexity in the
internal implementation of LLVM dialect types for removing larger-scale system
complexity. Previous commits to the LLVM dialect type system have adapted the
API to support types from other dialects.
Replace LLVMIntegerType with the built-in IntegerType plus additional checks
that such types are signless (these are isolated in a utility function that
replaced `isa<LLVMType>` and in the parser). Temporarily keep the possibility
to parse `!llvm.i32` as a synonym for `i32`, but add a deprecation notice.
Reviewed By: mehdi_amini, silvas, antiagainst
Differential Revision: https://reviews.llvm.org/D94178
This revision drops init_tensor arguments from Linalg on tensors and instead uniformizes the output buffers and output tensors to be consistent.
This significantly simplifies the usage of Linalg on tensors and is a stepping stone for
its evolution towards a mixed tensor and shape abstraction discussed in https://llvm.discourse.group/t/linalg-and-shapes/2421/19.
Differential Revision: https://reviews.llvm.org/D93469
The InlineAsmOp mirrors the underlying LLVM semantics with a notable
exception: the embedded `asm_string` is not allowed to define or reference
any symbol or any global variable: only the operands of the op may be read,
written, or referenced.
Attempting to define or reference any symbol or any global behavior is
considered undefined behavior at this time.
The asm dialect syntax is currently specified with an integer (0 [default] for the "att dialect", 1 for the intel dialect) to circumvent the ODS limitation on string enums.
Translation to LLVM is provided and raises the fact that the asm constraints string must be well-formed with respect to in/out operands. No check is performed on the asm_string.
An InlineAsm instruction in LLVM is a special call operation to a function that is constructed on the fly.
It does not fit the current model of MLIR calls with symbols.
As a consequence, the current implementation constructs the function type in ModuleTranslation.cpp.
This should be refactored in the future.
The mlir-cpu-runner is augmented with the global initialization of the X86 asm parser to allow proper execution in JIT mode. Previously, only the X86 asm printer was initialized.
Differential revision: https://reviews.llvm.org/D92166
This change is required so that bufferization can properly identify
the linalg.yield as a terminator with an associated parent op.
Differential Revision: https://reviews.llvm.org/D92173
Depends On D89963
**Automatic reference counting algorithm outline:**
1. `ReturnLike` operations forward the reference counted values without
modifying the reference count.
2. Use liveness analysis to find blocks in the CFG where the lifetime of
reference counted values ends, and insert `drop_ref` operations after
the last use of the value.
3. Insert `add_ref` before the `async.execute` operation capturing the
value, and pairing `drop_ref` before the async body region terminator,
to release the captured reference counted value when execution
completes.
4. If the reference counted value is passed only to some of the block
successors, insert `drop_ref` operations in the beginning of the blocks
that do not have reference coutned value uses.
Reviewed By: silvas
Differential Revision: https://reviews.llvm.org/D90716
Depends On D89958
1. Adds `async.group`/`async.awaitall` to group together multiple async tokens/values
2. Rewrite scf.parallel operation into multiple concurrent async.execute operations over non overlapping subranges of the original loop.
Example:
```
scf.for (%i, %j) = (%lbi, %lbj) to (%ubi, %ubj) step (%si, %sj) {
"do_some_compute"(%i, %j): () -> ()
}
```
Converted to:
```
%c0 = constant 0 : index
%c1 = constant 1 : index
// Compute blocks sizes for each induction variable.
%num_blocks_i = ... : index
%num_blocks_j = ... : index
%block_size_i = ... : index
%block_size_j = ... : index
// Create an async group to track async execute ops.
%group = async.create_group
scf.for %bi = %c0 to %num_blocks_i step %c1 {
%block_start_i = ... : index
%block_end_i = ... : index
scf.for %bj = %c0 t0 %num_blocks_j step %c1 {
%block_start_j = ... : index
%block_end_j = ... : index
// Execute the body of original parallel operation for the current
// block.
%token = async.execute {
scf.for %i = %block_start_i to %block_end_i step %si {
scf.for %j = %block_start_j to %block_end_j step %sj {
"do_some_compute"(%i, %j): () -> ()
}
}
}
// Add produced async token to the group.
async.add_to_group %token, %group
}
}
// Await completion of all async.execute operations.
async.await_all %group
```
In this example outer loop launches inner block level loops as separate async
execute operations which will be executed concurrently.
At the end it waits for the completiom of all async execute operations.
Reviewed By: ftynse, mehdi_amini
Differential Revision: https://reviews.llvm.org/D89963
We lower them to a std.global_memref (uniqued by constant value) + a
std.get_global_memref to produce the corresponding memref value.
This allows removing Linalg's somewhat hacky lowering of tensor
constants, now that std properly supports this.
Differential Revision: https://reviews.llvm.org/D91306
It was incorrect in the presence of a tensor argument with multiple
uses.
The bufferization of subtensor_insert was writing into a converted
memref operand, but there is no guarantee that the converted memref for
that operand is safe to write into. In this case, the same converted
memref is written to in-place by the subtensor_insert bufferization,
violating the tensor-level semantics.
I left some comments in a TODO about ways forward on this. I will be
working actively on this problem in the coming days.
Differential Revision: https://reviews.llvm.org/D91371
This patch converts elementwise ops on tensors to linalg.generic ops
with the same elementwise op in the payload (except rewritten to
operate on scalars, obviously). This is a great form for later fusion to
clean up.
E.g.
```
// Compute: %arg0 + %arg1 - %arg2
func @f(%arg0: tensor<?xf32>, %arg1: tensor<?xf32>, %arg2: tensor<?xf32>) -> tensor<?xf32> {
%0 = addf %arg0, %arg1 : tensor<?xf32>
%1 = subf %0, %arg2 : tensor<?xf32>
return %1 : tensor<?xf32>
}
```
Running this through
`mlir-opt -convert-std-to-linalg -linalg-fusion-for-tensor-ops` we get:
```
func @f(%arg0: tensor<?xf32>, %arg1: tensor<?xf32>, %arg2: tensor<?xf32>) -> tensor<?xf32> {
%0 = linalg.generic {indexing_maps = [#map0, #map0, #map0, #map0], iterator_types = ["parallel"]} ins(%arg0, %arg1, %arg2 : tensor<?xf32>, tensor<?xf32>, tensor<?xf32>) {
^bb0(%arg3: f32, %arg4: f32, %arg5: f32): // no predecessors
%1 = addf %arg3, %arg4 : f32
%2 = subf %1, %arg5 : f32
linalg.yield %2 : f32
} -> tensor<?xf32>
return %0 : tensor<?xf32>
}
```
So the elementwise ops on tensors have nicely collapsed into a single
linalg.generic, which is the form we want for further transformations.
Differential Revision: https://reviews.llvm.org/D90354
The pass combines patterns of ExpandAtomic, ExpandMemRefReshape,
StdExpandDivs passes. The pass is meant to legalize STD for conversion to LLVM.
Differential Revision: https://reviews.llvm.org/D91082
Previously, linalg-bufferize was a "finalizing" bufferization pass (it
did a "full" conversion). This wasn't great because it couldn't be used
composably with other bufferization passes like std-bufferize and
scf-bufferize.
This patch makes linalg-bufferize a composable bufferization pass.
Notice that the integration tests are switched over to using a pipeline
of std-bufferize, linalg-bufferize, and (to finalize the conversion)
func-bufferize. It all "just works" together.
While doing this transition, I ran into a nasty bug in the 1-use special
case logic for forwarding init tensors. That logic, while
well-intentioned, was fundamentally flawed, because it assumed that if
the original tensor value had one use, then the converted memref could
be mutated in place. That assumption is wrong in many cases. For
example:
```
%0 = some_tensor : tensor<4xf32>
br ^bb0(%0, %0: tensor<4xf32>, tensor<4xf32>)
^bb0(%bbarg0: tensor<4xf32>, %bbarg1: tensor<4xf32>)
// %bbarg0 is an alias of %bbarg1. We cannot safely write
// to it without analyzing uses of %bbarg1.
linalg.generic ... init(%bbarg0) {...}
```
A similar example can happen in many scenarios with function arguments.
Even more sinister, if the converted memref is produced by a
`std.get_global_memref` of a constant global memref, then we might
attempt to write into read-only statically allocated storage! Not all
memrefs are writable!
Clearly, this 1-use check is not a local transformation that we can do
on the fly in this pattern, so I removed it.
The test is now drastically shorter and I basically rewrote the CHECK
lines from scratch because:
- the new composable linalg-bufferize just doesn't do as much, so there
is less to test
- a lot of the tests were related to the 1-use check, which is now gone,
so there is less to test
- the `-buffer-hoisting -buffer-deallocation` is no longer mixed in, so
the checks related to that had to be rewritten
Differential Revision: https://reviews.llvm.org/D90657
Fix semantic in the distribute integration test based on offline feedback. This
exposed a bug in block distribution, we need to make sure the id is multiplied
by the stride of the vector. Fix the transformation and unit test.
Differential Revision: https://reviews.llvm.org/D89291
TensorConstantOp bufferization currently uses the vector dialect to store constant data into memory.
Due to natural vector size and alignment properties, this is problematic with n>1-D vectors whose most minor dimension is not naturally aligned.
Instead, this revision linearizes the constant and introduces a linalg.reshape to go back to the desired shape.
Still this is still to be considered a workaround and a better longer term solution will probably involve `llvm.global`.
Differential Revision: https://reviews.llvm.org/D89311
The TensorConstantOp bufferize conversion pattern has a bug that
makes it incorrect in the case of vectors whose alignment is not
the natural alignment. Circumvent it temporarily by using a power of 2.
Differential Revision: https://reviews.llvm.org/D89265
This revision introduces support for buffer allocation for any named linalg op.
To avoid template instantiating many ops, a new ConversionPattern is created to capture the LinalgOp interface.
Some APIs are updated to remain consistent with MLIR style:
`OwningRewritePatternList * -> OwningRewritePatternList &`
`BufferAssignmentTypeConverter * -> BufferAssignmentTypeConverter &`
Differential revision: https://reviews.llvm.org/D89226
This revision also inserts an end-to-end test that lowers tensors to buffers all the way to executable code on CPU.
Differential revision: https://reviews.llvm.org/D88998
Current setup for conv op vectorization does not enable user to specify tile
sizes as well as dimensions for vectorization. In this commit we change that by
adding tile sizes as pass arguments. Every dimension with corresponding tile
size > 1 is automatically vectorized.
Differential Revision: https://reviews.llvm.org/D88533
Recently, restrictions on vector reductions were made more relaxed by
accepting any width signless integer and floating-point. This CL relaxes
the restriction even more by including unsigned and signed integers.
Reviewed By: bkramer
Differential Revision: https://reviews.llvm.org/D88442
(1) simplify integer printing logic by always using 64-bit print
(2) add index support (since vector<16xindex> is planned to be added)
(3) adjust naming convention print_x -> printX
Reviewed By: bkramer
Differential Revision: https://reviews.llvm.org/D88436
