Retry landing https://github.com/llvm/llvm-project/pull/153373
## Major changes from previous attempt
- remove the test in CAPI because no existing tests in CAPI deal with
sanitizer exemptions
- update `mlir/docs/Dialects/GPU.md` to reflect the new behavior: load
GPU binary in global ctors, instead of loading them at call site.
- skip the test on Aarch64 since we have an issue with initialization there
---------
Co-authored-by: Mehdi Amini <joker.eph@gmail.com>
This PR implements "automatic" location inference in the bindings. The
way it works is it walks the frame stack collecting source locations
(Python captures these in the frame itself). It is inspired by JAX's
[implementation](523ddcfbca/jax/_src/interpreters/mlir.py (L462))
but moves the frame stack traversal into the bindings for better
performance.
The system supports registering "included" and "excluded" filenames;
frames originating from functions in included filenames **will not** be
filtered and frames originating from functions in excluded filenames
**will** be filtered (in that order). This allows excluding all the
generated `*_ops_gen.py` files.
The system is also "toggleable" and off by default to save people who
have their own systems (such as JAX) from the added cost.
Note, the system stores the entire stacktrace (subject to
`locTracebackFramesLimit`) in the `Location` using specifically a
`CallSiteLoc`. This can be useful for profiling tools (flamegraphs
etc.).
Shoutout to the folks at JAX for coming up with a good system.
---------
Co-authored-by: Jacques Pienaar <jpienaar@google.com>
This PR melds https://github.com/llvm/llvm-project/pull/150137 and
https://github.com/llvm/llvm-project/pull/149414 *and* partially reverts
https://github.com/llvm/llvm-project/pull/124832.
The summary is the `PyDenseResourceElementsAttribute` finalizer/deleter
has/had two problems
1. wasn't threadsafe (can be called from a different thread than that
which currently holds the GIL)
2. can be called while the interpreter is "not initialized"
https://github.com/llvm/llvm-project/pull/124832 for some reason decides
to re-initialize the interpreter to avoid case 2 and runs afoul of the
fact that `Py_IsInitialized` can be false during the finalization of the
interpreter itself (e.g., at the end of a script).
I don't know why this decision was made (I missed the PR) but I believe
we should never be calling
[Py_Initialize](https://docs.python.org/3/c-api/init.html#c.Py_Initialize):
> In an application \*\*\*\***embedding Python**\*\*\*\*, this should be
called before using any other Python/C API functions
**but we aren't embedding Python**!
So therefore we will only be in case 2 when the interpreter is being
finalized and in that case we should just leak the buffer.
Note,
[lldb](548ca9e976/lldb/source/Plugins/ScriptInterpreter/Python/PythonDataObjects.cpp (L81-L93))
does a similar sort of thing for its finalizers.
Co-authored-by: Anton Korobeynikov <anton@korobeynikov.info>
Co-authored-by: Max Manainen <maximmanainen@gmail.com>
Co-authored-by: Anton Korobeynikov <anton@korobeynikov.info>
Co-authored-by: Max Manainen <maximmanainen@gmail.com>
In order to access and modify resetOffset and boundsCheck of
RawBufferCastOp in pythonic binding, we will have to use Attrs instead
of Property. This is because we do not have python binding support for
property yet. We should move back to property once we add pythonic
binding support for it.
---------
Signed-off-by: Stanley Winata <stanley.winata@amd.com>
This patch specializes the Python bindings for ForallOp and
InParallelOp, similar to the existing one for ForOp. These bindings
create the regions and blocks properly and expose some additional
helpers.
- Introduces a `large_resource_limit` parameter across Python bindings,
enabling the eliding of resource strings exceeding a specified character
limit during IR printing.
- To maintain backward compatibilty, when using `operation.print()` API,
if `large_resource_limit` is None and the `large_elements_limit` is set,
the later will be used to elide the resource string as well. This change
was introduced by https://github.com/llvm/llvm-project/pull/125738.
- For printing using pass manager, the `large_resource_limit` and
`large_elements_limit` are completely independent of each other.
This is mentioned as a "must" in
https://nanobind.readthedocs.io/en/latest/porting.html#type-casters when
implementing type casters.
While most of the existing `from_cpp` methods were already marked
noexcept, many of the `from_python` methods were not. This commit adds
the missing noexcept declarations to all type casters found in
`NanobindAdaptors.h`.
---------
Co-authored-by: Maksim Levental <maksim.levental@gmail.com>
A new transform op to represent that an attribute is to be chosen from a
set of alternatives and that this choice is made available as a
`!transform.param`. When a `selected` argument is provided, the op's
`apply()` semantics is that of just making this selected attribute
available as the result. When `selected` is not provided, `apply()`
complains that nothing has resolved the non-determinism that the op is
representing.
The motivation is to avoid having to negate `isDynamic*` checks, avoid
double negations, and allow for `ShapedType::isStaticDim` to be used in
ADT functions without having to wrap it in a lambda performing the
negation.
Also add the new functions to C and Python bindings.
RFC:
https://discourse.llvm.org/t/rfc-deprecate-linalg-elemwise-unary-and-elemwise-binary/87144
Remove the two operations and fix the tests by:
* Cleaning simple operation tests of the old ops
* Changing `linalg.elemwise_{u|bi}nary` with `linalg.{exp|add}` on
transform tests
* Changing some of the tests with `linalg.elementwise` instead, to
broaden test coverage
* Surgically removing the `elemwise_*` part in the Python tests
* Update MLIR transform examples (text and tests) with
`linalg.elementwise` instead
Nothing else changed.
Removes the Debug... prefix on the ops in tablegen, in line with pretty
much all other Transform-dialect extension ops. This means that the ops
in Python look like
`debug.EmitParamAsRemarkOp`/`debug.emit_param_as_remark` instead of
`debug.DebugEmitParamAsRemarkOp`/`debug.debug_emit_param_as_remark`.
Interpret an option value with multiple values, either in the form of an
`ArrayAttr` (either static or passed through a param) or as the multiple
attrs associated to a param, as a comma-separated list, i.e. as a
ListOption on a pass.
Improve ApplyRegisteredPassOp's support for taking options by taking
them as a dict (vs a list of string-valued key-value pairs).
Values of options are provided as either static attributes or as params
(which pass in attributes at interpreter runtime). In either case, the
keys and value attributes are converted to strings and a single
options-string, in the format used on the commandline, is constructed to
pass to the `addToPipeline`-pass API.
…_reduce_matmul.
This patch exposes broadcast and transpose semantics on
'batch_reduce_matmul'. This is the last one in continuation of other two
variant of matmul ops.
The broadcast and transpose semantic are as follows:
Broadcast and Transpose semantics can be appiled by specifying the
explicit attribute 'indexing_maps' as shown below. This is a list
attribute, so must include maps for all arguments if specified.
Example Transpose:
```
linalg.batch_reduce_matmul indexing_maps = [
affine_map<(d0, d1, d2, d3) -> (d0, d3, d1)>, // transpose
affine_map<(d0, d1, d2, d3) -> (d0, d3, d2)>,
affine_map<(d0, d1, d2, d3) -> (d1, d2)>
]
ins(%arg0, %arg1 : memref<2x5x3xf32>,memref<2x5x7xf32>)
outs(%arg2: memref<3x7xf32>)
```
Example Broadcast:
```
linalg.batch_reduce_matmul indexing_maps = [
affine_map<(d0, d1, d2, d3) -> (d3)>, // broadcast
affine_map<(d0, d1, d2, d3) -> (d0, d3, d2)>,
affine_map<(d0, d1, d2, d3) -> (d1, d2)>
]
ins(%arg0, %arg1 : memref<5xf32>, memref<2x5x7xf32>)
outs(%arg2: memref<3x7xf32>)
```
Example Broadcast and Transpose:
```
linalg.batch_reduce_matmul indexing_maps = [
affine_map<(d0, d1, d2, d3) -> (d1, d3)>, // broadcast
affine_map<(d0, d1, d2, d3) -> (d0, d2, d3)>, // transpose
affine_map<(d0, d1, d2, d3) -> (d1, d2)>
]
ins(%arg0, %arg1 : memref<3x5xf32>, memref<2x7x5xf32>)
outs(%arg2: memref<3x7xf32>)
```
RFCs and related PR:
https://discourse.llvm.org/t/rfc-linalg-opdsl-constant-list-attribute-definition/80149https://discourse.llvm.org/t/rfc-op-explosion-in-linalg/82863https://discourse.llvm.org/t/rfc-mlir-linalg-operation-tree/83586https://github.com/llvm/llvm-project/pull/115319https://github.com/llvm/llvm-project/pull/122275
let constructor is legacy (do not use in tree!) since the tableGen
backend emits most of the glue logic to build a pass.
Note: The following constructor has been retired:
```cpp
std::unique_ptr<Pass> createExpandReallocPass(bool emitDeallocs = true);
```
To update your codebase, replace it with the new options-based API:
```cpp
memref::ExpandReallocPassOptions expandAllocPassOptions{
/*emitDeallocs=*/false};
pm.addPass(memref::createExpandReallocPass(expandAllocPassOptions));
```
Ops that are already snake case (like [`ROCDL_wmma_*`
ops](66b0b0466b/mlir/include/mlir/Dialect/LLVMIR/ROCDLOps.td (L411)))
produce python "value-builders" that collide with the class names:
```python
class wmma_bf16_16x16x16_bf16(_ods_ir.OpView):
OPERATION_NAME = "rocdl.wmma.bf16.16x16x16.bf16"
...
def wmma_bf16_16x16x16_bf16(res, args, *, loc=None, ip=None) -> _ods_ir.Value:
return wmma_bf16_16x16x16_bf16(res=res, args=args, loc=loc, ip=ip).result
```
and thus cannot be emitted (because of recursive self-calls).
This PR fixes that by affixing `_` to the value builder names.
I would've preferred to just rename the ops but that would be a breaking
change 🤷.
This PR is mainly about exposing the python bindings for
`linalg::isaConvolutionOpInterface` and `linalg::inferConvolutionDims`.
---------
Signed-off-by: Bangtian Liu <liubangtian@gmail.com>
This PR is mainly about exposing the python bindings for`
linalg::isaContractionOpInterface` and` linalg::inferContractionDims`.
---------
Signed-off-by: Bangtian Liu <liubangtian@gmail.com>
This commit extends the MLIR vector type to support pointer-like types
such as `!llvm.ptr` and `!ptr.ptr`, as indicated by the newly added
`VectorTypeElementInterface`. This makes the LLVM dialect closer to LLVM
IR. LLVM IR already supports pointers as vector element type.
Only integers, floats, pointers and index are valid vector element types
for now. Additional vector element types may be added in the future
after further discussions. The interface is still evolving and may
eventually turn into one of the alternatives that were discussed on the
RFC.
This commit also disallows `!llvm.ptr` as an element type of
`!llvm.vec`. This type exists due to limitations of the MLIR vector
type.
RFC:
https://discourse.llvm.org/t/rfc-allow-pointers-as-element-type-of-vector/85360
This is an implementation for [RFC: Supporting Sub-Channel Quantization
in
MLIR](https://discourse.llvm.org/t/rfc-supporting-sub-channel-quantization-in-mlir/82694).
In order to make the review process easier, the PR has been divided into
the following commit labels:
1. **Add implementation for sub-channel type:** Includes the class
design for `UniformQuantizedSubChannelType`, printer/parser and bytecode
read/write support. The existing types (per-tensor and per-axis) are
unaltered.
2. **Add implementation for sub-channel type:** Lowering of
`quant.qcast` and `quant.dcast` operations to Linalg operations.
3. **Adding C/Python Apis:** We first define he C-APIs and build the
Python-APIs on top of those.
4. **Add pass to normalize generic ....:** This pass normalizes
sub-channel quantized types to per-tensor per-axis types, if possible.
A design note:
- **Explicitly storing the `quantized_dimensions`, even when they can be
derived for ranked tensor.**
While it's possible to infer quantized dimensions from the static shape
of the scales (or zero-points) tensor for ranked
data tensors
([ref](https://discourse.llvm.org/t/rfc-supporting-sub-channel-quantization-in-mlir/82694/3)
for background), there are cases where this can lead to ambiguity and
issues with round-tripping.
```
Consider the example: tensor<2x4x!quant.uniform<i8:f32:{0:2, 0:2}, {{s00:z00, s01:z01}}>>
```
The shape of the scales tensor is [1, 2], which might suggest that only
axis 1 is quantized. While this inference is technically correct, as the
block size for axis 0 is a degenerate case (equal to the dimension
size), it can cause problems with round-tripping. Therefore, even for
ranked tensors, we are explicitly storing the quantized dimensions.
Suggestions welcome!
PS: I understand that the upcoming holidays may impact your schedule, so
please take your time with the review. There's no rush.
* `PyRegionList` is now sliceable. The dialect bindings generator seems
to assume it is sliceable already (!), yet accessing e.g. `cases` on
`scf.IndexedSwitchOp` raises a `TypeError` at runtime.
* `PyBlockList` and `PyOperationList` support negative indexing. It is
common for containers to do that in Python, and most container in the
MLIR Python bindings already allow the index to be negative.
Updated the Python diagnostics handler to emit notes (in addition to
errors) into the output stream so that users have more context as to
where in the IR the error is occurring.
In some projects like JAX ir.Context are used with disabled multi-threading to avoid
caching multiple threading pools:
623865fe95/jax/_src/interpreters/mlir.py (L606-L611)
However, when context has enabled multithreading it also uses locks on
the StorageUniquers and this can be helpful to avoid data races in the
multi-threaded execution (for example with free-threaded cpython,
https://github.com/jax-ml/jax/issues/26272).
With this PR user can enable the multi-threading: 1) enables additional
locking and 2) set a shared threading pool such that cached contexts can
have one global pool.
This PR extends the python bindings for CallSiteLoc, FileLineColRange,
FusedLoc, NameLoc with field accessors. It also adds the missing
`value.location` accessor.
I also did some "spring cleaning" here (`cast` -> `dyn_cast`) after
running into some of my own illegal casts.
The current `write_bytecode` implementation necessarily requires the
serialized module to be duplicated in memory when the python `bytes`
object is created and sent over the binding. For modules with large
resources, we may want to avoid this in-memory copy by serializing
directly to a file instead of sending bytes across the boundary.
This PR https://github.com/llvm/llvm-project/pull/123902 broke python
bindings for `tensor.pack`/`unpack`. This PR fixes that. It also
1. adds convenience wrappers for pack/unpack
2. cleans up matmul-like ops in the linalg bindings
3. fixes linalg docs missing pack/unpack
As linalg.batch_matmul has been moved into tablegen from OpDSL, its
derived python wrapper no longer exist.This patch adds the required
python wrapper.
Also refactors the BatchmatmulOp printer to make it consistent with its
parser.
For extremely large models, it may be inefficient to load the model into
memory in Python prior to passing it to the MLIR C APIs for
deserialization. This change adds an API to parse a ModuleOp directly
from a file path.
Re-lands
[4e14b8a](4e14b8afb4).
Implement the feature about perf by stage(llvm-ir -> isa, isa->binary).
The results will be stored into the properties, then users can use them
after using GpuModuleToBinary Pass.
Now that linalg.matmul is in tablegen, "hand write" the Python wrapper
that OpDSL used to derive. Similarly, add a Python wrapper for the new
linalg.contract op.
Required following misc. fixes:
1) make linalg.matmul's parsing and printing consistent w.r.t. whether
indexing_maps occurs before or after operands, i.e. per the tests cases
it comes _before_.
2) tablegen for linalg.contract did not state it accepted an optional
cast attr.
3) In ODS's C++-generating code, expand partial support for `$_builder`
access in `Attr::defaultValue` to full support. This enables access to
the current `MlirContext` when constructing the default value (as is
required when the default value consists of affine maps).
For extremely large models, it may be inefficient to load the model into
memory in Python prior to passing it to the MLIR C APIs for
deserialization. This change adds an API to parse a ModuleOp directly
from a file path.
If the large element limit is specified, large elements are hidden from
the asm but large resources are not. This change extends the large
elements limit to apply to printed resources as well.
Model the `IndexType` as `uint64_t` when converting to a python integer.
With the python bindings,
```python
DenseIntElementsAttr(op.attributes["attr"])
```
used to `assert` when `attr` had `index` type like `dense<[1, 2, 3, 4]>
: vector<4xindex>`.
---------
Co-authored-by: Christopher McGirr <christopher.mcgirr@amd.com>
Co-authored-by: Tiago Trevisan Jost <tiago.trevisanjost@amd.com>
