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>
- 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.
* `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.
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.
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).
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.
This logic is in the critical path for constructing an operation from
Python. It is faster to compute this in C++ than it is in Python, and it
is a minor change to do this.
This change also alters the API contract of
_ods_common.get_op_results_or_values to avoid calling
get_op_result_or_value on each element of a sequence, since the C++ code
will now do this.
Most of the diff here is simply reordering the code in IRCore.cpp.
Currently we make two memory allocations for each PyOperation: a Python
object, and the PyOperation class itself. With some care we can allocate
the PyOperation inline inside the Python object, saving us a malloc()
call per object and perhaps improving cache locality.
Previously ODS-generated Python operations had code like this:
```
super().__init__(self.build_generic(attributes=attributes, operands=operands, successors=_ods_successors, regions=regions, loc=loc, ip=ip))
```
we change it to:
```
super().__init__(self.OPERATION_NAME, self._ODS_REGIONS, self._ODS_OPERAND_SEGMENTS, self._ODS_RESULT_SEGMENTS, attributes=attributes, operands=operands, successors=_ods_successors, regions=regions, loc=loc, ip=ip)
```
This:
a) avoids an extra call dispatch (to `build_generic`), and
b) passes the class attributes directly to the constructor. Benchmarks
show that it is faster to pass these as arguments rather than having the
C++ code look up attributes on the class.
This PR improves the timing of the following benchmark on my workstation
from 5.3s to 4.5s:
```
def main(_):
with ir.Context(), ir.Location.unknown():
typ = ir.IntegerType.get_signless(32)
m = ir.Module.create()
with ir.InsertionPoint(m.body):
start = time.time()
for i in range(1000000):
arith.ConstantOp(typ, i)
end = time.time()
print(f"time: {end - start}")
```
Since this change adds an additional overload to the constructor and
does not alter any existing behaviors, it should be backwards
compatible.
In JAX, I observed a race between two PyOperation destructors from
different threads updating the same `liveOperations` map, despite not
intentionally sharing the context between different threads. Since I
don't think we can be completely sure when GC happens and on which
thread, it seems safest simply to add locking here.
We may also want to explicitly support sharing a context between threads
in the future, which would require this change or something similar.
This is a companion to #118583, although it can be landed independently
because since #117922 dialects do not have to use the same Python
binding framework as the Python core code.
This PR ports all of the in-tree dialect and pass extensions to
nanobind, with the exception of those that remain for testing pybind11
support.
This PR also:
* removes CollectDiagnosticsToStringScope from NanobindAdaptors.h. This
was overlooked in a previous PR and it is duplicated in Diagnostics.h.
---------
Co-authored-by: Jacques Pienaar <jpienaar@google.com>
Relands #118583, with a fix for Python 3.8 compatibility. It was not
possible to set the buffer protocol accessers via slots in Python 3.8.
Why? https://nanobind.readthedocs.io/en/latest/why.html says it better
than I can, but my primary motivation for this change is to improve MLIR
IR construction time from JAX.
For a complicated Google-internal LLM model in JAX, this change improves
the MLIR
lowering time by around 5s (out of around 30s), which is a significant
speedup for simply switching binding frameworks.
To a large extent, this is a mechanical change, for instance changing
`pybind11::` to `nanobind::`.
Notes:
* this PR needs Nanobind 2.4.0, because it needs a bug fix
(https://github.com/wjakob/nanobind/pull/806) that landed in that
release.
* this PR does not port the in-tree dialect extension modules. They can
be ported in a future PR.
* I removed the py::sibling() annotations from def_static and def_class
in `PybindAdapters.h`. These ask pybind11 to try to form an overload
with an existing method, but it's not possible to form mixed
pybind11/nanobind overloads this ways and the parent class is now
defined in nanobind. Better solutions may be possible here.
* nanobind does not contain an exact equivalent of pybind11's buffer
protocol support. It was not hard to add a nanobind implementation of a
similar API.
* nanobind is pickier about casting to std::vector<bool>, expecting that
the input is a sequence of bool types, not truthy values. In a couple of
places I added code to support truthy values during casting.
* nanobind distinguishes bytes (`nb::bytes`) from strings (e.g.,
`std::string`). This required nb::bytes overloads in a few places.
Why? https://nanobind.readthedocs.io/en/latest/why.html says it better
than I can, but my primary motivation for this change is to improve MLIR
IR construction time from JAX.
For a complicated Google-internal LLM model in JAX, this change improves
the MLIR
lowering time by around 5s (out of around 30s), which is a significant
speedup for simply switching binding frameworks.
To a large extent, this is a mechanical change, for instance changing
`pybind11::`
to `nanobind::`.
Notes:
* this PR needs Nanobind 2.4.0, because it needs a bug fix
(https://github.com/wjakob/nanobind/pull/806) that landed in that
release.
* this PR does not port the in-tree dialect extension modules. They can
be ported in a future PR.
* I removed the py::sibling() annotations from def_static and def_class
in `PybindAdapters.h`. These ask pybind11 to try to form an overload
with an existing method, but it's not possible to form mixed
pybind11/nanobind overloads this ways and the parent class is now
defined in nanobind. Better solutions may be possible here.
* nanobind does not contain an exact equivalent of pybind11's buffer
protocol support. It was not hard to add a nanobind implementation of a
similar API.
* nanobind is pickier about casting to std::vector<bool>, expecting that
the input is a sequence of bool types, not truthy values. In a couple of
places I added code to support truthy values during casting.
* nanobind distinguishes bytes (`nb::bytes`) from strings (e.g.,
`std::string`). This required nb::bytes overloads in a few places.
In the tablegen-generated Python bindings, we typically see a pattern
like:
```
class ConstantOp(_ods_ir.OpView):
...
def __init__(self, value, *, loc=None, ip=None):
...
super().__init__(self.build_generic(attributes=attributes, operands=operands, successors=_ods_successors, regions=regions, loc=loc, ip=ip))
```
i.e., the generated code calls `OpView.__init__()` with the output of
`build_generic`. The purpose of `OpView` is to wrap another operation
object, and `OpView.__init__` can accept any `PyOperationBase` subclass,
and presumably the intention is that `build_generic` returns a
`PyOperation`, so the user ends up with a `PyOpView` wrapping a
`PyOperation`.
However, `PyOpView::buildGeneric` calls `PyOperation::create`, which
does not just build a PyOperation, but it also calls `createOpView` to
wrap that operation in a subclass of `PyOpView` and returns that view.
But that's rather pointless: we called this code from the constructor of
an `OpView` subclass, so we already have a view object ready to go; we
don't need to build another one!
If we change `PyOperation::create` to return the underlying
`PyOperation`, rather than a view wrapper, we can save allocating a
useless `PyOpView` object for each ODS-generated Python object.
This saves approximately 1.5s of Python time in a JAX LLM benchmark that
generates a mixture of upstream dialects and StableHLO.
The MLIR C and Python Bindings expose various methods from
`mlir::OpPrintingFlags` . This PR adds a binding for the `skipRegions`
method, which allows to skip the printing of Regions when printing Ops.
It also exposes this option as parameter in the python `get_asm` and
`print` methods
The PR implements MLIR Python Bindings for a few simple edit operations
on Block arguments, namely, `add_argument`, `erase_argument`, and
`erase_arguments`.
When an operation is erased in Python, its children may still be in the
"live" list inside Python bindings. After this, if some of the newly
allocated operations happen to reuse the same pointer address, this will
trigger an assertion in the bindings. This assertion would be incorrect
because the operations aren't actually live. Make sure we remove the
children operations from the "live" list when erasing the parent.
This also concentrates responsibility over the removal from the "live"
list and invalidation in a single place.
Note that this requires the IR to be sufficiently structurally valid so
a walk through it can succeed. If this invariant was broken by, e.g, C++
pass called from Python, there isn't much we can do.
If the python callback throws an error, the c++ code will throw a
py::error_already_set that needs to be caught and handled in the c++
code .
This change is inspired by the similar solution in
PySymbolTable::walkSymbolTables.
This commit adds `walk` method to PyOperationBase that uses a python
object as a callback, e.g. `op.walk(callback)`. Currently callback must
return a walk result explicitly.
We(SiFive) have implemented walk method with python in our internal
python tool for a while. However the overhead of python is expensive and
it didn't scale well for large MLIR files. Just replacing walk with this
version reduced the entire execution time of the tool by 30~40% and
there are a few configs that the tool takes several hours to finish so
this commit significantly improves tool performance.
Currently, a method exists to get the count of the operation objects
which are still alive. This helps for sanity checking, but isn't
terribly useful for debugging. This new method returns the actual
operation objects which are still alive.
This allows Python code like the following:
```
gc.collect()
live_ops = ir.Context.current._get_live_operation_objects()
for op in live_ops:
print(f"Warning: {op} is still live. Referrers:")
for referrer in gc.get_referrers(op)[0]:
print(f" {referrer}")
```
This reverts commit bbc29768683b394b34600347f46be2b8245ddb30.
This change seems to be at odds with the non-owning part semantics of
MlirOperation in C API. Since downstream clients can only take and
return MlirOperation, it does not sound correct to force all returns of
MlirOperation transfer ownership. Specifically, this makes it impossible
for downstreams to implement IR-traversing functions that, e.g., look at
neighbors of an operation.
The following patch triggers the exception, and there does not seem to
be an alternative way for a downstream binding writer to express this:
```
diff --git a/mlir/lib/Bindings/Python/IRCore.cpp b/mlir/lib/Bindings/Python/IRCore.cpp
index 39757dfad5be..2ce640674245 100644
--- a/mlir/lib/Bindings/Python/IRCore.cpp
+++ b/mlir/lib/Bindings/Python/IRCore.cpp
@@ -3071,6 +3071,11 @@ void mlir::python::populateIRCore(py::module &m) {
py::arg("successors") = py::none(), py::arg("regions") = 0,
py::arg("loc") = py::none(), py::arg("ip") = py::none(),
py::arg("infer_type") = false, kOperationCreateDocstring)
+ .def("_get_first_in_block", [](PyOperation &self) -> MlirOperation {
+ MlirBlock block = mlirOperationGetBlock(self.get());
+ MlirOperation first = mlirBlockGetFirstOperation(block);
+ return first;
+ })
.def_static(
"parse",
[](const std::string &sourceStr, const std::string &sourceName,
diff --git a/mlir/test/python/ir/operation.py b/mlir/test/python/ir/operation.py
index f59b1a26ba48..6b12b8da5c24 100644
--- a/mlir/test/python/ir/operation.py
+++ b/mlir/test/python/ir/operation.py
@@ -24,6 +24,25 @@ def expect_index_error(callback):
except IndexError:
pass
+@run
+def testCustomBind():
+ ctx = Context()
+ ctx.allow_unregistered_dialects = True
+ module = Module.parse(
+ r"""
+ func.func @f1(%arg0: i32) -> i32 {
+ %1 = "custom.addi"(%arg0, %arg0) : (i32, i32) -> i32
+ return %1 : i32
+ }
+ """,
+ ctx,
+ )
+ add = module.body.operations[0].regions[0].blocks[0].operations[0]
+ op = add.operation
+ # This will get a reference to itself.
+ f1 = op._get_first_in_block()
+
+
# Verify iterator based traversal of the op/region/block hierarchy.
# CHECK-LABEL: TEST: testTraverseOpRegionBlockIterators
```
This fixes a longstanding bug in the `Context._CAPICreate` method
whereby it was not taking ownership of the PyMlirContext wrapper when
casting to a Python object. The result was minimally that all such
contexts transferred in that way would leak. In addition, counter to the
documentation for the `_CAPICreate` helper (see
`mlir-c/Bindings/Python/Interop.h`) and the `forContext` /
`forOperation` methods, we were silently upgrading any unknown
context/operation pointer to steal-ownership semantics. This is
dangerous and was causing some subtle bugs downstream where this
facility is getting the most use.
This patch corrects the semantics and will only do an ownership transfer
for `_CAPICreate`, and it will further require that it is an ownership
transfer (if already transferred, it was just silently succeeding).
Removing the mis-aligned behavior made it clear where the downstream was
doing the wrong thing.
It also adds some `_testing_` functions to create unowned context and
operation capsules so that this can be fully tested upstream, reworking
the tests to verify the behavior.
In some torture testing downstream, I was not able to trigger any memory
corruption with the newly enforced semantics. When getting it wrong, a
regular exception is raised.
This reverts commit 0d109035c29408f06efc148d67aab7e4b2aada5d.
Changes make Python bindings unbuildable without additional cmake
modifications (or modified `$PATH`).
```
/llvm-project/mlir/lib/Bindings/Python/IRCore.cpp:33:10: fatal error: 'funcobject.h' file not found
```
This header is provided by cpython, and we are not looking for that in
cmake.
Moreover, the nature of this change is not very clear to me. Seems to
replace one include with two dozens, presumably because the code is only
using transitively included headers, but the value for readability is
dubious. LLVM is also not strictly following IWYU.
Enables reusing the AsmState when printing from Python. Also moves the
fileObject and binary to the end (pybind11::object was resulting in the
overload not working unless `state=` was specified).
---------
Co-authored-by: Maksim Levental <maksim.levental@gmail.com>
