* This extracts metadata from python invocations (nearly) sufficient to generate ODS and a Torch IR translation table for most of the ops.
* It also has the side effect of creating a data structure with meaningfully runnable examples suitable for an automated regression test.
* There are some ops that are sufficiently complex/weird (like _convolution) that we'll just manually handle those.
* See example output: https://gist.github.com/stellaraccident/60a58457b15e9184e224fa98a2658769
* Make code that depends on the legacy "type dispatch" mechanism optional.
* This code is fairly tied to a specific ~1.3 version and uses a legacy dispatch mechanism.
* Moving it and making it optional allows the project to build with PyTorch 1.6 and makes it possible for us to start building out a more modern interface mechanism in parallel.
* Some of the moved code will be brought back into the more modern path, but isolating it now lets this be done incrementally.
* Tests are left failing since the entire frontend is optional and the next step involves reworking the interface mechanism to get them to passing in both regimes.
* Fix a few bogons to get things building
* Add Dockerfile with pytorch
Also, I configure with:
-DCMAKE_PREFIX_PATH="/opt/pytorch/pytorch"
(which is where pytorch is installed in this container)
* Make a dep conditional.
Co-authored-by: stephenneuendorffer <stephen.neuendorffer@xilinx.com>
This patch adds a pytorch interface to npcomp. This interface is modeled
after pytorch_xla and exposes the MLIR-based flow as a virtual device (similar
to a gpu device or the xla backend). Usage is intended to be something like:
dev = torch_mlir.mlir_device()
t0 = torch.randn((4,4), device=dev)
t1 = torch.randn((4,4), device=dev)
t2 = t0 + t1
t2_mlir = torch_mlir.get_mlir( t2 )
t2_cpu = t2.to('cpu')
In this case t2_cpu would contain the result of the computation, and t2_mlir
contains the mlir description of the computation. Note that this also
properly returns backward paths synthesized by pytorch. There are several
parts of this:
1) A tensor type (implemented by tensor.* and tensor_impl.*)
2) The device modeling (aten_mlir_bridge.*, aten_mlir_device.*, aten_mlir_type*)
3) a temporary IR (implemented by ir.cpp)
There is also a reference lowering directly from the ATen dialect to C
function calls consisting of two parts:
1) The driver that uses the IR to generate MLIR, run Passes and compile the
result using mlir::ExecutionEngine (implemented by jit.cpp and
mlir_gen.cpp)
2) A runtime library implemented by lib/aten_ops.cpp. Most of the operations
are implemented by callbacks into the torch C++ libraries.
Some aspects of this are known to be less than optimal, in particular:
1) There's some function definitions that don't live in the file corresponding
to their declaration.
2) More aspects of this (e.g. the IR) seem like they should be automatically
generated.
3) It's unclear to me how much of the 'IR' is actually necessary, or whether
MLIR could be created on the fly.
Note that this code is licensed in a way similar to pytorch, with the
intention that eventually (when npcomp reaches some maturity) it should be
pushed there. (see frontends/pytorch/LICENSE) The code is also structured
much closer to the pytorch coding style than the LLVM coding style.
* I expect that mlir-hlo will be a non-optional dependency of the project, so adding as a sub-module.
* IREE is an optional dependency and I'm keeping this as an out-of-tree checkout for the moment.
* The script will compute the join across both iree and mlir-hlo to find a common LLVM version.
* The script needs some more work (like a flag that says to update the version, etc). Likely needs more testing through an integrate or two.
This patch adds a dialect intended to be used as a frontend dialect
to facilitate lowering from "A Tensor Library" in torch/pytorch.
This patch includes several passes that are useful in conjuction with the
dialect:
--aten-layer-name: Generates layer names for each operation, which are not
present in the original pytorch.
--aten-to-std: Lower the ATen dialect into standard dialect function calls.
--return-elimination-pass: convert functions (primarily the toplevel function)
to pass return values by reference. This simplifies pytorch integration.
--aten-op-report: generate a textual report about the model
--liveness-report
Future patches will implement actual integration with the pytorch jit to
intercept and generates MLIR in this dialect, then lower the resulting MLIR
into function calls through aten-layer-name -> aten-to-std ->
return-elimination -> std-to-llvm. The result would then jitted using the LLVM
jit, linked against a runtime library which makes calls back into pytorch to
implement all the layers.
Co-authored-by: Jeff Fifield <jeff.fifield@xilinx.com>
Co-authored-by: Jeff Fifield <jeff.fifield@xilinx.com>
Mostly this is CMake cleanup. Several library dependencies are missing, which
is often revealed with shared library builds. Also, it's generally bad to
link directly against LLVM libraries because it fails when using
LLVM_LINK_LLVM_DYLIB. MLIR will pull in libLLVM.so, and there will be
duplicate linkage with the the explicit libraries. There may need to be more
refactoring here.
* Since the manylinux images do not hard-link against python libs (resolving them at runtime), the module must be built without resolving undefined references.
* For some reason, builds on this platform are stricter, exposing dependency ordering issues.
* The CMake bits to build the extension are still somewhat of a mess. I have better versions both upstream and in IREE and will be reconciling. For now, don't look too closely.
* These were separated originally for layering reasons that no longer apply.
* Most of the python extension code is under lib/ with just the module setup in python/.
* Uses local configs and unsupported annotation to disable optional tests.
* This separation was just an artifact of having initial trouble getting lit setup.
* Primarily, the upstream shape dialect now uses tensor<?xindex> for non-erroring, immediate shape calculations (and will return this for shape_of of a tensor or memref).
* In addition, upstream passes do not yet exist for fully lowering to standard ops, so the passes here need to be extended to handle this new convention.
* This should be seen as an intermediate state, necessary to integrate a new LLVM version and needs more work and cleanup for generality.
* There is a good deal of awkwardness in these conversions. The hope is that additional upstream work will yield better defined conversion paths once out of this intermediate state.
* Updates the LICENSE to the same verbiage as used in the circt project.
* Adds the incubator disclaimer to the README.
* Reworks the introduction of the README to more accurately reflect the eventual scope.
* There is a fair amount of further rework of the repo that needs to take place. This is just the minimal cosmetic changes now that it is part of LLVM.
My main interest in this is that tweaking the default of this flag is a
quick way to check for miscompiling canonicalizations / op definitions
not annotated properly (e.g. marked NoSideEffect when in fact it is not
safe to do so).
* Enables e2e test.
* With what I've learned in upstream about test directory layout, I can consolidate most of the separate directories we have for these things. Will do that in a followup.
* Not pleased with the LLVM global initialization depends but serviceable for now.
This required making module descriptors hold a FuncDescriptor* instead
of a pointer to array of FuncDescriptors as it previously did, which is
innocuous (just requires an llvm.bitcast after the llvm.mlir.addressof).
* Rewrites public function signatures to operate on tensors (vs ndarray).
* Most of our backends presume immutable tensors at public function boundaries.
Also, the previous code had a special case for deleting this op when it
had no uses. This is subsumed by the change in this commit since now
shape.const_shape is properly lowered.
With this change, the included test case with multiple serially
dependent ops works!
This specific issue was related to the scalar argument to that
function. We needed to compute a broadcast of a scalar shape (which is a
shape.const_shape) with another shape.
This ~totally reworks the existing "runtime" stuff to be more
principled and usable, such as from Python. It's still not fully
production-quality, mainly in the department of memory management (e.g.
it currently leaks memory; we need to figure out "who frees memrefs" +
the analysis and transformation needed to do that (maybe use upstream
buffer allocation pass?)).
The user API is in include/npcomp/runtime/UserAPI.h, though
include/npcomp/JITRuntime/JITModule.h is a friendlier wrapper.
The stuff under {include,lib}/runtime is totally firewalled from the
compiler and tiny (<6kB, though no attention has gone into optimizing
that size). For example, we don't link in libSupport into the runtime,
instead having our own bare bones replacements for basics like ArrayRef
(the JITRuntime helps with bridging that gap, since it *can* depend on
all common LLVM utilities).
The overall features of npcomprt is that it exposes a module that
with multiple function entry points. Each function has arguments and
results that are tensor-valued, and npcomprt::Tensor is the runtime type
that is used to interact with that (and a npcomprt::Ref<T>
reference-counting wrapper is provided to wrap npcomprt::Tensor in the
common case).
From an implementation perspective, an npcomprt module at the
LLVM/object/binary level exposes a single module descriptor struct that
has pointers to other metadata (currently just a list of function
metadata descriptors). All interactions with the npcomp runtime are
keyed off of that module descriptor, including function lookups and
dispatching. This is done to dodge platform ABI issues and also allow
enough reflection to e.g. verify provided arguments.
Most of the compiler-side work here was in LowerToNpcomprtABI and
LowerToLLVM.
Also,
- Rename npcomp_rt/NpcompRt to npcomprt/Npcomprt; it was getting
annoying to type the underscores/caps.
- misc improvements to bash_helpers.sh
* This elides the very common code the compiler adds for chaining otherwise tensor-related numpy ops together.
* More aggressive canonicalizations would require more advanced analysis.
Stella Laurenzo
stephenneuendorffer
stephenneuendorffer
Phoenix Meadowlark
Sean Silva