This commit adds decomposition support into the core aten operators
before importing the module...

Just a minor update to skip the space char if included in the equation
string

---------

C...

this introduces an additional soft dependency on the python ml_dtypes
python packages in order ...

This op adds two FIntegerType operands to produce an FIntegerType
result. This defines the usua...

This is currently always defining the inferReturnTypes method from
InferTypeOpInterface on all ...

Signed-off-by: Julian Oppermann <[email protected]>

This version of pytorch includes a patch to enable dynamo support on
Windows, so I would like t...

These currently default to calling emitError(), which uses the current
token's location. In som...

On modules we keep an array attribute `port_locs` for port location
information. On regular HWM...

Also renamed `targetNames` to `caseNames` to better match the FIRRTL-level terminology

Now that we're not using libpgmath, these tests pass on AArch64.

They may work on Windows now...

Which is what I tried to do in 330340b14f6325d76c96cb1713d29c4258e541b8,
but forgot one detail....

This commit adds the OnnxToTorch lowering for cosh, acosh, asin, asinh,
and atanh op.
This com...

The previous fix in https://github.com/llvm/circt/pull/6572 did not
get to the root of the issue...

Some operations include a backend matcher for specialized operations. We
map these back to gene...

* Change isMovableDeclaration to allow operations to be lifted if all their operands are constan...

Fixes https://github.com/llvm/torch-mlir/issues/2866

Some backends / downstream projects expe...

To accomodate upcoming slang 4 builds. Changes the version of clang in
the integration tests to 16.

This test exposes issues that need fixing
(1) propagate sparsity into the FX graph (over elt-wi...

Updates to LowerLayers to support driving layer-associated probes.

Signed-off-by: Schuyler El...

The investigation is largely recorded in
https://github.com/llvm/torch-mlir/pull/2881, but this...

This test was added with a small patch recently, but the test didn't
really check anything beyon...

LLVMPoller can be configured to use other `repourl` for test/debug
purpose. But LLVMPoller alwa...

This commit adds the OnnxToTorch support for Mean, IsInf, IsNaN, and
PRelu ops. All high priori...

The previous logic would create classes for any FIRRTL modules that
had property ports. However...

Various improvements on sparsity metadata:

(1) define single data structure for all sparsity ...

This patch makes the Protobuf package mandatory in addition to forcing a
config mode search. T...

This patch copies Fuchsia's staging build bots configurations to the
Fuchsia prod builders, whi...

This PR introduces a sparse_jit wrapper that can run simple models with
sparse tensor inputs en...

The clang-ppc64le-rhel buildbot uses a different external compiler
toolchain to the system defa...

Co-authored-by: Mike Urbach <[email protected]>

By updating convertScalarToDtype invocation pass original source and
destination datatypes for ...

* PrintStateInfo refactoring

* remove printinfopass

* fix failing tests

* extract arcil...

After https://reviews.llvm.org/D140236 (and other changes I'm sure),
flang doesn't need pgmath ...

This disables host/device and device/host transfers via the SDMA engines
for the OpenMP Offload...

…dcc7e89.
* to involve `chlo-legalize-to-stablehlo` pass.

Added a folded for a equals operator. This allows an equivalent
comparison folder, primarily fo...

Includes some minor first for `AffineMap::inferFromExprList`

If the input or result is a splat value we can just constant fold the
result. This is common fo...

Make `ImportVerilog` honor the `singleUnit` option by either adding all
source files to a singl...

Fix an error if a 0-bit was shifted right by any amount and this made it
all the way to LowerTo...

The lowering decomposes AtenTraceOp into an AtenDiagonalOp followed by
AtenSumOp.

The progre...

A hidden command-line argument is added to re-enable the pass on possible backports or during th...

Previously the builder class argument "jobs" defaulted to a string to
use with `WithProperties`...

Emit public FIRRTL dialect modules with the "public" keyword. This brings
the FIRRTL emitter fu...

Add some missing 'sstream' includes that were need on a fresh build on
macOS.

Signed-off-by: Sc...

Add python wheel instructions to README

Accept `EICG_wrapper` extmodules without `test_en` port in
`LowerIntrinsics`, and simply leave ...

Found via the (TIL) LLVM utility llvm/utils/filecheck_lint/filecheck_lint.py.

We don't need the runtime variants on build bots. This patch disable it
so the build time can b...

There is no lowering support for math::AbsIOp, so if the operand is an
integer type, it will fa...

Fix support for refs into extmodule that's a public module within a separate circuit.

Prepare...

These are used in contexts that are not SymbolTables, so these cannot
be Symbols. It appears th...

The FullAsynchronousResetAnnotation is used to mark an asynchronous
reset signal in a module to...

Fleshing out support for standard services and addition to the manifest.

These are used in contexts that are not SymbolTables, so these cannot
be Symbols. It appears th...

This test appears to be testing ExportVerilog for a pattern that isn't
actually legal, since a `...

This patch updates the configuration for the Fuchsia staging builder to
use a single stage buil...

Also, ban capturing outer properties inside a layerblock.

Add missing requirements on slang for recently added circt-verilog
tests. This gets my local bui...

Add the `ImportVerilogOptions` struct to capture various options that
can be passed to the Slan...

This adds a few passes that will ensure linalg with sparse tensors are
properly lowered to loop...

This was unused. Fixes #6660.

Provide intrinsics to capture what today is encoded via printf + when/stop/verif-op pattern matc...

Reuses [build_linux_packages.sh](https://github.com/llvm/torch-mlir/blob/main/build_tools/python...

shared build seems to require a specific version of glibc and causes a build failure of downstre...

The enabled layers of a module act as requirements on the instance op--the
module may only be i...

rewriteModuleBody is ran parallely so we can ignore the overhead of
plain getFieldRefFromValue.

Bumps [tar](https://github.com/isaacs/node-tar) to 6.1.11 and updates
ancestor dependency [gulp...

This PR contains three commits to update the validation checks in the
ONNX -> Torch conversion ...

This commit modifies the OnnxToTorch lowering of Onnx.Reshape op by
creating the result shape l...