-
We are finally starting to use oneDPL in earnest in the [traccc](https://github.com/acts-project/traccc) project, now that #1060 is not an issue anymore.
Now I ran into a different, pretty interest…
-
* Describe the scenario and benefits that the feature supports.
> Today all errors and warnings are in the `Data` assembly. This is fine for errors and warnings shared across the WiX Toolset and .w…
-
This ticket tracks the creation of a new compiler backend for bpftrace to replace LLVM.
## What's wrong with LLVM?
LLVM produces great bytecode, but it is a very large library. This makes it diffi…
ajor updated
10 months ago
-
**Is your feature request related to a problem? Please describe.**
[Extending cubes](https://cube.dev/docs/product/data-modeling/concepts/code-reusability-extending-cubes) brings great code reusabili…
-
### 🐛 Describe the bug
When using torch.compile, parameter freezing can be enabled on Inductor with torch._inductor.config.freezing flag. Dynamo checks this flag and adds the model parameters and buf…
-
Please check ScalableTestGrids_NB_SymbolicJacobian_ScalableTestGrids.Models.Type1.Type1_N_1_M_1 with the new backend and -d=mergeComponents, see [report](https://libraries.openmodelica.org/branches/he…
-
### What happened?
Hello, I'm the author of [IREE.gd](https://github.com/iree-gd/iree.gd), a port of IREE to Godot game engine. Recently, I tried to use a pose detection model, specifically [singlepo…
-
Have GCC output tagha bytecode.
-
### 🐛 Describe the bug
`out=` variant of `native_batch_norm` operation does not work with `aot_eager` backend.
```python
args = (
torch.rand(5, 5, 5),
torch.rand(5),
torch.rand(5),…
-