Fix integer overflow causing incorrect results. Kernels perform the addition of the destination zero_point in int16. This addition needed to be saturating to avoid wrapping around. Thanks to Marat Dukhan for reporting and debugging this issue. Additionally, this commit: - makes the new Cortex-X1 tuned kernels tested. - adds Context::get_runtime_enabled_paths() to query the runtime CPU detection from the public Context interface. - updates the Bazel-to-CMake converter to support some minor recent BUILD changes. PiperOrigin-RevId: 379778779
This is not an officially supported Google product.
ruy is a matrix multiplication library. Its focus is to cover the matrix multiplication needs of neural network inference engines. Its initial user has been TensorFlow Lite, where it is used by default on the ARM CPU architecture.
ruy supports both floating-point and 8bit-integer-quantized matrices.
ruy is designed to achieve high performance not just on very large sizes, as is the focus of many established libraries, but on whatever are the actual sizes and shapes of matrices most critical in current TensorFlow Lite applications. This often means quite small sizes, e.g. 100x100 or even 50x50, and all sorts of rectangular shapes. It's not as fast as completely specialized code for each shape, but it aims to offer a good compromise of speed across all shapes and a small binary size.
Some documentation will eventually be available in the doc/ directory, see doc/README.md.