As an open source deep learning compiler driven by the community, TVM is evolving quickly and well received by the industry. In this session, the architecture of the TVM stack will be introduced first, including some important features added recently such as AutoTVM and VTA (Versatile Tensor Accelerator) support. Then the build and deployment of deep learning models with TVM will be talked about, and ONNX (Open Neural Network eXchange format) is one of the model formats supported by TVM stack. Besides unified model format and operator definitions, ONNXIFI (ONNX Interface for Framework Integration) is another initiative from the ONNX community to define a cross-platform API, and how to fit TVM stack into ONNXIFI seems an interesting topic to discuss as well.
YVR18-332: TVM compiler stack and ONNX support
BKK19-111 - DRM HW Composer for Beagle X15 BoardTuesday, April 16, 2019
Describing the process of adaptation AOSP DRM HWC to be used on Beagle X15 Board (4.14 kernel).
This can be used as an example of launching the external/drm_hwc on a board: a simple "How to" with the minimun steps required to get the drm_hwc functional.
SAN19-413 - TEE based Trusted Keys in LinuxFriday, October 4, 2019
Protecting key confidentiality is essential for many kernel security use-cases such as disk encryption, file encryption and protecting the integrity of file metadata. Trusted and encrypted keys provides a mechanism to export keys to user-space for storage as an encrypted blob and for the user-space to later reload them onto Linux keyring without the user-space knowing the encryption key. The existing Trusted Keys implementation relied on a TPM device but what if you are working on a system without one?
This session will introduce a Trusted Keys implementation which relies on a much simpler trusted application running in a Trusted Execution Environment (TEE) for sealing and unsealing of Trusted Keys using a hardware unique key provided by the TEE.