VeriSilicon today announced significant milestones have been achieved for its versatile and highly scalable neural network inference engine family VIP8000. The fully programmable VIP8000 processors reach the performance and memory efficiency of dedicated fixed-function logic with the customizability and future proofing of full programmability in OpenCL, OpenVX, and a wide range of NN frameworks including NNEF. “The biggest thing to happen in the computer industry since the PC is AI and machine learning, it will truly revolutionize, empower, and improve our lives. It can be done in giant machines from IBM and Google, and in tiny chips made with VeriSilicon’s neural network processors,” said Dr. Jon Peddie, president Jon Peddie Research. “By 2020 we will wonder how we ever lived without our AI assistants,” he added.
The Khronos Group announces that the Vulkan Working Group’s Portability Initiative has been working with Khronos members Valve, LunarG, and The Brenwill Workshop to enable Vulkan applications to be ported to Apple platforms. The Vulkan Portability resource page links to a collection of free and open source set of tools, SDKs, and runtime libraries to enable Vulkan development on macOS and deployment on macOS and iOS platforms. Valve is extending Dota 2 using the Vulkan tools on macOS to achieve significantly higher performance than native OpenGL drivers. Vulkan support for Dota 2 on macOS will be released in the coming months as a free update.
Valve, LunarG, and the Brenwill Workshop join forces with Khronos to release the first open source SDK for Mac. For more information about the partnership that made this SDK possible, read the Khronos press release. This SDK supports Vulkan API revision 1.0.69 on macOS (MoltenVK subset).
A new milestone of the Magnum C++11/C++14 graphics engine brings WebGL 2.0 and WebAssembly, VR support, lots of niceties for Windows users, iOS port, new experimental UI library, improved testing capabilities, support for over 80 new asset formats, new examples and much more.
Sundog Software released version 4.0 of its ocean water simulation library for OpenGL, the Triton Ocean SDK. Triton 4 features a re-architecture to align it with modern rendering architectures, and uses OpenGL 4.5 and certain NVidia extensions to implement a Vulkan-like approach to rendering water. Multi-threaded processing of command lists, bindless rendering, and bindless uniform buffer objects all work to maximize performance, especially when rendering multiple views concurrently in VR applications. Triton has also updated to use NVidia’s CUDA Toolkit 9.1 under the hood for accelerating the Fast Fourier Transforms that power its ocean wave model. Triton allows you to simulate physically-accurate seas for any sea state or swell conditions, and supports ship wakes, reflections, rotor wash, coastal effects, and more. It’s used worldwide in hundreds of maritime training systems and games.
CG Internals published a blog article covering screen-filling rasterization using graphics hardware and modern OpenGL. The findings are applicable to OpenGL ES, Vulkan, and WebGL as well. For rendering screen-filling geometry we usually have to choose between a screen-aligned quad and a screen-aligned triangle. But - is there a difference? If so, which approach is better than the other? In this article we want to show you the differences between both approaches and offer an alternative. Following the theoretical analysis we introduce a demo program and evaluate screencasts together with multiple performance measures.
Verge3D is based on WebGL and integrate a glTF exporter. Verge3D enables developing and publishing models, scenes and entire 3D web applications online. Verge3D includes a visual editor called Puzzles which allows for setting up interactive scenarios for your web apps. This tool is based on Google’s Blockly framework used in education and other industries. If you are a 3D artist, you will appreciate Puzzles which gives you the power to directly express your creativity in the realm of interactive 3D Web.
LunarG creates tools to help simplify Vulkan development. We leveraged the new Vulkan Layer Factory to create the Vulkan Assistant Layer, a layer that helps developers identify Vulkan best practices. The Vulkan Assistant Layer — VK_LAYER_LUNARG_assistant_layer — functions as a Vulkan best practices layer and is intended to highlight potential performance issues, questionable usage patterns, common mistakes, and items that may lead to application problems that are not specifically prohibited by the Vulkan specification. The Vulkan Assistant Layer can be found as part of the LunarG Vulkan SDK.
Starting today, Facebook is rolling out support for the industry standard glTF 2.0 file format for Facebook 3D posts. 3D objects or scenes saved in glTF can be dragged straight to a browser window to add to your Facebook account. The company is also adding the feature to its platform tools so developers can build ways to export creations to Facebook from various apps. With glTF 2.0 support, Facebook is opening up even more ways to share 3D content on Facebook from more creation tools and platforms. They’re introducing new Graph API endpoints with 3D Post support so developers can build seamless 3D sharing into any app — letting people share interactive objects or scenes directly to Facebook with just a click. Learn more about glTF and what Facebook is doing here, and check out a cool example of glTF in action here.
The next AEC Hackathon will be held in San Francisco on February 23-25, 2018. Four challenges await: Best overall project; Best project that solves a big AEC (Architecture, Engineering, Construction) problem; Best hack from a past event and Best mashup project. Tickets are now online. Get your Hackathon hat on and head over to the event registration today! The Khronos Group is a proud supporting organization of AEC Hackathon.
The Khronos Group announces the release of a geometry compression extension to glTF 2.0 using Google Draco technology to significantly reduce the size of glTF models and scenes. The Khronos glTF Draco extension specification is accompanied by optimized, open source compression and decompression libraries on the Draco GitHub site to enable the rapid deployment of glTF compressed geometry into tools, engines, applications, and browsers everywhere.
The Khronos Neural Network Exchange Format, among other technologies, go a long way to enable highly optimized implementations of inference for systems trained on a range of systems. Explains Chris Rowen, CEO for Babblabs, “This is extremely valuable to opening up the path to exploit optimized high-volume inference engines in phones, cars, cameras and other IoT devices. This higher-level robust set of interfaces breaks the tyranny of instruction set compatibility as a standard for exchange and allows for greater levels of re-optimization as the inference execution hardware evolves over time.” Read more on the Semiconductor Engineering blog.
Standards make life easier, and we depend on them for more than we might realize — from knowing exactly how to drive any car, to knowing how to get hot or cold water from a faucet. Balancing the need for a stable standard, while at the same time allowing technology advances to be rapidly exploited, is a big part of what Khronos does. There are two ways a Khronos standard can be extended: Vendor Extensions and Khronos Extensions. Read on to learn how both of these work within Khronos.
This podcast episode of “The Interview” with The Next Platform focuses on an effort to standardize key neural network features to make development and innovation easier and more productive. To explore this topic, The Next Platform was joined by Neil Trevett. Listen to the podcast and read the write up.
NNEF and ONNX are two similar open formats to represent and interchange neural networks among deep learning frameworks and inference engines. At the core, both formats are based on a collection of often used operations from which networks can be built. Because of the similar goals of ONNX and NNEF, we often get asked for insights into what the differences are between the two. Read the Khronos blog to learn more about the similarities and differences between NNEF and ONNX.