Intel Corporation

Telefonaktiebolaget Lm Ericsson (publ)

Nvidia Corporation

Swiftfoot Graphics Ab

Jon N Hasselgren

Tomas G Akenine-Moller

Carl Jacob Munkberg

Robert M Toth

Franz Petrik Clarberg

Magnus Andersson

Jim K Nilsson

Tomas Guy Akenine-Möller

Jacob J Munkberg

Bjorn Johnsson

Petrik Clarberg

Prasoonkumar Surti

Jacob Munkberg

Marco Salvi

Ignacio Llamas

Christopher Ryan Wyman

Shiqiu Liu

Jacob Ström

Karthik Vaidyanathan

Saurabh Sharma

Jim Rasmusson

Rahul P Sathe

Bimal Poddar

Subramaniam M Maiyuran

Travis T Schluessler

Peter L Doyle

Thomas A Piazza

Robert L Farrell

Thomas F Raoux

Kalyan Kumar Bhiravabhatla

Ariel Berkovits

Tomas Akenine-Moller

Paul A Johnson

Björn Johnsson

Tomas Akenine-Möller

Matt Pharr

Carl Munkberg

Kevin Rogovin

Harri Syrja

Tomas G Akenine-Moeller

Abhishek Ventakesh

Tomas Akenine-Moeller

Robert T Toth

Tomas G Akenine-MÖLLER

Carl J Munkberg

Tomas Akenine-MÖLLER

The performance of a neural network is improved by applying quantization to data at various points in the network. In an embodiment, a neural network includes two paths. A quantization is applied to each path, such that when an output from each path is combined, further quantization is not required. In an embodiment, the neural network is an autoencoder that includes at least one skip connection. In an embodiment, the system determines a set of quantization parameters based on the characteristics of the data in the primary path and in the skip connection, such that both network paths produce output data in the same fixed point format. As a result, the data from both network paths can be combined without requiring an additional quantization.

Quantizing autoencoders in a neural network

Disclosed approaches may leverage the actual spatial and reflective properties of a virtual environment—such as the size, shape, and orientation of a bidirectional reflectance distribution function (BRDF) lobe of a light path and its position relative to a reflection surface, a virtual screen, and a virtual camera—to produce, for a pixel, an anisotropic kernel filter having dimensions and weights that accurately reflect the spatial characteristics of the virtual environment as well as the reflective properties of the surface. In order to accomplish this, geometry may be computed that corresponds to a projection of a reflection of the BRDF lobe below the surface along a view vector to the pixel. Using this approach, the dimensions of the anisotropic filter kernel may correspond to the BRDF lobe to accurately reflect the spatial characteristics of the virtual environment as well as the reflective properties of the surface.

Reflection denoising in ray-tracing applications

An apparatus and method are described for a non-uniform rasterizer. For example, one embodiment of an apparatus comprises: a graphics processor to process graphics data and render images using the graphics data; and a non-uniform rasterizer within the graphics processor to determine different resolutions to be used for different regions of an image, the non-uniform rasterizer to receive a plurality of polygons to be rasterized and to responsively rasterize the polygons in accordance with the different resolutions.

Apparatus and method for non-uniform frame buffer rasterization

Disclosed approaches may leverage the actual spatial and reflective properties of a virtual environment&#x2014;such as the size, shape, and orientation of a bidirectional reflectance distribution function (BRDF) lobe of a light path and its position relative to a reflection surface, a virtual screen, and a virtual camera&#x2014;to produce, for a pixel, an anisotropic kernel filter having dimensions and weights that accurately reflect the spatial characteristics of the virtual environment as well as the reflective properties of the surface. In order to accomplish this, geometry may be computed that corresponds to a projection of a reflection of the BRDF lobe below the surface along a view vector to the pixel. Using this approach, the dimensions of the anisotropic filter kernel may correspond to the BRDF lobe to accurately reflect the spatial characteristics of the virtual environment as well as the reflective properties of the surface.

A method for improving performance of generation of digitally represented graphics. The method comprises: receiving a first representation of a base primitive; providing a set of instructions associated with vertex position determination; executing said retrieved set of instructions on said first representation of said base primitive using bounded arithmetic for providing a second representation of said base primitive, and subjecting said second representation of said base primitive to a culling process. A corresponding apparatus and computer program product are also presented.

Method, apparatus, and computer program product for improved graphics performance

In accordance with some embodiments, a zero coverage test may determine whether a primitive such as a triangle relies on lanes between rows or columns or lines of samples. If so, the primitive can be culled in a zero coverage culling test.

Zero-coverage rasterization culling

Embodiments described herein enable a hierarchical-Z unit of a graphics processor to be primed using Hi-Z data generated by occlusion culling operations performed on a general purpose processor. One embodiment provides for instructions to cause operations including performing occlusion culling for a scene via the general purpose processor and storing generated hierarchical-Z data. The Hierarchical-Z data generated during the occlusion culling operations can be shared with the graphics processor and used to prime a hierarchical-Z unit of the graphics processor. The at least a portion of the scene can then be rendered using the hierarchical-Z data after priming the hierarchical-Z unit, improving the effectiveness of hierarchical-Z operations of the graphics processor for the scene.

Priming hierarchical depth logic within a graphics processor

Techniques related to graphics rendering including techniques for compression and/or decompression of graphics data by use of indexed subsets are described. In one example, compression in graphics rendering may include determining a plurality of color values associated with individual pixels of a tile of pixels, generating a subset of the plurality of color values such that the subset of the plurality of color values include one or more distinct color values from the plurality of color values, associating an index value with each color value of the subset of the plurality of color values, determining, for each of the individual pixels, an associated pixel index value to generate a plurality of pixel index value associated with the individual pixels of the tile of pixels, storing, in memory, graphics data including the subset of the plurality of color values, the associated index values, and the plurality of pixel values.

Subset based compression and decompression of graphics data

Methods and apparatus relating to techniques for provision of hierarchical Z-Culling (HiZ) optimization for texture-dependent discard operations are described. In an embodiment, a processor performs one or more operations (such as HiZ or Hierarchical Stencil test) on depth data of an image tile in response to a determination that texture space bounds of the image tile is fully opaque. The processor performs the one or more operations regardless of whether a discard operation is enabled. Other embodiments are also disclosed and claimed.

Hierarchical Z-culling (HiZ) optimization for texture-dependent discard operations

In accordance with some embodiments, the knowledge that a capped frame time is used can be exploited to reduce power consumption. Generally a capped frame time is a pre-allocated amount of time to apply power for rendering in graphics processing. Generally the frame time involves the application of power and some down time in which only idle power is applied pending the next frame time. By making better use of that down time, power consumption reductions can be achieved in some embodiments.

Reducing power consumption during graphics rendering

Growing community of inventors

Bulkelflostrand, Sweden

Jon N Hasselgren