<inline-formula><tex-math notation="LaTeX">$k^+$</tex-math><alternatives> <inline-graphic xlink:type="simple" xlink:href="vasilakis-ieq1-2417581.gif"/></alternatives></inline-formula>-buffer: An Efficient, Memory-Friendly and Dynamic <inline-formula><tex-math notation="LaTeX">$k$</tex-math><alternatives> <inline-graphic xlink:type="simple" xlink:href="vasilakis-ieq2-2417581.gif"/></alternatives></inline-formula>-buffer Fram

Abstract

Depth-sorted fragment determination is fundamental for a host of image-based techniques which simulates complex rendering effects. It is also a challenging task in terms of time and space required when rasterizing scenes with high depth complexity. When low graphics memory requirements are of utmost importance, k-buffer can objectively be considered as the most preferred framework which advantageously ensures the correct depth order on a subset of all generated fragments. Although various alternatives have been introduced to partially or completely alleviate the noticeable quality artifacts produced by the initial k-buffer algorithm in the expense of memory increase or performance downgrade, appropriate tools to automatically and dynamically compute the most suitable value of k are still missing. To this end, we introduce k(+)-buffer, a fast framework that accurately simulates the behavior of k-buffer in a single rendering pass. Two memory-bounded data structures: (i) the max-array and (ii) the max-heap are developed on the GPU to concurrently maintain the k-foremost fragments per pixel by exploring pixel synchronization and fragment culling. Memory-friendly strategies are further introduced to dynamically (a) lessen the wasteful memory allocation of individual pixels with low depth complexity frequencies, (b) minimize the allocated size of k-buffer according to different application goals and hardware limitations via a straightforward depth histogram analysis and (c) manage local GPU cache with a fixed-memory depth-sorting mechanism. Finally, an extensive experimental evaluation is provided demonstrating the advantages of our work over all prior k-buffer variants in terms of memory usage, performance cost and image quality.

Z‐Thickness Blending: Effective Fragment Merging for Multi‐Fragment Rendering

“…• The k + -buffer [VPF15] is an efficient k-buffer approach that uses both static and dynamic k values with an optimal GPU implementation. In our experiments, we employed the static version of the k + -buffer (SKB).…”

“…In certain applications where only limited memory resources are allowed, bounded memory must be used throughout the entire rendering pipeline. For example, general k ‐buffer algorithms [BCL∗07, MCTB11, VPF15] use the bounded memory that stores k fragments. The k ‐buffer algorithm can efficiently select the k ‐foremost fragments among the unsorted fragments that concurrently enter the same pixel using the GPU built‐in atomic operators during the store pass.…”

“…Several studies have developed many k ‐buffer variants [MCTB11] by performing multiple passes [LWXW09] or exploiting the GPU‐accelerated pixel synchronization operations [Sal13]. Vasilakis et al [VPF15] developed an efficient and memory‐friendly k ‐buffer algorithm called k + ‐buffer , which uses a dynamic k value decision scheme and GPU‐optimized implementation techniques.…”

“…Specifically, this technique performs two rendering processes in a common graphics processing units (GPU) pipeline (i.e., raster‐based graphics), including the storage of three‐dimensional (3D) geometric information into framebuffers and the generation of final images by exploiting the stored geometric information (i.e., screen‐space information). The memory access capabilities of modern graphics hardware allow multi‐fragment rendering to utilize screen‐space geometric information, which helps increase the demand for screen‐space or deferred rendering algorithms [KG13, MY18] in real‐time such as order‐independent‐transparency (OIT) [VPF15, MKKP18], dynamic photorealistic rendering techniques [BKKB13, MMNL14, FHSS18], and interactive scientific visualizations (e.g., multivariate surface data representations [RASS16] or hybrid data representations for the fusion of multiple volumetric and potentially transparent surface data [LFS∗15]).…”

Efficient Illumination Algorithms for Global Illumination In Interactive and Real-Time Rendering

“…To this end, various k-buffer variants have been proposed in the literature, exploiting stencil routing operations [BM08a], hardware-dependent spin-lock mechanisms [VF14], or even employing both a kand an A-buffer for the purposes of hair rendering [Yu+12]. We refer readers to the recent work of Vasilakis et al [VPF15] for a more general overview in k-buffer rendering.…”