Visibility Rendering Order: Improving Energy Efficiency on Mobile GPUs through Frame Coherence

Abstract

During real-time graphics rendering, objects are processed by the GPU in the order they are submitted by the CPU, and occluded surfaces are often processed even though they will end up not being part of the final image, thus wasting precious time and energy. To help discard occluded surfaces, most current GPUs include an Early-Depth test before the fragment processing stage. However, to be effective it requires that opaque objects are processed in a front-to-back order. Depth sorting and other occlusion culling techniques at the object level incur overheads that are only offset for applications having substantial depth and/or fragment shading complexity, which is often not the case in mobile workloads. We propose a novel architectural technique for GPUs, Visibility Rendering Order (VRO), which reorders objects front-to-back entirely in hardware by exploiting the fact that the objects in graphics animated applications tend to keep its relative depth order across consecutive frames (temporal coherence). Since order relationships are already tested by the Depth Test, VRO incurs minimal energy overheads because it just requires adding a small hardware to capture that information and use it later to guide the rendering of the following frame. Moreover, unlike other approaches, this unit works in parallel with the graphics pipeline without any performance overhead. We illustrate the benefits of VRO using various unmodified commercial 3D applications for which VRO achieves 27% speed-up and 15.8% energy reduction on average over a state-of-the-art mobile GPU.

Omega-Test: A Predictive Early-Z Culling to Improve the Graphics Pipeline Energy-Efficiency

“…This utilization of a Deferred Rendering scheme in a TBR pipeline is known as Tile-Based Deferred Rendering (TBDR). Recent academic work explored different alternatives of TBDR [8]. The main drawback of the best performing alternative is that it rasterizes all primitives twice, and so fragments are processed twice as well: once for calculating the depths in the visibility determination phase, and again for calculating the rest of attributes to continue down the pipeline.…”

“…Contrarily, our proposal does not introduce timing overheads in the pipeline, and the added structures (Ω-Table, E-Buffer, correction queues; see Section 3) are small, on-chip buffers. As a reference comparison point, in [8] it is proposed Visibility Rendering Order (VRO), a technique that attacks overdraw by sorting objects in a front-to-back order, where authors quantitatively show that VRO outperforms TBDR. We have compared Ω-Test with VRO (see Section 6) showing that our approach beats VRO.…”

“…In particular, fragments that appear behind others (for a given camera viewpoint) are not visible in the final scene. The solution to the visibility problem is not unique and multiple approaches can be found in the literature [8], [9], [10] being the so-called Depth Test (or Z-Test) [11], which performs a visibility test at the pixel granularity, the most widely implemented technique in contemporary GPUs.…”

Mobile Augmented Reality for Low-End Devices Based on Planar Surface Recognition and Optimized Vertex Data Rendering

“…A performance amelioration strategy is the configuration of the number ration of the shading processing units which substantially balances the load [38]. Occlusion or visibility culling methods drastically reduce redundant fragment processing computations by avoiding the processing of occluded surfaces of the 3D model, but are not suitable for low-power devices [39,40].…”

A Parallel Divide-and-Conquer-Based Evolutionary Algorithm for Large-Scale Optimization

“…Specifically, the parallelism of NPDC can be quantified in terms of the speed-up [55]. The speed-up is a key indicator for measuring parallel efficiency [56].…”