首页> 外文期刊>ACM Transactions on Graphics >Automatic Shader Simplification Using Surface Signal Approximation
【24h】

Automatic Shader Simplification Using Surface Signal Approximation

机译:使用表面信号逼近自动简化着色器

获取原文
获取原文并翻译 | 示例
           

摘要

In this paper, we present a new automatic shader simplificationrnmethod using surface signal approximation. We regard the entirernmulti-stage rendering pipeline as a process that generates signalsrnon surfaces, and we formulate the simplification of the fragmentrnshader as a global simplification problem across multi-shaderrnstages. Three new shader simplification rules are proposed to solvernthe problem. First, the code transformation rule transforms fragmentrnshader code to other shader stages in order to redistribute computationsrnon pixels up to the level of geometry primitives. Second,rnthe surface-wise approximation rule uses high-order polynomial basisrnfunctions on surfaces to approximate pixel-wise computations inrnthe fragment shader. These approximations are pre-cached and simplifyrncomputations at runtime. Third, the surface subdivision rulerntessellates surfaces into smaller patches. It combines with the previousrntwo rules to approximate pixel-wise signals at different levelsrnof tessellations with different computation times and visual errors.rnTo evaluate simplified shaders using these simplification rules, wernintroduce a new cost model that includes the visual quality, renderingrntime and memory consumption. With these simplification rulesrnand the cost model, we present an integrated shader simplificationrnalgorithm that is capable of automatically generating variants ofrnsimplified shaders and selecting a sequence of preferable shaders.rnResults show that the sequence of selected simplified shaders balancernperformance, accuracy and memory consumption well.
机译:在本文中,我们提出了一种使用表面信号逼近的新的自动着色器简化方法。我们将整个多阶段渲染管线视为生成信号表面的过程,并将片段着色器的简化公式化为跨多着色器阶段的全局简化问题。提出了三种新的着色器简化规则来解决该问题。首先,代码转换规则将片段着色器代码转换到其他着色器阶段,以便将计算像素重新分配到几何图元的级别。其次,逐面逼近规则使用曲面上的高阶多项式基函数来近似片段着色器中的逐象素计算。这些近似值已预先缓存,并在运行时简化了计算。第三,曲面细分规则将曲面细分为较小的小块。它与先前的两个规则相结合,可以在不同的计算时间和视觉误差下细分不同级别的像素信号。为了使用这些简化规则评估简化的着色器,我们引入了一个新的成本模型,其中包括视觉质量,渲染时间和内存消耗。利用这些简化规则和成本模型,我们提出了一种集成的着色器简化算法,该算法能够自动生成简化的着色器的变体并选择优先的着色器序列。结果表明,所选简化着色器的序列很好地平衡了性能,准确性和内存消耗。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号