Figure 1: (left) Our approach automatically identifies opportunities for placing surrogate rules ... more Figure 1: (left) Our approach automatically identifies opportunities for placing surrogate rules (α) in shape grammar derivation trees, either directly or only for partial subtrees (β). At runtime, surrogate rules can evaluate to simple impostors that replace detailed geometry in procedurally generated scenes. (center) Our approach uses view-dependent factors to decide on the applicability of impostors, achieving high visual fidelity (original/ours). (right) Wireframes indicate the significantly reduced geometry in the rendered scene (original/ours).
Proceedings of the 2018 International Conference on Supercomputing, 2018
Harnessing the power of massively parallel devices like the graphics processing unit (GPU) is dif... more Harnessing the power of massively parallel devices like the graphics processing unit (GPU) is difficult for algorithms that show dynamic or inhomogeneous workloads. To achieve high performance, such advanced algorithms require scalable, concurrent queues to collect and distribute work. We show that previous queuing approaches are unfit for this task, as they either (1) do not work well in a massively parallel environment, or (2) obstruct the use of individual threads on top of single-instruction-multiple-data (SIMD) cores, or (3) block during access, thus prohibiting multi-queue setups. With these issues in mind, we present the Broker Queue, a highly efficient, fully linearizable FIFO queue for fine-granular parallel work distribution on the GPU. We evaluate its performance and usability on modern GPU models against a wide range of existing algorithms. The Broker Queue is up to three orders of magnitude faster than non-blocking queues and can even outperform significantly simpler techniques that lack desired properties for fine-granular work distribution.
In this paper, we examine the challenges of implementing priority-based task management with resp... more In this paper, we examine the challenges of implementing priority-based task management with respect to userdefined preferential attributes on a graphics processing unit (GPU). Previous approaches usually rely on constant synchronization with the CPU to determine the proper chronological sequence for execution. We transfer the responsibility for evaluating and arranging planned tasks to the GPU where they are continuously processed in a persistent kernel. By implementing a dynamic work queue with segments of variable size, we introduce possibilities for gradually improving the overall order and identify necessary meta data required to avoid write-read conflicts in a massively parallel environment. We employ an autonomous controller module to allow queue management to run concurrently with task execution. We revise Batcher’s bitonic merge sort and show its eligibility for sorting partitioned work queues. The performance of the algorithm for tasks with constant execution time is evalu...
Symposium on Interactive 3D Graphics and Games, 2020
With the introduction of hardware-supported ray tracing and deep learning for denoising, computer... more With the introduction of hardware-supported ray tracing and deep learning for denoising, computer graphics has made a considerable step toward real-time global illumination. In this work, we present an alternative global illumination method: The stochastic substitute tree (SST), a hierarchical structure inspired by lightcuts with light probability distributions as inner nodes. Our approach distributes virtual point lights (VPLs) in every frame and efficiently constructs the SST over those lights by clustering according to Morton codes. Global illumination is approximated by sampling the SST and considers the BRDF at the hit location as well as the SST nodes’ intensities for importance sampling directly from inner nodes of the tree. To remove the introduced Monte Carlo noise, we use a recurrent autoencoder. In combination with temporal filtering, we deliver real-time global illumination for complex scenes with challenging light distributions.
In this paper, we present several compute‐based point cloud rendering approaches that outperform ... more In this paper, we present several compute‐based point cloud rendering approaches that outperform the hardware pipeline by up to an order of magnitude and achieve significantly better frame times than previous compute‐based methods. Beyond basic closest‐point rendering, we also introduce a fast, high‐quality variant to reduce aliasing. We present and evaluate several variants of our proposed methods with different flavors of optimization, in order to ensure their applicability and achieve optimal performance on a range of platforms and architectures with varying support for novel GPU hardware features. During our experiments, the observed peak performance was reached rendering 796 million points (12.7GB) at rates of 62 to 64 frames per second (50 billion points per second, 802GB/s) on an RTX 3090 without the use of level‐of‐detail structures. We further introduce an optimized vertex order for point clouds to boost the efficiency of GL_POINTS by a factor of 5× in cases where hardware ...
Efficient rendering of multiple views can be a critical performance factor for real-time renderin... more Efficient rendering of multiple views can be a critical performance factor for real-time rendering applications. Generating more than one view multiplies the amount of rendered geometry, which can cause a huge performance impact. Minimizing that impact has been a target of previous research and GPU manufacturers, who have started to equip devices with dedicated acceleration units. However, vendor-specific acceleration is not the only option to increase multi-view rendering (MVR) performance. Available graphics API features, shader stages and optimizations can be exploited for improved MVR performance, while generally offering more versatile pipeline configurations, including the preservation of custom tessellation and geometry shaders. In this paper, we present an exhaustive evaluation of MVR pipelines available on modern GPUs. We provide a detailed analysis of previous techniques, hardware-accelerated MVR and propose a novel method, leading to the creation of an MVR catalogue. Our ...
Following recent advances in GPU hardware development and newly introduced rendering pipeline ext... more Following recent advances in GPU hardware development and newly introduced rendering pipeline extensions, the segmentation of input geometry into small geometry clusters-so-called meshlets-has emerged as an important practice for efficient rendering of complex 3D models. Meshlets can be processed efficiently using mesh shaders on modern graphics processing units, in order to achieve streamlined geometry processing in just two tightly coupled shader stages that allow for dynamic workload manipulation in-between. The additional granularity layer between entire models and individual triangles enables new opportunities for fine-grained visibility culling methods. However, in contrast to static models, view frustum and backface culling on a per-meshlet basis for skinned, animated models are difficult to achieve while respecting the conservative spatio-temporal bounds that are required for robust rendering results. In this paper, we describe a solution for computing and exploiting relevant conservative bounds for culling meshlets of models that are animated using linear blend skinning. By enabling visibility culling for animated meshlets, our approach can help to improve rendering performance and alleviate bottlenecks in the notoriously performanceand memory-intensive skeletal animation pipelines of modern real-time graphics applications.
To e ectively utilize an ever increasing number of processors during parallel rendering, hardware... more To e ectively utilize an ever increasing number of processors during parallel rendering, hardware and so ware designers rely on sophisticated load balancing strategies. While dynamic load balancing is a powerful solution, it requires complex work distribution and synchronization mechanisms. Graphics hardware manufacturers have opted to employ static load balancing strategies instead. Speci cally, triangle data is distributed to processors based on its overlap with screenspace tiles arranged in a xed pa ern. While the current strategy of using simple pa erns for a small number of fast rasterizers achieves formidable performance, it is questionable how this approach will scale as the number of processors increases further. To address this issue, we analyze real-world rendering workloads, derive requirements for e ective pa erns, and present ten di erent pa ern design strategies based on these requirements. In addition to a theoretical evaluation of these design strategies, we compare the performance of select pa erns in a parallel sort-middle so ware rendering pipeline on an extensive set of triangle data captured from eight recent video games. As a result, we are able to identify a set of pa erns that scale well and exhibit signi cantly improved performance over naïve approaches.
Proceedings of the ACM on Computer Graphics and Interactive Techniques, 2018
Due to its flexibility, compute mode is becoming more and more attractive as a way to implement m... more Due to its flexibility, compute mode is becoming more and more attractive as a way to implement many of the algorithms part of a state-of-the-art rendering pipeline. A key problem commonly encountered in graphics applications is streaming vertex and geometry processing. In a typical triangle mesh, the same vertex is on average referenced six times. To avoid redundant computation during rendering, a post-transform cache is traditionally employed to reuse vertex processing results. However, such a vertex cache can generally not be implemented efficiently in software and does not scale well as parallelism increases. We explore alternative strategies for reusing per-vertex results on-the-fly during massively-parallel software geometry processing. Given an input stream divided into batches, we analyze the effectiveness of sorting, hashing, and intra-thread-group communication for identifying and exploiting local reuse potential. We design and present four vertex reuse strategies tailored...
Proceedings of the ACM on Computer Graphics and Interactive Techniques, 2018
In this paper, we question the premise that graphics hardware uses a post-transform cache to avoi... more In this paper, we question the premise that graphics hardware uses a post-transform cache to avoid redundant vertex shader invocations. A large body of existing work on optimizing indexed triangle sets for rendering speed is based upon this widely-accepted assumption. We conclusively show that this assumption does not hold up on modern graphics hardware. We design and conduct experiments that demonstrate the behavior of current hardware of all major vendors to be inconsistent with the presence of a common post-transform cache. Our results strongly suggest that modern hardware rather relies on a batch-based approach, most likely for reasons of scalability. A more thorough investigation based on these initial experiments allows us to partially uncover the actual strategies implemented on graphics processors today. We reevaluate existing mesh optimization algorithms in light of these new findings and present a new mesh optimization algorithm designed from the ground up to target archit...
In this paper, we present a real-time graphics pipeline implemented entirely in software on a mod... more In this paper, we present a real-time graphics pipeline implemented entirely in software on a modern GPU. As opposed to previous work, our approach features a fully-concurrent, multi-stage, streaming design with dynamic load balancing, capable of operating efficiently within bounded memory. We address issues such as primitive order, vertex reuse, and screen-space derivatives of dependent variables, which are essential to real-world applications, but have largely been ignored by comparable work in the past. The power of a software approach lies in the ability to tailor the graphics pipeline to any given application. In exploration of this potential, we design and implement four novel pipeline modifications. Evaluation of the performance of our approach on more than 100 real-world scenes collected from video games shows rendering speeds within one order of magnitude of the hardware graphics pipeline as well as significant improvements over previous work, not only in terms of capabilit...
Harnessing the power of massively parallel devices like the graphics processing unit (GPU) is dif... more Harnessing the power of massively parallel devices like the graphics processing unit (GPU) is difficult for algorithms that show dynamic or inhomogeneous workloads. To achieve high performance, suc...
Figure 1: (left) Hierarchical bucket queues enable the implementation of foveated, Reyes-style mi... more Figure 1: (left) Hierarchical bucket queues enable the implementation of foveated, Reyes-style micropolygon rendering in real-time on the GPU. By dynamically reducing the number of split recursions in areas distant from the user's current fixation point (red), rendering quality can adaptively be decreased where not needed, leading to speedups of more than 100%. (right) Using prioritization in a GPU path tracer, compute power can be directed primarily to areas of high variance to achieve faster convergence. The overlay visualizes the number of traced paths.
Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, 2015
{ mohr | kerbl | donoser | schmalstieg | kalkofen }@icg.tugraz.at (a) (c) (d) Figure 1. Retargete... more { mohr | kerbl | donoser | schmalstieg | kalkofen }@icg.tugraz.at (a) (c) (d) Figure 1. Retargeted 2D documentation. (a) The input documentation consists of an annotated explosion diagram and a sequence of images presenting disassembly instructions. (b) From analyzing the 2D explosion diagram, our system is able to generate a 3D explosion diagram presented in AR. Moreover, our system generates 3D annotations in AR based on the input 2D documentation. (c) In addition, our system is able to analyze image sequences in order to create 3D animations from it. This allows to present animated 3D documentations in AR. Here we show six key-frames from the resulting AR animation.
In this paper, we examine the challenges of implement-ing priority-based task management with res... more In this paper, we examine the challenges of implement-ing priority-based task management with respect to user-defined preferential attributes on a graphics processing unit (GPU). Previous approaches usually rely on constant synchronization with the CPU to determine the proper chronological sequence for execution. We transfer the re-sponsibility for evaluating and arranging planned tasks to the GPU where they are continuously processed in a per-sistent kernel. By implementing a dynamic work queue with segments of variable size, we introduce possibili-ties for gradually improving the overall order and iden-tify necessary meta data required to avoid write-read con-flicts in a massively parallel environment. We employ an autonomous controller module to allow queue manage-ment to run concurrently with task execution. We revise Batcher's bitonic merge sort and show its eligibility for sorting partitioned work queues. The performance of the algorithm for tasks with constant execution t...
In this paper, we present Whippletree, a novel approach to scheduling dynamic, irregular workload... more In this paper, we present Whippletree, a novel approach to scheduling dynamic, irregular workloads on the GPU. We introduce a new programming model which offers the simplicity and expressiveness of task-based parallelism while retaining all aspects of the multi-level execution hierarchy essential to unlocking the full potential of a modern GPU. At the same time, our programming model lends itself to efficient implementation on the SIMD-based architecture typical of a current GPU. We demonstrate the practical utility of our model by providing a reference implementation on top of current CUDA hardware. Furthermore, we show that our model compares favorably to traditional approaches in terms of both performance as well as the range of applications that can be covered. We demonstrate the benefits of our model for recursive Reyes rendering, procedural geometry generation and volume rendering with concurrent irradiance caching.
Figure 1: Key frames from a disassembly sequence for the Leaning Tower CAD model. Our approach id... more Figure 1: Key frames from a disassembly sequence for the Leaning Tower CAD model. Our approach identifies reusable blocking information and evaluates the geometric feasibility of part removals in parallel. Thus, we are able to create disassembly sequences for models such as this 4,193 part assembly within minutes instead of days. Furthermore, unstable configurations are avoided by using structural constraints to elevate reproducibility on real-world objects. We facilitate the visual assessment of sequences by combining previewing techniques with visual cues to highlight smaller details.
We present a software solution for planning and simulating radio-frequency ablation (RFA) treatme... more We present a software solution for planning and simulating radio-frequency ablation (RFA) treatment for patients suffering from hepatocellular carcinoma. Our software provides the graphical front-end for the results of the EU FP7 project IMPPACT. The main planning application was designed to assist with the identification of minimum-risk setups for RFA probes and generation of evaluable 3D representations of the predicted necrosis zones. Patient-specific mesh data describing the involved anatomic structures are used to individually parameterize the simulation procedure for personalized results. Our software supplies tools for extracting the required anatomic meshes from computed tomography (CT) images and for specifying custom intervention protocols. Data collected during simulations allow for detailed illustration of expected effectiveness and progress of heat-induced necrosis over time. Our software was evaluated positively by medical personnel and has been successfully put into practice at two independent European clinical sites.
Figure 1: (left) Our approach automatically identifies opportunities for placing surrogate rules ... more Figure 1: (left) Our approach automatically identifies opportunities for placing surrogate rules (α) in shape grammar derivation trees, either directly or only for partial subtrees (β). At runtime, surrogate rules can evaluate to simple impostors that replace detailed geometry in procedurally generated scenes. (center) Our approach uses view-dependent factors to decide on the applicability of impostors, achieving high visual fidelity (original/ours). (right) Wireframes indicate the significantly reduced geometry in the rendered scene (original/ours).
Proceedings of the 2018 International Conference on Supercomputing, 2018
Harnessing the power of massively parallel devices like the graphics processing unit (GPU) is dif... more Harnessing the power of massively parallel devices like the graphics processing unit (GPU) is difficult for algorithms that show dynamic or inhomogeneous workloads. To achieve high performance, such advanced algorithms require scalable, concurrent queues to collect and distribute work. We show that previous queuing approaches are unfit for this task, as they either (1) do not work well in a massively parallel environment, or (2) obstruct the use of individual threads on top of single-instruction-multiple-data (SIMD) cores, or (3) block during access, thus prohibiting multi-queue setups. With these issues in mind, we present the Broker Queue, a highly efficient, fully linearizable FIFO queue for fine-granular parallel work distribution on the GPU. We evaluate its performance and usability on modern GPU models against a wide range of existing algorithms. The Broker Queue is up to three orders of magnitude faster than non-blocking queues and can even outperform significantly simpler techniques that lack desired properties for fine-granular work distribution.
In this paper, we examine the challenges of implementing priority-based task management with resp... more In this paper, we examine the challenges of implementing priority-based task management with respect to userdefined preferential attributes on a graphics processing unit (GPU). Previous approaches usually rely on constant synchronization with the CPU to determine the proper chronological sequence for execution. We transfer the responsibility for evaluating and arranging planned tasks to the GPU where they are continuously processed in a persistent kernel. By implementing a dynamic work queue with segments of variable size, we introduce possibilities for gradually improving the overall order and identify necessary meta data required to avoid write-read conflicts in a massively parallel environment. We employ an autonomous controller module to allow queue management to run concurrently with task execution. We revise Batcher’s bitonic merge sort and show its eligibility for sorting partitioned work queues. The performance of the algorithm for tasks with constant execution time is evalu...
Symposium on Interactive 3D Graphics and Games, 2020
With the introduction of hardware-supported ray tracing and deep learning for denoising, computer... more With the introduction of hardware-supported ray tracing and deep learning for denoising, computer graphics has made a considerable step toward real-time global illumination. In this work, we present an alternative global illumination method: The stochastic substitute tree (SST), a hierarchical structure inspired by lightcuts with light probability distributions as inner nodes. Our approach distributes virtual point lights (VPLs) in every frame and efficiently constructs the SST over those lights by clustering according to Morton codes. Global illumination is approximated by sampling the SST and considers the BRDF at the hit location as well as the SST nodes’ intensities for importance sampling directly from inner nodes of the tree. To remove the introduced Monte Carlo noise, we use a recurrent autoencoder. In combination with temporal filtering, we deliver real-time global illumination for complex scenes with challenging light distributions.
In this paper, we present several compute‐based point cloud rendering approaches that outperform ... more In this paper, we present several compute‐based point cloud rendering approaches that outperform the hardware pipeline by up to an order of magnitude and achieve significantly better frame times than previous compute‐based methods. Beyond basic closest‐point rendering, we also introduce a fast, high‐quality variant to reduce aliasing. We present and evaluate several variants of our proposed methods with different flavors of optimization, in order to ensure their applicability and achieve optimal performance on a range of platforms and architectures with varying support for novel GPU hardware features. During our experiments, the observed peak performance was reached rendering 796 million points (12.7GB) at rates of 62 to 64 frames per second (50 billion points per second, 802GB/s) on an RTX 3090 without the use of level‐of‐detail structures. We further introduce an optimized vertex order for point clouds to boost the efficiency of GL_POINTS by a factor of 5× in cases where hardware ...
Efficient rendering of multiple views can be a critical performance factor for real-time renderin... more Efficient rendering of multiple views can be a critical performance factor for real-time rendering applications. Generating more than one view multiplies the amount of rendered geometry, which can cause a huge performance impact. Minimizing that impact has been a target of previous research and GPU manufacturers, who have started to equip devices with dedicated acceleration units. However, vendor-specific acceleration is not the only option to increase multi-view rendering (MVR) performance. Available graphics API features, shader stages and optimizations can be exploited for improved MVR performance, while generally offering more versatile pipeline configurations, including the preservation of custom tessellation and geometry shaders. In this paper, we present an exhaustive evaluation of MVR pipelines available on modern GPUs. We provide a detailed analysis of previous techniques, hardware-accelerated MVR and propose a novel method, leading to the creation of an MVR catalogue. Our ...
Following recent advances in GPU hardware development and newly introduced rendering pipeline ext... more Following recent advances in GPU hardware development and newly introduced rendering pipeline extensions, the segmentation of input geometry into small geometry clusters-so-called meshlets-has emerged as an important practice for efficient rendering of complex 3D models. Meshlets can be processed efficiently using mesh shaders on modern graphics processing units, in order to achieve streamlined geometry processing in just two tightly coupled shader stages that allow for dynamic workload manipulation in-between. The additional granularity layer between entire models and individual triangles enables new opportunities for fine-grained visibility culling methods. However, in contrast to static models, view frustum and backface culling on a per-meshlet basis for skinned, animated models are difficult to achieve while respecting the conservative spatio-temporal bounds that are required for robust rendering results. In this paper, we describe a solution for computing and exploiting relevant conservative bounds for culling meshlets of models that are animated using linear blend skinning. By enabling visibility culling for animated meshlets, our approach can help to improve rendering performance and alleviate bottlenecks in the notoriously performanceand memory-intensive skeletal animation pipelines of modern real-time graphics applications.
To e ectively utilize an ever increasing number of processors during parallel rendering, hardware... more To e ectively utilize an ever increasing number of processors during parallel rendering, hardware and so ware designers rely on sophisticated load balancing strategies. While dynamic load balancing is a powerful solution, it requires complex work distribution and synchronization mechanisms. Graphics hardware manufacturers have opted to employ static load balancing strategies instead. Speci cally, triangle data is distributed to processors based on its overlap with screenspace tiles arranged in a xed pa ern. While the current strategy of using simple pa erns for a small number of fast rasterizers achieves formidable performance, it is questionable how this approach will scale as the number of processors increases further. To address this issue, we analyze real-world rendering workloads, derive requirements for e ective pa erns, and present ten di erent pa ern design strategies based on these requirements. In addition to a theoretical evaluation of these design strategies, we compare the performance of select pa erns in a parallel sort-middle so ware rendering pipeline on an extensive set of triangle data captured from eight recent video games. As a result, we are able to identify a set of pa erns that scale well and exhibit signi cantly improved performance over naïve approaches.
Proceedings of the ACM on Computer Graphics and Interactive Techniques, 2018
Due to its flexibility, compute mode is becoming more and more attractive as a way to implement m... more Due to its flexibility, compute mode is becoming more and more attractive as a way to implement many of the algorithms part of a state-of-the-art rendering pipeline. A key problem commonly encountered in graphics applications is streaming vertex and geometry processing. In a typical triangle mesh, the same vertex is on average referenced six times. To avoid redundant computation during rendering, a post-transform cache is traditionally employed to reuse vertex processing results. However, such a vertex cache can generally not be implemented efficiently in software and does not scale well as parallelism increases. We explore alternative strategies for reusing per-vertex results on-the-fly during massively-parallel software geometry processing. Given an input stream divided into batches, we analyze the effectiveness of sorting, hashing, and intra-thread-group communication for identifying and exploiting local reuse potential. We design and present four vertex reuse strategies tailored...
Proceedings of the ACM on Computer Graphics and Interactive Techniques, 2018
In this paper, we question the premise that graphics hardware uses a post-transform cache to avoi... more In this paper, we question the premise that graphics hardware uses a post-transform cache to avoid redundant vertex shader invocations. A large body of existing work on optimizing indexed triangle sets for rendering speed is based upon this widely-accepted assumption. We conclusively show that this assumption does not hold up on modern graphics hardware. We design and conduct experiments that demonstrate the behavior of current hardware of all major vendors to be inconsistent with the presence of a common post-transform cache. Our results strongly suggest that modern hardware rather relies on a batch-based approach, most likely for reasons of scalability. A more thorough investigation based on these initial experiments allows us to partially uncover the actual strategies implemented on graphics processors today. We reevaluate existing mesh optimization algorithms in light of these new findings and present a new mesh optimization algorithm designed from the ground up to target archit...
In this paper, we present a real-time graphics pipeline implemented entirely in software on a mod... more In this paper, we present a real-time graphics pipeline implemented entirely in software on a modern GPU. As opposed to previous work, our approach features a fully-concurrent, multi-stage, streaming design with dynamic load balancing, capable of operating efficiently within bounded memory. We address issues such as primitive order, vertex reuse, and screen-space derivatives of dependent variables, which are essential to real-world applications, but have largely been ignored by comparable work in the past. The power of a software approach lies in the ability to tailor the graphics pipeline to any given application. In exploration of this potential, we design and implement four novel pipeline modifications. Evaluation of the performance of our approach on more than 100 real-world scenes collected from video games shows rendering speeds within one order of magnitude of the hardware graphics pipeline as well as significant improvements over previous work, not only in terms of capabilit...
Harnessing the power of massively parallel devices like the graphics processing unit (GPU) is dif... more Harnessing the power of massively parallel devices like the graphics processing unit (GPU) is difficult for algorithms that show dynamic or inhomogeneous workloads. To achieve high performance, suc...
Figure 1: (left) Hierarchical bucket queues enable the implementation of foveated, Reyes-style mi... more Figure 1: (left) Hierarchical bucket queues enable the implementation of foveated, Reyes-style micropolygon rendering in real-time on the GPU. By dynamically reducing the number of split recursions in areas distant from the user's current fixation point (red), rendering quality can adaptively be decreased where not needed, leading to speedups of more than 100%. (right) Using prioritization in a GPU path tracer, compute power can be directed primarily to areas of high variance to achieve faster convergence. The overlay visualizes the number of traced paths.
Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, 2015
{ mohr | kerbl | donoser | schmalstieg | kalkofen }@icg.tugraz.at (a) (c) (d) Figure 1. Retargete... more { mohr | kerbl | donoser | schmalstieg | kalkofen }@icg.tugraz.at (a) (c) (d) Figure 1. Retargeted 2D documentation. (a) The input documentation consists of an annotated explosion diagram and a sequence of images presenting disassembly instructions. (b) From analyzing the 2D explosion diagram, our system is able to generate a 3D explosion diagram presented in AR. Moreover, our system generates 3D annotations in AR based on the input 2D documentation. (c) In addition, our system is able to analyze image sequences in order to create 3D animations from it. This allows to present animated 3D documentations in AR. Here we show six key-frames from the resulting AR animation.
In this paper, we examine the challenges of implement-ing priority-based task management with res... more In this paper, we examine the challenges of implement-ing priority-based task management with respect to user-defined preferential attributes on a graphics processing unit (GPU). Previous approaches usually rely on constant synchronization with the CPU to determine the proper chronological sequence for execution. We transfer the re-sponsibility for evaluating and arranging planned tasks to the GPU where they are continuously processed in a per-sistent kernel. By implementing a dynamic work queue with segments of variable size, we introduce possibili-ties for gradually improving the overall order and iden-tify necessary meta data required to avoid write-read con-flicts in a massively parallel environment. We employ an autonomous controller module to allow queue manage-ment to run concurrently with task execution. We revise Batcher's bitonic merge sort and show its eligibility for sorting partitioned work queues. The performance of the algorithm for tasks with constant execution t...
In this paper, we present Whippletree, a novel approach to scheduling dynamic, irregular workload... more In this paper, we present Whippletree, a novel approach to scheduling dynamic, irregular workloads on the GPU. We introduce a new programming model which offers the simplicity and expressiveness of task-based parallelism while retaining all aspects of the multi-level execution hierarchy essential to unlocking the full potential of a modern GPU. At the same time, our programming model lends itself to efficient implementation on the SIMD-based architecture typical of a current GPU. We demonstrate the practical utility of our model by providing a reference implementation on top of current CUDA hardware. Furthermore, we show that our model compares favorably to traditional approaches in terms of both performance as well as the range of applications that can be covered. We demonstrate the benefits of our model for recursive Reyes rendering, procedural geometry generation and volume rendering with concurrent irradiance caching.
Figure 1: Key frames from a disassembly sequence for the Leaning Tower CAD model. Our approach id... more Figure 1: Key frames from a disassembly sequence for the Leaning Tower CAD model. Our approach identifies reusable blocking information and evaluates the geometric feasibility of part removals in parallel. Thus, we are able to create disassembly sequences for models such as this 4,193 part assembly within minutes instead of days. Furthermore, unstable configurations are avoided by using structural constraints to elevate reproducibility on real-world objects. We facilitate the visual assessment of sequences by combining previewing techniques with visual cues to highlight smaller details.
We present a software solution for planning and simulating radio-frequency ablation (RFA) treatme... more We present a software solution for planning and simulating radio-frequency ablation (RFA) treatment for patients suffering from hepatocellular carcinoma. Our software provides the graphical front-end for the results of the EU FP7 project IMPPACT. The main planning application was designed to assist with the identification of minimum-risk setups for RFA probes and generation of evaluable 3D representations of the predicted necrosis zones. Patient-specific mesh data describing the involved anatomic structures are used to individually parameterize the simulation procedure for personalized results. Our software supplies tools for extracting the required anatomic meshes from computed tomography (CT) images and for specifying custom intervention protocols. Data collected during simulations allow for detailed illustration of expected effectiveness and progress of heat-induced necrosis over time. Our software was evaluated positively by medical personnel and has been successfully put into practice at two independent European clinical sites.
Uploads
Papers by Bernhard Kerbl