- the paper presents a new technique that allows a massive number of dynamic lights to be rendered in real-time
- the presented implementation shows 3.4 million dynamic, emissive triangles in under 50 ms
- this is archived by reusing information from spatially and temporally adjacent pixels to filter probabilities to guide rays
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/spatiotemporal.png)
- AMD virtual event that contains talks about
- Ryzen™ Processor Software Optimization
- optimizing for Radeon™ RDNA Architecture
- From Source to ISA: A Trip Down the Shader Compiler Pipeline
- A Review of GPUOpen Effects
- Curing Amnesia and Other GPU Maladies With AMD Developer Tools
- Radeon™ ProRender Full Spectrum Rendering 2.0: The Universal Rendering API
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/lets_build.png)
- overview of what Constructive Solid Geometry (CSG) is and workflows
- presents the algorithm that allows for iterative updates including how to calculate intersections and mesh generation
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/csg.png)
- the post provides an overview of the new AMD memory profiler for Vulkan and D3D12
- this tool allows gaining a deeper understanding of memory usage
- shows all allocations, tracked on a timeline, will enable comparisons of snapshots and gives warnings about possible problems
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/allocations.png)
- the article shows how to calculate the shadows from the sun
- presents the results of the shadow quality with white and blue noise
- the implementation is then extended to Spherical Positional and spotlights
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/spot.png)
- the article presents how using blue noise can improve ray marching results
- provides a shadertoy that compares against white noise
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/stochasticfog.png)
- the author presents his ideas about how the Nanite technology in the UE5 tech demo might be implemented
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/unreal-engine-5.jpg)
- the article that discusses possible ideas on how the UE5 nanite technology might be implemented
- contains links to exciting techniques that might be related, or inspired the development
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/unreal-engine-5-cover-image.jpg)
- an overview of the new Nvidia Ampere Architecture aimed at compute workloads
- new tensor cores for f32 data, IEE compliant fp64, and BF16 throughput improvements
- adds the ability to run multiple processes on the same GPU with error isolation
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/GA100SM.png)
- overview of new features in CUDA 11 and how the NVIDIA Ampere GPU hardware features are exposed
![](https://www.jendrikillner.com/img/posts/graphics-programming-weekly-132/cuda-11-1.jpg)
Thanks to Aras Pranckevičius for support of this series.
Would you like to see your name here too? Become a Patreon of this series.