Explore >> Select a destination
|
You are here 
|
www.reedbeta.com | ||
| | | | |
fgiesen.wordpress.com
|
|
| | | | | This post is part of the series "A trip through the Graphics Pipeline 2011". Welcome back! This post deals with the second half of pixel processing, the "join phase". The previous phase was all about taking a small number of input streams and turning them into lots of independent tasks for the shader units. Now... | |
| | | | |
timur.hu
|
|
| | | | | Mesh and task shaders (amplification shaders in D3D jargon) are a new way to process geometry in 3D applications. First proposed by NVidia in 2018 and initially available in the "Turing" series, they are now supported on RDNA2 GPUs and are part of the D3D12 API. There is also an extension in Vulkan (and a vendor-specific one in OpenGL). This post is about what mesh shadig is and next time I'm going to talk about how mesh/task shaders are implemented on the driver side. | |
| | | | |
raphlinus.github.io
|
|
| | | | | The GPU in your computer is about 10 to 100 times more powerful than the CPU, depending on workload. For real-time graphics rendering and machine learning, you are enjoying that power, and doing those workloads on a CPU is not viable. Why aren't we exploiting that power for other workloads? What prevents a GPU from being a more general purpose computer? | |
| | | | |
logins.github.io
|
|
| | | Why Talking about DX12 Resource Handling Among the many changes that happened in the transition between Direct3D version 11 and 12, the idea of shader resources has been reworked quite a lot and the process to handle resources has become much more granular than before. D3D12 introduces a new binding model for resources compared to D3D11. Objects called resource views are created to reference buffer resources to the graphics pipeline, but now all the operations that were happening when binding a resource in D3D11 (e.g. make it resident in GPU memory, lifetime in memory checks, state handling, CPU-to-GPU memory mapping checks) are now fragmented in different calls to the SDK so that we can execute just a subset of them depending on the current needs. I talked ... | ||
