Running a Fragment Shader on a GPU Core.
The article discusses the efficient execution of fragment shrader on a single-core graphics processing unit (GPU), detailing shrader compilation to SIMD (single instruction, multiple date) instruction sequences. Shrader compilation by the graphics driver produces a GPU binary from high-level fragmen...
| Publicado en: | Communications of the ACM Vol. 51; no. 10; p. 55 |
|---|---|
| Formato: | Artículo |
| Publicado: |
Association for Computing Machinery
Oct2008
|
| Materias: | |
| Acceso en línea: | Ver este registro en EBSCOhost |
| Sumario: | The article discusses the efficient execution of fragment shrader on a single-core graphics processing unit (GPU), detailing shrader compilation to SIMD (single instruction, multiple date) instruction sequences. Shrader compilation by the graphics driver produces a GPU binary from high-level fragment shrader source, the article states. Topics include four vector registers, GPUs executing a copy of the shrader binary on each of the four thread contexts, and a pool of general-purpose vector registers. |
|---|