Single instruction, multiple threads: Difference between revisions

Content deleted Content added
Vortex GPU: add Jeff busy nyuzi
Tags: Mobile edit Mobile web edit Advanced mobile edit
Nyuzi GPGPU: add nyuzi performance analysis link
Tags: Mobile edit Mobile web edit Advanced mobile edit
Line 76:
=== Nyuzi GPGPU ===
 
Nyuzi is not SIMT: it is worth listing for comparison. Nyuzi implemented a [[barrel processor]] strategy over a SIMT one.<ref>http://www.cs.binghamton.edu/~millerti/nyami-ispass2015.pdf</ref> The project is noteworthy for learning and adapting as it progressed, providing comprehensive documentation and research papers.<ref>https://github.com/jbush001/NyuziProcessor/wiki</ref> as well as valuable insights into performance analysis.<ref>https://github.com/jbush001/NyuziProcessor/wiki/Performance-Analysis</ref> The floating-point pipelines are Predicated [[SIMD]].
 
== See also ==