Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Check out the TRIPS/EDGE architecture: http://www.cs.utexas.edu/~trips/overview.html

It uses dataflow to efficiently execute individual instructions while atomically committing groups of instructions to preserve understandable semantics. While tiled architectures are common now, most of them have one core per tile. TRIPS built a large, high-performance core out of simple tiles so it can run different types of code efficiently.



That link is missing their evaluation of TRIPS, which they did in 2008 after they built the chips: http://www.cs.utexas.edu/~bmaher/pubs/asplos09.pdf

My impression is that TRIPS is much too complicated to be clocked at comparable frequencies to commercial processors. Their ISA also appears very hard to compile to.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: