HN2new | past | comments | ask | show | jobs | submitlogin

Surely auto-parallelization is exactly what CPU pipelining and branch prediction do quite effectively?


Well, sort of, but then the moment you run into hazards [1], bad branch predictions [2] or any other problems, the CPU will either stall the pipeline a few cycles or just flush the whole thing, so it's not like it's a magic solution just waiting to be adapted.

[1] http://en.wikipedia.org/wiki/Pipeline_hazard [2] http://en.wikipedia.org/wiki/Branch_predictor


Yes, but being inside the CPU means that it can be much finer grained which makes the problem much easier.




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

Search: