Comment by christkv

I wonder how they handle potential differences in memory barriers, instruction order scheduling and other stuff and do they run the core in one mode continuously or do the mix instruction streams from different instruction sets? Anybody got a link to an article?