The real question, then, is whether this actually results in increased parallelism. Threads are a good way to isolate program tasks (and therefore simplify the programming and reduce bugs) - but if the application-level task just switches back and forth between the emulated thread and the Rosetta thread, there is no extra parallelism and a Duo and a Solo would perform at the same speed.mrgreen4242 said:I swear I read that Rosetta loads WITH each thread that normally, so it'd be closest to #3...
The real clincher would be to look at application benchmarks - if an app that never uses more than 100% of one CPU on a dual G4, and the same app under Rosetta on a Duo does use more than 100% of a CPU - then one should question the performance of Rosetta on a Solo.