As far as concrete numbers go for folders, this alone is a big reason to stay away from Bulldozer. Folders will often run both the cpu and gpu at the same time at 100%. If Bulldozer came in the same as a 2600K in folding numbers, but more efficiently then i could have been happy. But I don't see that happening, even when official PPD numbers start to roll in.
The only hope I see for folders here is if the gromacs folks can optimize for the new SSE instructions. But those guys can take quite a while to get these optimizations done due to the needed reliability of submitted results. Maybe these Bulldozer chips will do well in the multi-processor boards.
Even then, with these power numbers....
And who wouldn't want better performance in 198% of tasks? :bleh:
Seriously, though - Bulldozer's failure to keep up in single- or lightly-threaded applications really oughtn't come as much of a surprise to anybody. It's failure to keep up in multi-threaded applications, though, is an absolutely shocking finding that undercuts the whole raison d'être of Bulldozer. here are my takeaways:
It's a poor chip for ordinary users; It's a poor chip for gamers; It's a poor chip for crunchers; and, from the indicators that hint at SMP performance (wPrime and Cinebench), It's a poor chip for servers. In sum, Bulldozer has failed to appeal to ANY market at all, not even AMD's traditional stronghold in the budget category. AMD has pulled some fail over the years, but this has got to be the most epic of them all! :sad:
Guess I was waiting for nothing. I shudder at the thought of ever folding with this. Hell, they should name it PussyCat. Expected a roar, got a tiny meow instead. What an utter travesty. :doh:
Probably not that much to be honest.
The cooler a processor is the less power it consumes, so supposedly super-cooled processors are extremely energy efficient.
Someone else can surely elaborate on the physics behind it.