Shouldn't a "twice the speed/half the cores" cpu (like the i5 3470, no turbo) have the same performance? let's say they have the same IPC, cache size, level and latencies. I think they switch threads fast enough.
Let me rephrase, in which way a "twice the speed/half the cores" cpu will be hindered while trying to do the same job as the "slower/more cores" cpu? Maybe by "competing for resources" inside a core, but you can assign the "competing" threads to lower loaded cores.
Slower cores my be hindered by having to synchronize cpu intensive jobs that have to be divided because 1 core isn't enough. That would be an advantage for the "twice the speed" cpu. IA and physics could be an example of this? Doing the compute of those tasks in the cloud could be a solution but offline play would suffer.
Let me rephrase, in which way a "twice the speed/half the cores" cpu will be hindered while trying to do the same job as the "slower/more cores" cpu? Maybe by "competing for resources" inside a core, but you can assign the "competing" threads to lower loaded cores.
Slower cores my be hindered by having to synchronize cpu intensive jobs that have to be divided because 1 core isn't enough. That would be an advantage for the "twice the speed" cpu. IA and physics could be an example of this? Doing the compute of those tasks in the cloud could be a solution but offline play would suffer.