INTERACT FORUM
Devices => PC's and Other Hardware => Topic started by: Manfred on October 20, 2022, 03:10:39 am
-
The 12th gen intel nuc generation has now Performance Cores, Efficient Cores:
https://www.quietpc.com/sys-ultranuc-pro-12 (https://www.quietpc.com/sys-ultranuc-pro-12)
Does that impact MC's behaviour and does MC work with this new intel architecture?
-
I would be highly surprised if any of that affected MC in any meaningful way. And considering that Windows 10 and Windows 11 has full support for Intel's 12th gen, it shouldn't be any issue. Though for 12th gen and newer it might be a good idea to run Windows 11 over Windows 10 (as knowing Microsoft, they probably add better support in Windows 11 to force users to upgrade).
However... Intel's 13th gen is coming out very soon (the benchmark embargo was lifted today and Intel's 13th gen beats AMD's Ryzen 7000 series!) with big gains and cheaper prices.
-
The 12th gen intel nuc generation has now Performance Cores, Efficient Cores:
https://www.quietpc.com/sys-ultranuc-pro-12 (https://www.quietpc.com/sys-ultranuc-pro-12)
Does that impact MC's behaviour and does MC work with this new intel architecture?
Short answer no. There was a bunch of worry about how this would impact performance. Yes, Windows 11 has better support for scheduling P / E core architecture (and probably generally better handling of Big / Little architectures). However, in most things, there hasn't been serious side effects in Windows 10.
I too was a little skeptical of the P/E core architecture. I had it explained to me quite recently and it actually makes a bunch of sense for a wide range of applications / situations.
P/E cores aren't *just* a speed difference, and indeed the speed difference might be much less than the numbers suggest. What's really differentiating their strengths is latency apparently. So your P cores aren't just 'fast', they're snappy. They 'know' or rather are assigned, time sensitive tasks. E cores? They're not as slow as we think, and they're being assigned things that don't need to be done RIGHT NOW. It's intelligent thread scheduling pretty much. How much of that is on chip and how much of that is OS, I'm not sure. That's out of my depth. We already do thread scheduling obviously, but now we're 'picking' threads and specific 'tasks' to go on a core with characteristics that suit the job at hand.