Considering the differences between 12700K and 13/14600K for a new build.
Considering the differences between 12700K and 13/14600K for a new build.
They perform almost identically, yet in most tests the 8 P-cores underperform compared to 6 P-cores with more E-cores. For regular tasks—multiple browser tabs, virtual machines, or background CAD software—the results are similar. The price difference between the 3 CPUs is also comparable to the 8 vs 6 P-core setup. It’s hard to decide whether to go with the E-core version.
The issue is unclear to me. It seems VMs don't necessarily use physical cores directly, and the allocation often reflects thread counts instead. With three processors each having 20 threads, it's hard to pinpoint a clear advantage for VMs in terms of processor distribution.
I've only ever used VMs for running Android games or once I set up a W7 VM for some reason I can't remember now.
Not using VMWare Workstation either. If you leave the vmx enabled on them, your virtual machines will be extremely slow. For example, a Windows XP VM runs much slower than on a Pentium II 233. I tend to be quite frustrated with my e-cores, which makes me really want a 7800X3D. Running VMs can be a real hassle, Steam downloads are weird, but the rest works fine.
I checked the discussions and it seems many find E cores challenging for virtual machines. It looks like 12700K might be a better choice.
It's probably the route I would go. Hell, if I had a 12700K instead of my 12600KF I'd probably just disable the e-cores altogether since I'd have an extra two big boy cores/threads which would help "make up the difference" so to speak. To illustrate my point of how ass e-cores are for VMs, here's a freshly installed Xubuntu 21.10 VM booting from my MP510, with 6GB of RAM and 4 cores assigned. First part is with it using the e-cores, second part is me changing the affinity to only p-cores. And believe me, the difference with Windows guests is FAR greater. 2024-01-12 12-32-43.mp4
It came to my attention. I’m curious if this could free up more TDP for pushing the 8 P-cores even further. Have they tested this setup? I haven’t managed a consistent overclocked system that works every day yet. Still, experimenting with numbers sounds enjoyable for now.
You're in the same boat as me—using boost tables, voltage curves, and PL1/PL2 limits. It really works. Or it simply reduces overall power consumption, which can be useful.