Just curious as to performance for other users out there. I recently shut-off these two options for other reasons, but after playing a bit of VR after work my VR experience compared to when the settings were on seemed a lot smoother overall. Less stutter, less HMD jumps, no more fixed-screen freeze while loading (when your HMD view freezes on a frame and the screen doesn't change regardless of HMD position).
Also to note, I only have C0/C1 states enabled for the CPU, even before switching off Speedstep/EIST.
ps. For those replying, please include you experience with the settings, especially if you have tested their effects on VR performance.
pss. I am only familiar with the Intel versions of this power-saving tech, but the question still applies to AMD's chips (I would assume they have a similar type of power-saving tech).
psss. I am semi-new to the more technical elements of computing, and so if anyone has insight as to how these settings have an effect on VR performance (or lackthereof), I am all ears. Afaik, the main difference is that with the settings on there is some lag between when a CPU core is requiring a higher voltage for processing and when it actually receives the necessary voltage to carry it out (assuming the core was in a low-power state before the instruction).
I'd suggest you look at the thermals on your overclock before you start messing around with performance boosts from those CPU settings, make sure those hitches and pauses you're seeing are not from heat saturation after gaming sessions push your cpu or gpu to the thermal limits. Is the 1080 overclocked?
@HeadRusch1 The 1080 is overclocked, but I've had that overclock for over a year now and temps have been fine. The CPU and GPU temps are normal. CPU rarely hits above 70C and GPU hovers around 70C under max load. Even with the new constant voltage, the CPU still stays in the normal range (though idles 3-5C higher).
I just run a locked 4.7 GHz on my i7-8700k normally. Sometimes I'll turn on power saver mode in windows, which will down clock it when idle but it doesn't seem to have any impact on gaming.
I've not noticed any anomalies in performance just letting the CPU / motherboard change clock speeds to between 3.7 / 4.3 GHz and the 1080 managing itself.
CPU goes to about 60c max during gaming, GPU I'd have to check but never noticed anything amiss. The 1080 runs a small factory overclock but that is all.
I too have noticed that any cpu throttling seems to interfere with VR. I've turned them all off in BIOS and a mild overclock seems to make things much more stable. It really seemed to help my experience. Also, gpu temps and cpu temps are a totally different thing. gpus are designed to run at much higher temps, in the 80 to 90 deg C (high 90's or 100 is too high) but cpus can't handle those temps, they should generally run much lower, 50-60 deg C is pretty high. AMD chips run hotter, but just note that cpu's can't handle high temps without crashing.
@xor1337 - things with the CPU are not as they seem - as I don't monitor Tdie but CPU-Z's tCTRL feed there is a 10c temperature offset to be taken into consideration (so 60c reported is Tdie of 50c) so there is still 45c headroom (tMAXX being 95c Tdie) so those temperatures being reported are not too much of a concern. The PC is on a temporary water loop with only a 120mm radiator while I wait for a handful of fittings to arrive. Even with this setup it is still running 15c or so lower than the stock Prism air cooler.
That said, from past experience it is quite surprising just how long a processor running constantly high temperatures can last - thinking of a server in a tiny cupboard that ran for several years averaging 75c Tdie...
@xor1337 50-60C is high for CPU temps? I've always heard that you can run your CPU within voltage specification limits as long as you stay below 80C (or the temp in which throttling starts occurring, but even then throttling is only in anticipation of higher temps that can actually cause damage), and that CPU's were designed to handle the heat in the full range of temps, up to the throttle point.
Then again, I'm not a electrical engineer, and I've only read such things on overclocking websites/forums.
I do understand though the higher the average temps you run on your CPU theoretically lowers its lifespan, in which the idea of a 50-60C highpoint would be a good rule-of-thumb for persons needing the computer to never have issues due to heat.
Intel's i7's will run at 100C before they throttle when at stock settings. My laptop with an i7-2670QM idles at about 60C and runs 80-90c most of the time when actually in use. Given it's about 7 years old and still works just fine, the heat clearly isn't an issue other than making the fan sound like a hairdryer. These are core temps as reported by HWMonitor. CPU-Z reports motherboard temps for the CPU socket which are cooler. My desktop's i7-8700k seems to have similar limits, though I target 70-80C max temps.
AMD's Ryzen lineup is rated for 84C before throttling if memory serves. They are too new to say how such temperatures affect longevity but one would assume AMD fully tested them at those temps and set it to a long term safe limit. Not that Ryzen will ever hit those temps with even a basic air cooler. When I used a Ryzen 5 1600X for mining Monero, it hung out at around 50-60C with a cheap air cooler despite having all 6 cores loaded up 24/7.
Sounds like you have your temps under control, and others seem to notice those settings affecting VR. Good luck with your continued tweaking, you're probably on to something here.
