On the crosshair you can set custom curves, got mine quieter at full load than my FX was at iddle with clc. (dual loop on ryzen)
yeh i got a custom curve but asoon as it spikes it ramps it up is their a way to make it were it waits 1-2 seconds before actually ramping up insted of instantly
i found a fan delay i set it to 2.28 seconds and the fans arnt ramping up on the little spikes so
1 Like
The fan delay is nice. I just wish I could control CPU_FAN and CPU_OPT with the fluid temp like the other two fans in my system.
it sucks i can inky go to 75 the it auto defaults the fans to 100% not like it gets that hot but still
Which motherboard are you using?
And what particular voltage did you set and eventually which LLC setting?
If you use LLC then you need to understand how your particular motherboard,
handles those said levels.
ASUS ROG CROSSHAIR VI EXTREME is the board
thank u ive been busy tonight playing ESO lol
If you run Prime95 what does the voltage drop to?
If it stays at 1.41/1.42 you have a real problem.
i dont run prime 95 at all as its proven to be a trash program that casues more good then bad. the cpu voltage never goes past 1.42v as thats the max i have it set at for 100% load thats less then a .15 oc on the voltage side as my chip came to my door set at 1.412
it idles around 1.38 when doing nothing in game correction. if u looked at the photos of hwmonitor it shows u the min volages
That’s pretty opinionated for someone asking for help. Don’t bite the hand.
Be aware that Ryzen does a whole lot of internally segmented voltage regulation and power distribution and the sensors readout we get via PECI seen in HWmon/HWInfo are an average over short time periods, but not all the sensors output updates at the same rate. And the programs you use don’t actually report data as accurately as everyone perceives they do. So the data you see at any one time on screen is not ‘in sync’ at the same time with everything else and some data is calculated from other values which the program is reading wrongly. An example is the extremely low voltages of 0.033V reported on VIN5 with HWMonitor.
Now the wattage of the CPU power draw is calculated based on V(Voltage) * C(Amperes) = Watt.
Remember: The Voltage you set is driven by the Board VRM, but inside the CPU where it’s measured Voltage is regulated and managed differently in microsecond intervals.
If the internal voltage spiked, leading to increased current for a short time then you will get a higher wattage displayed even if the voltage to you is still displayed the same.
Same goes if the program gathered garbage numbers like 0.033V, suddenly your power draw will appear very low.
As for my 1700X at stock the voltage reads 1.18V under full load. Overlocked to 4Ghz it would be about 1.3V at least and still be stable.
This is normal due to how the internal sensor network reports voltage under heavy load.
Now as an edge case: If however the VRM keeps increasing Voltage to remain at an external 1.42V under full load, your actual internal voltage may be considerably higher.
That said basically everything is normal on your system. But you should really stability test that 4Ghz overclock. At least 80% of Ryzen chips are not RAM & Cache stable like that even though people will run windows and games with them without obvious issues only to then notice x265 encodes crash.
I know this because I have been testing and mucking about with several Ryzen desktops & Epyc servers since March.
3 Likes
i have an hour of aida 64 and from waht ive read 1 hour on aida 64 is 100% stable.
so ive been playing around in the bios for a bit and managed to get it to 4ghz off stock voltage and slowly went from LLC 1 to LLC4 seeing as 4 was the only stable at 4ghz. now indidnt adjust anything but LLC and Power phase to balance. is their anything else i can do to change? it seems pretty stable. but im seeing spikes while under full load on aida 64 to 77 degrees c and a base of 69 degrees c with my fan curve… note its not an agressive fan curve.
so yah stable 1.4125v ( my default voltage) at 4.0 ghz max under 100% load 1 hour aida 64 is 69degrees with 1 spike every 12-15 min
I realize 1.4125v may have been what your board was doing from first boot, but this IS an OC voltage, not a Ryzen stock, the stock voltatage range is 1.2v to 1.3625V depending on workload (although it often looks like 1.375 in many programs,as they round to the nearest 0.025v), and if you turn the multiplier over 34x on the 1700x (and it DOES vary based on which chip, the 1700 is lower, the 1800x higher in this regard) it will lock to the latter unless overridden and disable PB/XFR. I recommend you check for a bios update, but I think your ~1.4v is likely the max that your motherboard had for it auto-OC , and assuming your temps are not getting high, it may use it.
its a full custom loop it came out at a auto 1.39-1.41v out of the box i hard reset it today… and tinkered with a profile…
Nothing looks abnormal to me. Even the max temps seemed ok if the system was loaded. I run a stock cooler on a 1700 and can run 4.0Ghz @ 1.4V with similar number but not fully loaded because to cooler can’t keep up and mostly loaded like handbrake it’s fine but hotter than yours.
I dialed mined back to 3.75Ghz because my RX480 is the bottleneck not my CPU.
yeh im getting rid of my 2 fury xs for a vega 64 and a vertical gpu bracket… so ill have alot more air flow in my case so im planning to see positive changes and have better temps