High Idle Temp. ASUS Tuning Tool too aggressive? (i7-4790K + Z97i-plus + EVGA ACX)

DRGustav

Reputable
Dec 12, 2015
3
0
4,510
Hello y'all. I'm seeking a little help please.

I've a i7-4790K with Asus Z97i-plus on a Mini ITX Hadron Air, EVGA ACX Tower cooler. Went into BIOS EZ Tuning tool and it reported a 10% OC was possible, but since I've noticed a 45C @ idle and from what I take a 1.3 VCORE. But reading around that doesn't sound right. Here's the HWMonitor view of sensors. Is this tool known to be too aggressive with VCORE reason why I get that high temp? Any thoughts? Tks!
35U3qFS.jpg
 
I'm not as familiar with hwmonitor however looking at the info it appears your current oc is closer to 4.5ghz. See the max core clocks? Current says 4487mhz, min says 4380, max says 4586mhz which is almost 4.6ghz. Vid and ia aren't core voltage. If that vcore above is correct it says 1.68v which is WAY too high. Disable your auto overclock would by my suggestion, it's not doing you any favors and is way too aggressive.

I don't know if that vcore is correct or not, different monitoring software can display varying readings depending how they interpret the sensors. 1.68v will damage your cpu. 1.4v is considered high and generally not recommended for everyday use. 1.3v or lower is preferable. Look up on that screenshot under 'voltages' where you have nothing circled. 5v, 3.3v, 12v - at the bottom of that subset list is 'vcore' which is your core voltage.

It might be worth trying hwinfo64, I've had pretty good luck with it. It lists data in a similar way. Here's an example of hwinfo64 with my system and where it lists the cpu vcore status. (underlined in red)
http://prntscr.com/9df7gg
 

DRGustav

Reputable
Dec 12, 2015
3
0
4,510
Thanks for the input.

I did load the default settings in BIOS and left the EZ Tune on "normal" profile. It reads 1.099V and CPU Temp dropped ~10C

What reads the VCORE correctly then? Seems the VID from HWMonitor matches the Core Voltage from CPU-Z but I'm confused as it doesn't match BIOS reading.

Here are the readings I got with the "normal" profile from different monitor software:

EqC9S8N.jpg


Ca1byid.jpg


ILmIXv3.jpg


uBtXOXt.jpg


t2qJ7jP.jpg

B3hpzb2.jpg
 
Best I can figure so far is speccy is reading a different sensor/value. Hwmonitor seems to be reporting vid as vcore and down below under your cpu name it's saying 'vid' when it means 'vcore' from the looks of those values.

Hwinfo64 doesn't seem to be helping much as it's not listing vcore but instead vcorerefin which is some other sensor/reading and seems to be an asus motherboard oddity. Other people have reported the same odd vcorerefin value on their asus boards as well.

Cpu-z seems to be correct though I don't think cpu-z shows actual current vcore but instead the max or what it's set to in the bios. If you're using eist (speedstep) to allow the cpu to idle down under low load conditions the actual vcore should drop as well. When my cpu isn't under much load it drops to around 0.800v though I do have it manually set in bios to 1.28v for my overclock. When the cpu is under full load actual vcore will increase to 1.28v. Cpu-z is showing my vcore as a steady 1.28v though which tells me it doesn't adapt in real time.

I've shared this thread with some other folks so hopefully they can also weigh in and take a look at the monitoring software discrepancies. I'd like to help you get some monitoring software that reliable as a starting point so you can at least keep an eye on things. Then you'll have a solid starting point for overclocking.

Manual overclocking is preferred to auto overclocking. I know it's easy enough to click 'turbo' or something in the bios but they can be too aggressive sometimes. Each cpu reacts a bit different to voltages and one chip may need more or less voltage than another. In an attempt to get a successful overclock a lot of auto tune utilities will err on the side of applying more voltage than necessary for stability reasons. Additional voltage only adds more heat than you need and if the core voltage is pushed or allowed to auto-raise itself too high can be bad for the cpu.
 
Yeah, the first thing I'd do is go into the bios and see if what's reported there is similar to what HWinfo reports. I'd also highly recommend downloading Core Temp and look at the VID, Core frequency and CPU temp values there, as they tend to be pretty accurate using that utility, or at least as accurate as any other.


http://www.alcpu.com/CoreTemp/
 
VID and Vcore should probably be different nomenclature for the same value, except that in HWmonitor, which sucks, the Vcore value may simply be a different value since it's listed under the PSU group and not the CPU. Your CPU sensor value is showing 1.308, which is what I'd check against the bios. Keep in mind, this value is going to change/fluctuate so long as your Intel Speedstep is enabled and your min/max processor settings in the advanced power options section of power in control panel are set to anything other than 100% for both.

That's normal though, for a stock configuration, but you can lock it in place if this is throwing you off, then change it back later or just leave it. Setting the power profile to performance should automatically change them both to 100%, but might not on some systems so it's best to double check that manually by clicking "change profile plan settings" and then "advanced settings" and look under the processor settings.
 

DRGustav

Reputable
Dec 12, 2015
3
0
4,510
Thanks for the input Darkbreeze.

Interestingly after everything was back to default I noticed that enabling the XMP profile the CPU temperature goes from 35C to 45C on idle with or without running the EZ tuning tool.
In conclusion, without tinkering and dialing all the number manually there'll be no XMP neither OC for me.
 
All overclocking should be done manually anyhow. There are good reasons for this, not the least of which is the fact that the software developers for these manufacturers tend to err to the side of stabilty/caution with their overclocking settings or utilities. This means they generally increase the voltage significantly when it may not be necessary or necessary to the degree which they employ. As can be expected, increases beyond what's necessary/recommended when manually overclocking, in those kinds of automatic or one click utilities, increases voltage to the point where thermals are often out of control or just far too elevated to achieve anything realistic without having severe thermal issues, especially without an extreme cooling solution and phenomenal case cooling airflow.

I highly recommend doing any overclocking manually. I'm unsure as to why you would have an immediate temperature increase using the XMP profile unless you have very little airflow or something is wrong with your cooling solution. I'd make absolutely certain your heatsink or waterblock are mounted with the correct pressure, and that there is not too much or too little thermal paste. Double check operation of all fans and also orientation to ensure that front, bottom and side fans are configured as intake and that top or rear fans are configured as exhaust. Use of liquid cooling may change that methodology somewhat, depending on if you have a top mounted radiator and how it's configured.

You could also simply have a bad board. I'd double check the individual DRAM settings to make sure the XMP profile you've selected isn't using a voltage other than what it's supposed to and I'd probably manually set it to what the specs say the voltage should be for the speed you have configured.