Asrock z370 Extreme4 Stutter, Bad frametime spikes - i7 8700k, GTX 1070

[rsky]

Recently I bought new parts for my build, together with the 8700k I got:

EVGA GeForce GTX1070 FTW Gaming 8GB
G.Skill Ripjaws 3200Mhz 2x8GB
ASRock Z370 Extreme4
Scythe Mugen5 PCGH Edition
Corsair CS650M PSU

Only XMP setting and standard turbo OC (max 4,6 Ghz) in BIOS

Temps stay pretty low, but when I try to run programs like Chrome or a map from a game then the temps jump +10 - +20 in HWmonitor and Afterburner. Temps stay there for 1 second and then drop down to normal again.

Example between idle and starting chrome:

[img=628,906]https://i.imgur.com/nJ3WM0d.jpg[/img]

Ingame example (Quake Champions: Maxed) and launching chrome in the background. Loading a new map within the game gives the same results:

[img=302,318]https://i.imgur.com/eE810c1.jpg[/img]

Afaik temps shouldnt rise 10 - 20 degrees for 1 second for launching common applications, so im wondering if there something set wrong.

Voltage seems to go fairly high on some clocks (1.350) aswell.

Thanks in advance.

 

[CompuTronix]

rsky,

On behalf of Tom's Moderator Team, welcome aboard!

I just finished writing this in two other Threads when I saw yours; it's from the following Thread: Possible issue with Core i7 8700K Clock speed fluctuation on Idle - http://www.tomshardware.com/answers/id-3722730/issue-core-8700k-clock-speed-fluctuation-idle.html

Intel’s specification for Digital Thermal Sensor (Core temperature) response time is 256 milliseconds, or about 1/4th of a second, so Core temperatures increase and decrease instantly with changes in load.

As you're no doubt aware, since Windows has dozens of Processes and Services running in the background, it’s normal to see rapid and random Core temperature “spikes” or fluctuations, especially during the first few minutes after startup, which should eventually settle.

Any software activity will show some percentage of CPU Utilization in Task Manager, where unnecessary Tray items, Startups, Processes and Services that contribute to excessive or continued spiking can be disabled. However, as the workload increases to 100%, the spike levels are overtaken and become obscured by the load. Nonetheless, here's what's different about your 8th Generation processor:

Beginning with 6th Generation processors, Intel introduced "Speed Shift" technology in Windows 10, which responds much faster to changes in workload than "SpeedStep" due to having many more Core speed and Core voltage transition levels.

Since 7th and 8th Generation Speed Shift is twice as fast as 6th Generation, some users complain of Core temperature spikes, which in worst case scenarios, can also cause fluctuations in fan RPM at idle. Motherboard manufacturers are developing BIOS fixes that include separate SpeedStep and Speed Shift settings with more flexible fan curves and time delay options.

There's nothing wrong with your 8700K; it's behaving quite normally as per design.

CT

... the cpu is now reporting spikes to the heavens (for no good reason ?).
Example, just moving the mouse around ( not even clicking, just moving ) locks the CPU to 4.4GHz until I stop; as seen in this print - https://imgur.com/a/vnJ48ND - under Profile 1 in MSI AB.
Why ? Why do I need 4.4GHz to move the mouse (specifically on build 1803) ?
I've never seen something like this. I may accept it, but I still don't get it, makes no sense.

It may be normal, but it feels wrong, very wrong.

Most members experience the same problem with their 8700K that you've described. We see many of these threads every day. Since temperatures respond instantly to changes in load, most people express their concerns based on temperature spikes, while some, like yourself, express the problem in terms of load spikes. Nevertheless, it's the same problem, which is also shared by the 7700K and other high-end 7th and 8th generation processors.

There's a 74 page thread on Intel's Forum regarding this very issue - https://communities.intel.com/thread/110728?start=1095&tstart=0 - The title is a bit misleading as it doesn't accurately characterize the problem, but as the thread progresses, the actual problem becomes more clear.

As I attempted to explain to you in my above post, the root of the problem is Speed Shift, which is driver-integrated into Windows 10. As illustrated in the above chart, Intel implemented this technology to save power over time, by greatly reducing the time it takes to execute and complete processor tasks, then quickly return to idle. Since Intel has made impressive strides in reducing power consumption and increasing efficiency, the development of power saving features such as Speed Shift result in considerable energy savings.

However, users who observe Speed Shift's behavior graphically rather than numerically, such as you and I, whether thermal spikes or load spikes, are almost always taken aback by how rapid, pronounced and seemingly continuous the spiking actually is. All processors inherently produce load and thermal spikes when executing various workloads, especially those which are heavier with greater fluctuations. As load increases to 100% TDP, spiking from lesser background loads becomes obscured. Spiking can not be eliminated, only dampened and minimized by improving cooling, reducing core voltage and delidding.

Does this explanation make more sense?

I hope that clarifies the nature of the problem. Also, we have a Guide you should read:

Intel CPU Temperature Guide - http://www.tomshardware.com/forum/id-1800828/intel-temperature-guide.html

Once again, welcome aboard!

CT

 

[rsky]

Thanks for the post, I have changed a few things about my setup and it seems that it doesnt get as hot anymore.

However I am having an entirely different issue now. With games I get really bad frametime spikes. These happen in Windows aswell sometimes, however way less noticeable. Things I tried to avoid this:

-Reinstall Windows 10 (twice)
-Install motherboard drivers
-Flash motherboard to new version
-Update firmware on SSD
-Change SATA cables
-Disable XMP
-Remove NVIDIA drivers and reinstall latest
-Use specific settings in NVIDIA control panel
-Test CPU and RAM under stress
-Default settings motherboard

I am completely out of things to try, and this issue was not there with my older motherboard / cpu. Here is an image to show how inconsistent it is:

Is there a posibility that the PSU is defect?

Really hoping some has a solution to this, as im almost tempted to send everything back at this point. Its a bit strange that better hardware performs worse.