Anomalous sensor readings in HWiNFO64

whatever

New Member
As per title, I'm getting obviously wrong readings on the following sensors: BUS Clock/PCIe Clock and Memory Clock. These values fluctuate rather wildly from their normal, constant values of 100mhz and 800mhz respectively, which in the part of the base clock value causes core clock readings to be erroneous as well. Appart from common sense, I've verified with two other monitoring programs that HWiNFO readings are not correct on these items and that it's HWiNFO that's misreading the sensors or is somehow otherwise bugging out. The errors manifest themselves when taxing the system, for example playing a game or running a benchmark, in light desktop use such as web browsing these values appear more correct. I pretty much have to treat any readings with the program now as suspect since it can't read these correct.

I've attached two pictures showing the wrong readings.

My system specs are the following:

Intel i5 4690k stock (3.9ghz turbo)
Asus Z87C motherboard, almost all BIOS settings at their default values
Kingston HyperX Fury 1600mhz 16gb RAM, timings overcloked to 9-9-9-24
Windows 7 Pro SP1
HWiNFO64 version 6.10-3880

Regarding HWiNFO settings, they're pretty much at default. I tried disabling some and all of the 'CPU Clock Measurement' settings, but they just made the problem worse. I'd provide a debug log, but I've no idea how to take one with this program, I see the 'Debug mode' setting, but no idea how to start taking the log which is apparently supposed to appear in the program's directory. The sensors page 'Logging start' button produces a .csv file that's pretty much unintelligible when viewing in OpenOffice and doesn't include anything that looks like debug info.

I'll provide whatever else is needed to solve this problem if required.

Thanks.
 

Attachments

  • hwinfo1.png
    hwinfo1.png
    85.4 KB · Views: 32
  • hwinfo2.png
    hwinfo2.png
    79.6 KB · Views: 31
BCLK fluctuations especially under high load are known on some CPU families before Skylake, but not so extreme as you have described.
As a workaround I suggest to enable the "Bus Clock-based" option and disable "Periodic polling".
 
BCLK fluctuations especially under high load are known on some CPU families before Skylake, but not so extreme as you have described.
As a workaround I suggest to enable the "Bus Clock-based" option and disable "Periodic polling".

Thanks for the reply. However like I stated, I already tried toggling those options and they don't make any difference, in fact having none of them on makes it worse. However I've since remembered I have HPET disabled in Windows device manager, could this make a difference? Naturally I assume the program uses the motherboard's timer instead of Windows'.

Is there some information I could provide to help fix the issue with pre-Skylake processors, like the mentioned debug log? Or if there is no priority due to such old processors, could you recommend me an old version of the program that would function properly?
 
Yes, HPET can make a significant difference as it's used for more precise measurements.
Best option would be to enable HPET and use the settings I posted above.
 
Turning HPET on didn't fix this issue, the readings are still erroneously fluctuating like in the screenshots.

I ask again, is there some previous version of HWiNFO64 that I could use that has better support for Devil's Canyon/Haswell chipsets?
 
Ah, wait ! I didn't notice the version you're currently using. There is a specific bug in v6.10 that occurs on some systems with Windows 7.
Upgrade to the latest Beta version (6.11), that should work much better.
 
Screenshot (520).png
Getting weird readings too on my ASUS Prime x470-Z Pro. If my MOBO was really that hot then I would be in trouble lol and I have nothing plugged into my CPU OPT header
 
Nevermind I remade the debug file and it was smaller this time ;) BTW this is the beta version too.
 

Attachments

  • HWiNFO64.DBG
    1.8 MB · Views: 5
This is a BIOS bug. Did those erratic values appear after a recent BIOS update? I see you're using 5216 from 2-Sep.
Yeah. The odd thing is I haven't had any readings like that from MSI Afterburner or any other software I have used since
 
Back
Top