Potential Bug in v5.03-2595 Beta

parsec

Well-Known Member
Hi Martin,

I tried the new Beta version of HWiNFO64, v5.03-2595, which includes Enhanced sensor monitoring on ASRock X99 and Z170 series.

I ran it with my ASRock X99 Extreme6/3.1 board, and I noticed the VCore reading that was fine on the previous Beta version was now an impossible value, over 2V at times.

I found that the sensor item that was VCore in the previous Beta (VIN14), now has an Original name of 3VSB. The new Beta now has two 3VSB items/readings with this ASRock X99 board.

I'm attaching a debug file and Report file. I have some sensor reading edited out of the display, so let me know if you need a debug file with all the readings in it.

Thanks for your attention, I depend on the HWiNFO64 VCore reading, since everything else displays the VIDs. No pressure Martin... ;)
 

Attachments

  • DESKTOP-M6EADEK.HTM
    286.1 KB · Views: 1
  • HWiNFO64.DBG
    972 KB · Views: 1
Are you sure that VIN14 was the correct Vcore there? I don't see this being used for your board by ASRock config files.
 
Martin said:

Yes it is Martin, thank you for that! My precious VCore reading is back!

I treasure your VCore reading because nothing else provides it, all I get is the maximum VID.

When CPU stress testing using Haswell Adaptive Voltage, the actual VCore will be above the maximum VID. That is known to happen when using Adaptive voltage mode.

The VID will also increase beyond its usual limit, but the true VCore is even greater. That is important to know when dialing in an OC on a Haswell processor.


About my VIN14 comment, that 2596 version labels what is VCore, as... Vcore. I don't see a VIN14 reading among the sensor readings in 2596.

I think you know now that the 2595 version does not have a VIN14 reading, when used with my X99 board. But since I said I edited out some readings, there was no way for you to know if VIN14 was missing or edited out by me, in the debug file data. Sorry about that! Do you agree or am I missing something?

Some more (positive) feedback, I see three separate VR data displays in HWiNFO when used with my X99 board. One of course is the main CPU VR data. I also see two additional VR data sets, which are for the DRAM memory VRs. With four channels of memory on X99 boards, there is one memory VR for two channels, so two total. The memory VR data displays have been shown on earlier versions of HWiNFO, so not a recent addition.

I am curious what enhancements to ASRock X99 sensor monitoring you added in Beta 2595. Frankly, I did not notice anything new.

Thanks for the super quick fix Martin, I truly appreciate it! :)  As always, HWiNFO is the best PC monitoring program in the world! :cool:
 
What I did in Beta 2595 was that I changed some sensor names/values (mostly voltages) for ASRock X99 series. This was based on new ASRock configuration files used for their tools.
But as you noticed not all of them were correct. This is because I assumed that all their X99 boards used the same assignments, some were just hidden. But that is not true - for example the X99 Professional has Vcore at VIN12 and 3VSB at VIN14, while your board doesn't have Vcore assigned in the config files (or more precisely it seems to report VID instead of VCore).. But I know from the past, that ASRock has been monitoring more values (like VCore), which was not (properly) utilized via their tools.
So in 2596 I renamed VIN14 on your board to VCore based on your assumption, hoping that it's correct ;)
 
Martin said:
What I did in Beta 2595 was that I changed some sensor names/values (mostly voltages) for ASRock X99 series. This was based on new ASRock configuration files used for their tools.
But as you noticed not all of them were correct. This is because I assumed that all their X99 boards used the same assignments, some were just hidden. But that is not true - for example the X99 Professional has Vcore at VIN12 and 3VSB at VIN14, while your board doesn't have Vcore assigned in the config files (or more precisely it seems to report VID instead of VCore).. But I know from the past, that ASRock has been monitoring more values (like VCore), which was not (properly) utilized via their tools.
So in 2596 I renamed VIN14 on your board to VCore based on your assumption, hoping that it's correct ;)

Thanks for the explanation and the extra work for my ASRock X99 board. I really appreciate it. :)

Your VCore reading seems accurate to me, given my experience with HWinFO and my ASRock Z97 board.

On both boards with C States enabled, the HWiNFO VCore reading is not always identical to the VIDs. The VCore readings will show values well below the minimum VIDs shown in HWiNFO and in the Intel Extreme Tuning Utility, which I always use for comparison purposes. The HWiNFO VCore reading will drop to 0.00000 volts during CPU idle moments.

The program ThrottleStop 6.00 shows the CPU cores in the C6 state in the mid to high 90% range when the VCore reading in HWiNFO is zero or a fraction of a volt. I know they are not synchronized, but that's the best we can get. ASRock uses a less than ideal UEFI version on this new model board (USB 3.1 series) and C7 C State support does not exist in the UEFI yet...  :dodgy:

I'd like to think the HWiNFO VCore reading is accurate, its behavior when stress testing with adaptive voltage is also in general what it should be, greater than the VIDs. But if you think (know) it isn't, then I'd rather know the truth rather than fool myself.
 
I don't know more than you in this matter ;)
Asking ASRock which VIN is Vcore might provide the answer. I know they gave such information in the past about Vcore where certain boards were not showing them.
 
Back
Top