HWiNFO v6.23-4080 Beta released

[Martin]

HWiNFO v6.23-4080 Beta available.

Changes:

• Improved power monitoring of EVGA iCX 20-series GPUs.
• Added temperature monitoring of some legacy NVIDIA GPUs.
• Enhanced sensor monitoring on EVGA SR-3 DARK.
• Improved support of XGI Volari and SiS315/330 GPUs.
• Added option to truncate values in tray icons.
• Improved support of legacy systems with EISA bus.

 

[VictorVG]

Driver installation have small problem - then new driver have name HWiNFO150:

Old:

HWiNFO, HWiNFO Kernel Driver, Driver, Running, System start, \??\C:\Windows\system32\drivers\HWiNFO64A.SYS, Not trusted

New:

HWiNFO150, HWiNFO Kernel Driver, Driver, Running, System start, \??\C:\Windows\system32\drivers\HWiNFO64A_150.SYS, Not trusted

for old driver needed send SCM command:

SERVICES_STOP
SERVICES_DELETE

before install new driver or user see error messages whets OS cann't install new driver check AntiVirus software. I just use use Process Hacker for stop/delete old driver, but not all users know it.

Best regards, Victor.

 

[Martin]

This change in driver was intentional, but the error message during first run not
Let me check that now...

 

[VictorVG]

I met this phenomenon on two machines, though I myself ruined the screenshot of the window indicating the antivirus - the screen magnifier was accidentally turned on and as a result the screenshot came out "unreadable".

 

[Martin]

If you experience this again it would be great to capture that error in the Debug File.
I tried to reproduce such error, yet wasn't able to trigger it..

 

[VictorVG]

I will do it - there is one more machine under Win7 there that I will do. True, a dump can come out voluminous - through SeDebugPrivileges I will give a command to create a dump of the process memory.

 

[Zach]

Just a question for the new beta version, outside your discussion here

What is "Package C6 Residency". I Know what individual core C0/1/6 residency is, but the Package C6? Its appears always 0% for all cur/min/max/avg even though some cores are entering "Core C6 Residency" quiet often.

 

[Martin]

Just a question for the new beta version, outside your discussion here

What is "Package C6 Residency". I Know what individual core C0/1/6 residency is, but the Package C6? Its appears always 0% for all cur/min/max/avg even though some cores are entering "Core C6 Residency" quiet often.

This is the whole-CPU (entire CPU package) C6 (sleeping) state residency.

 

[VictorVG]

Yes, it was reproduced, the dump (MSVC format - make use dbcore.dll and Process Hacker v3.0.7436.2909 Git-6b1939f9 x64) is voluminous - 110 Mb. Put on the exchanger: https://www.upload.ee/files/11125006/dmp.7z.html

messages screenshot:

 

[VictorVG]

Zach

If simple, then state C6 is the state when the CPU is off, but it is ready to start by an external start signal. The CPU exit from this state is not much different from the initial startup state, except that the control logic skips part of the start-up checks.

Also, please, read Intel document POWER MANAGEMENT STATES: P-STATES, C-STATES, AND PACKAGE C-STATES for more details.

 

[Zach]

This is the whole-CPU (entire CPU package) C6 (sleeping) state residency.

Zach

If simple, then state C6 is the state when the CPU is off, but it is ready to start by an external start signal. The CPU exit from this state is not much different from the initial startup state, except that the control logic skips part of the start-up checks.

Also, please, read Intel document POWER MANAGEMENT STATES: P-STATES, C-STATES, AND PACKAGE C-STATES for more details.

I now more or less this...
But it should report something as long as individual cores are entering PowerDown C6 states right?

Take a look...
If I understood it correctly the "sensor" is a calculator and it should reported
Package C6 Residency: cur_7.7%, min_0.0%, max_52.02%, avg_19.12%
but its all 0%

 

[Zach]

I'm trying to upload the DBG file but the platform wont let me.
Says "The uploaded file is too large for the server to process"
Its 2523KB

Or it is not allowed in this section?

 

[Martin]

Yes, it was reproduced, the dump (MSVC format - make use dbcore.dll and Process Hacker v3.0.7436.2909 Git-6b1939f9 x64) is voluminous - 110 Mb. Put on the exchanger: https://www.upload.ee/files/11125006/dmp.7z.html

messages screenshot:

View attachment 4135

Thanks, but could you also create the HWiNFO Debug File (DBG) ?

 

[Martin]

I now more or less this...
But it should report something as long as individual cores are entering PowerDown C6 states right?

Take a look...
If I understood it correctly the "sensor" is a calculator and it should reported
Package C6 Residency: cur_7.7%, min_0.0%, max_52.02%, avg_19.12%
but its all 0%

Spoiler

View attachment 4136

I'm not quite sure how this works in AMD, but I assume all cores have to be in a lower power state at the same time for the entire package to reach C6.
Most likely the Windows Power Plan does affect this too.
Try to put the machine into sleep mode.

 

[Zach]

If thats the case, then ok! I thought it worked differently.
Thanks!

What about the upload, and that I couldnt do it?

 

[VictorVG]

So everything is simple - each CPU core can be turned off simply - just put a logic element (LE) 2I at the synchronization input and send a ban signal to one of its inputs, for LE 2I it is a logical 0 and the circuit will stop in the state in which the "ban" . And since the transistors in it operate in the key mode (class D), the transistor consumes almost all the energy in the process of switching between the states “open” (saturated and conducts current) and “closed” (in the cutoff and does not conduct current), but in states “open” / “closed” its consumption is ten times lower than during the switching process, which provides energy savings. Reducing the supply voltage to the minimum gives additional savings, but it will affect the heating of the power supply more so it uses the change in the duration of the open state of the power switch with a constant voltage at its input, and this switch in turn charges the capacitors supplying our circuit (PWM voltage stabilization circuit ) And the shorter the time the key is opened, the less voltage will be on the capacitor, and since the current through the key is more or less unchanged, then its heating will decrease.

 

[Martin]

If thats the case, then ok! I thought it worked differently.
Thanks!

What about the upload, and that I couldnt do it?

There was a server migration recently, probably some of the settings didn't update.. Will check that, please have patience...

 

[VictorVG]

Martin

There he forgot to enable logging in HWiNFO , but made a memory dump at the moment when it displayed a message. According to PH, the size of the occupied HWiNFO memory at that moment was approximately 105 MB, in usual conditions I did not see more than 39 - 40 MB.

.dbg can be removed, but in the afternoon, done?

 

[Martin]

Martin

There he forgot to enable logging in HWiNFO , but made a memory dump at the moment when it displayed a message. According to PH, the size of the occupied HWiNFO memory at that moment was approximately 105 MB, in usual conditions I did not see more than 39 - 40 MB.

.dbg can be removed, but in the afternoon, done?

Yes, you can remove that file. If it happens once again please make sure to create the HWiNFO Debug File.

 

[VictorVG]

OK! Debug log + stack + error window info and dump - https://www.upload.ee/files/11128248/HWiNFO64.exe.dmp.7z.html (WinDbg format)