Major BUG: GPU-Z causes Windows 10 PC to hard lock

[vegaio]

Maybe i should add, theres no stability problem at all with GPU-Z latest version running and 17.11.1 driver. But there are many crashes with 17.11.2 , but i did not test, if the crashes with 17.11.2 would persist without GPU-Z running.

 

[jaggerwild]

O

Sorry, those specs listed are very outdated, I am not using that at all. Will update in a bit, I am overloaded with work in recent weeks.

EDIT: Fixed.

Over loaded with work but enough time to come here n post that it happened once.
I think TOMSHARDWARE misses you!!! Go back towards the light!!!!

 

[eidairaman1]

Maybe i should add, theres no stability problem at all with GPU-Z latest version running and 17.11.1 driver. But there are many crashes with 17.11.2 , but i did not test, if the crashes with 17.11.2 would persist without GPU-Z running.

Thank you for affirming what works instead of complaining like some members here do and have a chip on their shoulder.

 

[Naki]

O

Over loaded with work but enough time to come here n post that it happened once.
I think TOMSHARDWARE misses you!!! Go back towards the light!!!!

My freeze was most definitely NOT once. It was many times.
Anyway, all fine now - in the recent 1 month, month and a half I have got no such freezes any more. Maybe one of the (semi-)recent AMD Radeon drivers updates helped, and/or same for Windows 10 updates/new builds I installed.
Using 17.11.2 drivers now, will update today to 17.11.4.

RE TomsHardware forums, please note I check out some of the eNews links they send me from time to time, but I do not usually do English forums. I use local native ones instead, being in Bulgaria.
The only exceptions are these GPU-Z forums, and Core Temp forums too.

Maybe i should add, theres no stability problem at all with GPU-Z latest version running and 17.11.1 driver. But there are many crashes with 17.11.2 , but i did not test, if the crashes with 17.11.2 would persist without GPU-Z running.

What crashes? Please be more specific. Does the whole PC crash with a BSOD/GSOD?
Do apps other than GPU-Z crash, or GPU-Z itself?
If other apps crash, list their names.

Also, kindly note this thread is not related to any crashing, but freezing. This is not the same thing. If you get crashes, please start a new thread for your issues, or bump a proper thread, which this is not.

 

[vegaio]

The monitor goes blank and sound comes to a halt after 5-10 seconds. So i have to manually hard reboot. But only with 17.11.2 . 17.11.1 there is no problem at all. I run 17.11.1 several weeks now without any problems and while running GPU-Zlatest version.

 

[Naki]

Thanks! Okay, very similar to what I got a few times. In my case no sound mostly, as issue happens on startup, so no music or music videos are playing, yet. And I do not get a black/blank screen, the image on the screen for me was whatever showed on screen when GPU-Z caused the freeze. (and as I said, the issue seems fixed for me now)
What you report is a freeze, not a crash. Freeze & crash are not the same thing, please note.

I did get what you say on my other PC with an older Radeon R9 290 card a while ago, but I rarely use that one directly (I use it for my work with remote access to it).

Check out 17.1.14 drivers - out since a few days. Maybe it could help?
It has some known issues, so be prepared for some:
http://support.amd.com/en-us/kb-art...son-ReLive-Edition-17.11.4-Release-Notes.aspx
(I think it may cause issues with MS Excel & MS Word. AMD do not say which exact software they mean, but usually it is that, as there were all kinds of issues previously with Microsoft Office with some of these non-WHQL drivers.)

Also, December rename/rebrand of the AMD Radeon drivers is coming soon! Hope the drivers will get better with every week.
It will be named: Adrenalin Edition

http://links.em.experience.amd.com/...MwNjQ2NDExNzAyS0&j=MTE2NDA3NzM1NQS2&mt=1&rt=0

Hmm, music sounds very Matrixy to me.

 

[vegaio]

Seems like there is a prob with HWinfo:

http://www.overclock.net/t/1634018/official-vega-frontier-rx-vega-owners-thread/4980#post_26482254
http://www.overclock.net/t/1634018/official-vega-frontier-rx-vega-owners-thread/4980#post_26482547

I had some crashes too with the latest driver and even one time with the formerly thought stable 17.11.1 (no HWINFO, but GPU-Z running).

 

[Naki]

Thanks for the update!
Finally AMD Radeon Adrenalin Edition is out, a few days ago. If you want, please check if it changes anything for you.

 

[Jon8RFC]

I wanted to chime in that I, too, have this issue. It occurs reliably with two monitors, gpu-z on the second monitor, and running Counterstrike: Global Offensive.

It happened reliably on an old q9550, HD5870, Win10, two monitors and CSGO. It also happens reliably with an i7-7700, GTX 1070, Win10, two monitors and CSGO. Just one monitor seems to not be an issue, and I have no problems until I run gpu-z while utilizing two monitors and occasional lock ups which I can't reproduce, but it will always lock up with CSGO. I can't recall if it happened on the q9550 system with Win7.

 

[W1zzard]

I wanted to chime in that I, too, have this issue. It occurs reliably with two monitors, gpu-z on the second monitor, and running Counterstrike: Global Offensive.

It happened reliably on an old q9550, HD5870, Win10, two monitors and CSGO. It also happens reliably with an i7-7700, GTX 1070, Win10, two monitors and CSGO. Just one monitor seems to not be an issue, and I have no problems until I run gpu-z while utilizing two monitors and occasional lock ups which I can't reproduce, but it will always lock up with CSGO. I can't recall if it happened on the q9550 system with Win7.

Are you using latest GPU-Z? I think the crashes were fixed a few versions ago

 

[Jon8RFC]

Are you using latest GPU-Z? I think the crashes were fixed a few versions ago

I'm using 2.17.0 at the moment, which is the version released 10 days ago, and it causes lock ups very reliably. Every version prior to 2.17.0, for about the past year that I can confidently recall (and likely longer than that), has also caused the lock ups.

 

[W1zzard]

thanks, i'll do more testing with csgo then

 

[Jon8RFC]

thanks, i'll do more testing with csgo then

Oh, I did some searching here and found this:
https://www.techpowerup.com/forums/...o-steam-valve-anti-cheat.248101/#post-3914194
That build, with SHA-1 hash E4E3FDDE1EB8ADAD8A8B1F3876A3401A7BC30E14 let me play a good 15-20 minutes and no lock ups, but I don't know if it's anti-cheat related and if it's coincidental that the mode I was playing didn't have it running, or if it only runs once per load to get a successful "it ran" logged and then not again for an hour, or until reloading the game, etc. I ran 2.17.0 immediately after, and it didn't lock up after about 4-5 minutes, and it would reliably lock up within about a minute prior to using the linked version, without even starting a game. So it may be coincidental. I tested earlier this morning and 2.17.0 reliably locked up anywhere between 30-45 seconds of pressing the play button to load the game.

EDIT:
I thought I was in the clear, so I fired up 2.17.0 and it ran for about 5-8 minutes but locked up during a competitive match, so I'm guessing it's somehow anti-cheat related.

EDIT2:
I tried again with 2.11.0 and it lasted all the time in the menu and in the warmup of the competitive match, but once the competitive match actually began, it locked up within a minute.

 

[W1zzard]

Just to confirm, you are seeing no issues with a single monitor, only on dual monitor?

Both HD 5870 and GTX 1070

 

[Jon8RFC]

On the older system with the hd5870 it seemed more reliably unstable with dual monitors, but I only really hooked up that monitor when I really needed it, like to see gpuz and task manager on a whim, and it was likely after I had been playing a game, such as csgo. Lots of smoke on a particular map would bog down my system, so I'd occasionally check with gpu-z mid-match and plug in my second monitor and it would often not be a problem, leading me to believe it was some dual-monitor configuration. If I preemptively plugged in my second monitor and ran gpu-z once I saw what map was going to be played, it would lock up rather quickly. Now, on my current system, I leave my monitor plugged in all the time and would run gpu-z often upon boot up, which would be before I'd run steam, and whenever I'd fire up csgo it would lock up. Just poor coincidences on my part, sorry.

I did more thorough testing today, with just the gtx 1070 system, to get you a better answer. It's seemingly not related to monitor setup. It's related to time when steam is run (and most likely, when VAC runs) and how long csgo is run before gpu-z is run...as of today, and how that software is designed.

1. If steam is started, gpu-z run, and then csgo run...within 120 seconds (it was between 15-65 seconds, but I used 120 to play it safe) of the game interface being displayed on the screen, it will lock up.
2. If steam is started, csgo run, and 120 seconds allowed to pass after the game interface is displayed on the screen, gpu-z can then be run, and then...
   1. ...it will not lock up for quite a long time. What the specific range of time is, I don't know.
   2. ...if CSGO alone is exited and run again while gpu-z was running the whole time, it will still not lock up for a long period of time.
   3. ...if steam is fully exited, run again, and CSGO run again, while gpu-z was running the whole time, it WILL lock up within 120 seconds of sitting once the game interface is displayed.
   4. (not fully tested) ...a competitive match is started, it will likely lock up within 120 seconds of a competitive match successfully passing the warmup phase and starting.

This is as of today, and how the steam-VAC functionality is tied together. It's a matter of timing to get it to lock up, and I'm very strongly inclined to think that it's related to when VAC runs, which appears to (as only one part I can observe) be tied to when steam was last run or logged in to, and either VAC or gpu-z get stuck when trying to access the same low-level hardware functions, maybe? I'm no programmer, just trying to provide information so that you can form your own, better ideas.

EDIT: edited item 2.3 to specify that CSGO has to be run again as well

 

[W1zzard]

Amazing testing, I'll try to reproduce

 

[Jon8RFC]

Wow, I'm sorry to be full of misleading information, but I'll leave the previous stuff rather than edit it, because maybe it'll be relevant if/when steam/vac changes things in the future.

Rather than hastily jumping into CSGO, it seems like it also locks up after a successful steam login, as long as gpu-z has been running before that event. Sitting at the Steam login prompt is fine for 2 minutes, sitting at the 2-factor authentication after entering name&password is fine for 2 minutes. Upon successful login is when it locks up reliably after a short period. This is unless or until they change how and when these apparent VAC scans occur.

If steam is run, and a short period is allowed to pass, gpu-z can be run. This explains why I've been able to play PUBG for an hour, or more, without an issue--I boot up, run steam, check my email, then run the game and decide I want to monitor it with gpu-z. However, if steam is run and that short period passes, gpu-z can be run without issue ONLY until CSGO (and maybe other current, VAC-enabled games?) is run for the first time within a particular steam login session, and the computer locks up shortly after. It's not necessarily exclusively a data integrity check, but a system memory or hardware check. I tinkered with this while doing a large steam download and as best I can tell, it doesn't care about download data integrity. I tested with a Factorio download, no issue, and then a Dota 2 download and it was fine throughout and after completion, so long as gpu-z was run after the initial 2-minute safety buffer when steam logged in. Again, I'm not doing anything other than observing, so this is all circumstantial speculation from a non-programmer and for anyone reading in the future, it's likely that this activity changes over time to circumvent cheaters, especially since it a lot of it seems to have very narrow-window time triggers.

EDIT:
I downloaded and tried two other VAC-enabled games Toxikk and Dota 2. Toxikk seems like an abandoned game, so it played fine for 10 minutes and I moved on to Dota 2. Dota 2, on the other hand, is pretty popular like CSGO and it did lock up in the sub 60-second range, even after Steam had run and passed what I assume are its initial VAC checks. I ran gpu-z after many minutes of a successful Steam login, and had no issue until shortly after Dota 2 had loaded.

 

[Sora]

You probably have something else running that implements the cpuid sdk such as corsair link / icue or one of the hwmon core based utilities that do not get along safely with gpu-z and the vac driver.

 

[Jon8RFC]

You probably have something else running that implements the cpuid sdk such as corsair link / icue or one of the hwmon core based utilities that do not get along safely with gpu-z and the vac driver.

I quit everything else to eliminate any chances of conflicts when confirming gpu-z and steam/vac, and don't run anything unusual like you mentioned, anyway. I'm a pretty thorough edge-case troubleshooter.

Are you not able to replicate it yourself with a non-integrated graphics chip and a current vac-enabled game (that's the key--such as csgo or dota 2) on your steam account, Sora?
Have steam completely exited, run gpu-z, then sign on to steam and wait 2 minutes.

I don't know if vac runs dependent upon the account having used current vac-enabled games within a certain time period or not, so it's important to keep the test case similar before declaring "no problems".

I did send the VAC team an email a few weeks ago, and mentioned this situation, but haven't heard back.

 

[Sora]

No, i can't.

 

[lexluthermiester]

Wow, this seems like one of those "perfect storm" of circumstances situations. This will be interesting to watch as it unfolds.

 

[Jon8RFC]

As I suspected, Steam/VAC runs scans based on the last successful scan or successful intervals of scans.

I've run Steam many times in recent weeks without gpu-z. I fired it up and no crash...yet. I fired up CSGO and no crash...yet. Then I decided to play a game of CSGO and after about 5-7 total minutes from having run steam, it locked up.

Upon reboot, because there was a previously failed VAC check, running gpu-z and then steam produced a lock up after about 40 seconds from successful logon to Steam. I was able to repeat this, and it was not a one-time occurrence. It needs a successful check within some time period before it goes into a limbo-state for the logon lock up until it's triggered by a different event.

So, all you need to do (for now, until the VAC team changes its intervals) is let it lock up just once (you have to be patient and let it do its thing in a current VAC game--play a full competitive match in CSGO while gpu-z has been running since before you ran steam), and then it will reliably lock up very quickly upon steam logon as long as you run gpu-z before running steam. Once you let it have a successful scan, then it will seemingly be ok unless you run a current, VAC-enabled game for a substantial period of time such that it triggers a new scan.

 

[W1zzard]

play a full competitive match in CSGO

I've been trying to reproduce but my CSGO level is too low for competitive, and normal quickmatch seems to be running fine.

Also, what graphics card did you use?

 

[Jon8RFC]

Since mid-February it's been a GTX 1070 and I've been installing updated drivers when it notifies me. Before that, it was an HD5870, which doesn't have any newer drivers than at least a year old. Two, independent computers.

If you have a debug-type build you want to send my way, I can run that and get some logs to you. This is pretty interesting to me.

 

[W1zzard]

The problem is that I need a way to reproduce locally, so I could investigate a bunch of things that could cause it

I've seen the problem before, but thought I had fixed it since a few versions ago.