Weird behavior with Chrome

[Vasllo]

I've had chrome tabs crash several times with my "new" PC, which never happened in my low-end notebook (i5 2450M, 6GB and GT540M). Why so? I recently upgraded to 16GB RAM, so that is not an issue at all. And to make matters worse, I just had an unusual crash right now, I went to Facebook, with other tabs open, and opened 7 notifications, one in each tab, as I usually do, and my PC froze for a second, then went to a black screen then seemed like restarted the explorer and everything else, it went to my desktop and reopened everything I had open. What the f* is up with that?

 

And an extra question, my PC boots really fast sometimes and not so fast others (quite often) and have been slower to both boot and shut down lately, with a 2 month old legit Windows installation. Should I consider formatting considering all this BS? I really don't want to go through that, but, if I need to...

 

My whole setup is in my signature.

[Tabs]

If you disable gpu acceleration, do the crashes stop happening? The freeze you mentioned sounds like it may have been the gpu driver crashing and being recovered by Windows - or as you said, perhaps explorer crashing.

 

When you set the machine up with Windows 2 months ago, was it a fresh install using a clean Microsoft image? I'm assuming you're on Windows 10, but if not, it may be helpful to know which Windows version you're running.

 

I notice you're also overclocking - if these crashes started since installing more memory, it's possible you may need to increase your uncore voltage to handle twice as many modules, assuming the new modules are confirmed good.

[Vasllo]

36 minutes ago, Tabs said:

If you disable gpu acceleration, do the crashes stop happening? The freeze you mentioned sounds like it may have been the gpu driver crashing and being recovered by Windows - or as you said, perhaps explorer crashing.

 

When you set the machine up with Windows 2 months ago, was it a fresh install using a clean Microsoft image? I'm assuming you're on Windows 10, but if not, it may be helpful to know which Windows version you're running.

 

I notice you're also overclocking - if these crashes started since installing more memory, it's possible you may need to increase your uncore voltage to handle twice as many modules, assuming the new modules are confirmed good.

Haven't tried to disable GPU acceleration yet, but that wouldn't be a good long-term solution, I think, rigtht? But I'll try it. It makes sense to be the GPU. But only in this one instance, the other ones are just tab crash with the sad face and all.

 

Yes, it's Windows 10 x64 Home and it was a clean install. And I'm pretty sure my modules are fine, but I'll check them aswell. And voltage is actually a possibility, because I set my voltage to as tight as possible when I had 8GB. But I hope I don't have to increase it, I don't really want to go over the 70mV extra I have atm, I don't really want to go farther then the current 1.328-1.36v.

 

Edit: May that GPU crash be due to my undervolting aswell? I mean, it's pretty damn stable so far, so that would be weird, but possible I guess. That would end my plan to go for 1.1v

[Tabs]

10 minutes ago, Mateus Campello said:

Haven't tried to disable GPU acceleration yet, but that wouldn't be a good long-term solution, I think, rigtht? But I'll try it. It makes sense to be the GPU. But only in this one instance, the other ones are just tab crash with the sad face and all.

 

Yes, it's Windows 10 x64 Home and it was a clean install. And I'm pretty sure my modules are fine, but I'll check them aswell. And voltage is actually a possibility, because I set my voltage to as tight as possible when I had 8GB. But I hope I don't have to increase it, I don't really want to go over the 70mV extra I have atm, I don't really want to go farther then the current 1.328-1.36v.

 

Edit: May that GPU crash be due to my undervolting aswell? I mean, it's pretty damn stable so far, so that would be weird, but possible I guess. That would end my plan to go for 1.1v

I wasn't suggesting you increase your core voltage, I was suggesting you look at your IMC voltages (uncore, VCCSA, VCCIO etc) to ensure that your system remains stable with extra modules. I doubt voltage changes to your gpu could be at play here, but it's worth checking anyway - most of the voltage changes people make to their gpu are to the P0 performance state, so they only affect when it's in 3d rendering mode (like a game), not in any of the interim P-states like when rendering windows or playing video.

 

As for chrome and gpu rendering - gpu rendering helps chrome use less power and be more responsive, so you'll want to re-enable it once you determine if it is fixed by being turned off. If turning it off fixes it then you know you should more closely scrutinise your gpu setup. 

[Vasllo]

1 hour ago, Tabs said:

I wasn't suggesting you increase your core voltage, I was suggesting you look at your IMC voltages (uncore, VCCSA, VCCIO etc) to ensure that your system remains stable with extra modules. I doubt voltage changes to your gpu could be at play here, but it's worth checking anyway - most of the voltage changes people make to their gpu are to the P0 performance state, so they only affect when it's in 3d rendering mode (like a game), not in any of the interim P-states like when rendering windows or playing video.

 

As for chrome and gpu rendering - gpu rendering helps chrome use less power and be more responsive, so you'll want to re-enable it once you determine if it is fixed by being turned off. If turning it off fixes it then you know you should more closely scrutinise your gpu setup. 

Well, fml, just got a WHEA_Uncorrectable_Error BSOD... Any idea of how much I should increase those voltages?

[Tabs]

33 minutes ago, Mateus Campello said:

Well, fml, just got a WHEA_Uncorrectable_Error BSOD... Any idea of how much I should increase those voltages?

Are you definitely, 100% sure that the new sticks you added are completely stable/good? Have you been able to successfully run any memtest on them? 

 

Would you consider removing the extra memory for a day or two to see if the BSOD's disappear? Perhaps run only the new memory to see the same.

 

WHEA_Uncorrectable_Error means there's a hardware issue. WHEA stands for Windows Hardware Error Architecture, and is the core component of enabling things like error correction for ECC memory modules, and handling hardware faults. When it encounters an error it cannot correct... it causes that bluescreen to prevent data corruption.

 

You can generally go +0.1v on the voltages I listed earlier safely, and often much further, but I'd recommend taking a few small steps at first to see if stability improves.

[Vasllo]

2 hours ago, Tabs said:

Are you definitely, 100% sure that the new sticks you added are completely stable/good? Have you been able to successfully run any memtest on them? 

 

Would you consider removing the extra memory for a day or two to see if the BSOD's disappear? Perhaps run only the new memory to see the same.

 

WHEA_Uncorrectable_Error means there's a hardware issue. WHEA stands for Windows Hardware Error Architecture, and is the core component of enabling things like error correction for ECC memory modules, and handling hardware faults. When it encounters an error it cannot correct... it causes that bluescreen to prevent data corruption.

 

You can generally go +0.1v on the voltages I listed earlier safely, and often much further, but I'd recommend taking a few small steps at first to see if stability improves.

So, I just gave VCSSA and VCCIO 90mV and ran MemTest86, which took 1.5h, and got no errors. I'll try again tomorrow with default voltages to check if I need the 90mV bump and reinstall my graphics driver with DDU. I just updated it, but I'll reinstall it anyways just to make sure. So, tomorrow I hope I get everything fixed, because now I gotta sleep, because this test took seriously long and I'm seriously late, lol, but my anxiety couldn't wait. Thanks for all the help, mate, for real.

[Vasllo]

13 hours ago, Tabs said:

Are you definitely, 100% sure that the new sticks you added are completely stable/good? Have you been able to successfully run any memtest on them? 

 

Would you consider removing the extra memory for a day or two to see if the BSOD's disappear? Perhaps run only the new memory to see the same.

 

WHEA_Uncorrectable_Error means there's a hardware issue. WHEA stands for Windows Hardware Error Architecture, and is the core component of enabling things like error correction for ECC memory modules, and handling hardware faults. When it encounters an error it cannot correct... it causes that bluescreen to prevent data corruption.

 

You can generally go +0.1v on the voltages I listed earlier safely, and often much further, but I'd recommend taking a few small steps at first to see if stability improves.

Just finished the MemTest86 again with stock VCCIO and VCSSA voltages and also got no errors. Also reinstalled my graphics driver with DDU, I hope it's fixed now. Thanks for all the help.