MP 1,1-5,1 Windows 10 Locking Up When Gaming - Temps?

[dataharvested]

So, I just installed Windows 10 on its own SSD for gaming purposes and it will crash eventually mid-game. I did some reading around these forums around temperatures and I believe these temps may not be the issue causing the crashing. I play for about 15-20 minutes and entire system will freeze up with a buzzing sound coming through the speakers and I'm forced to power off the computer manually and restart it.

I did upgrade the CPU's and there is new thermal paste on them. But I did not learn about re-pasting the Northbridge until afterward and have not done that yet.

Just seeing if anyone thinks these temps (pictured) are of any concern. I'm going to have to find the replacement pins for the Northbridge before I can re-paste it.

 

[dataharvested]

I decided to forgo the new pins for the Northbridge. Honestly, if you are careful removing them they do go back in just fine. Just don't overly pinch or damage the plastic on the pins during removal and be sure to double check they have "spread" on the backside upon reinstallation. So they don't slip back through the board. I then applied new Arctic Silver thermal paste to the CPU's and the Northbridge. The thermal paste on the Northbridge was very dry and came off easily with just a dry cotton swab. I did go over the area with a little alcohol too to make sure the surface was properly prepared for the new paste.

I will report back if there are any changes in temps or the system crashing.

 

[dataharvested]

It got worse. :/

 

[dataharvested]

Another temp monitoring app, DIFFERENT results!

Not sure why that is. The troubleshooting continues....

BTW, both this and the previous screenshots were while I was playing the same game.

 

[dataharvested]

Installed MacsFanControl and put the fans on full blast. Temps are down now, but still crashing the computer to a complete standstill when playing a game. Installing a different game to try to narrow down the problem here. I also removed the USB 3.2 expansion card and the ATTO FastFrame card for troubleshooting purposes.

:|

 

[tsialex]

Your problem is not exactly temperature related, Mac Pros in rendering farms work for months/years without any hiccups with CPU temps around 90 to 95ºC - you are nowhere near that. You could have a defective PSU, or a dying GPU.

Also, your game could require an instruction that is not supported by your Xeon CPU, like AVX, that is used just somewhere in the game and not in the whole game, my son had found this before.

 

[dataharvested]

Your problem is not exactly temperature related, Mac Pros in rendering farms work for months/years without any hiccups with CPU temps around 90 to 95ºC - you are nowhere near that. You could have a defective PSU, or a dying GPU.

Also, your game could require an instruction that is not supported by your Xeon CPU, like AVX, that is used just somewhere in the game and not in the whole game, my son had found this before.

Ahh, thanks for the info.

 

[dataharvested]

I did some more reading on the AVX instruction set and it's being used in various software titles. Kind of a bummer, but what can you do. With the help of this community I resurrected a 2010 computer for use with current software. I shouldn't expect perfection with the modifications to get the computer where it is today (OCLP, macOS Monterey, Windows 10 Pro). I did have my hopes up for gaming some of my favorite titles in Windows, but it is what it is. I'm still going to try some of my other Steam games and see how they do though. I'm happy that the install is working so well otherwise!

 

[dataharvested]

It's seeming as though it may be the Xeon's inability to process some of the data from the games as mentioned by tsialex. GTA V fails to even load the game and I'm still getting random system lock-ups in PUBG. I have even tried a different graphics card with the same results. As a side note, I applied new thermal paste on my RX 580 this morning as more of a preventative measure. Other temps and voltages seem to be inline with the norm from what I've read. I might have to just build a Windows PC for gaming unfortunately.

 

[jscipione]

The Northbridge overheating typically happens on dual socket Mac Pros especially when combined with high watt CPUs, however it is rarely an issue with single socket models. Your temperatures also seem fine.

My guess is that there is a problem with your GPU overheating, either because of a hardware issue like it needs thermal paste or more likely a driver issue. I can’t tell what your GPU is from what you have posted but I would try to see if that could be the problem instead.

 

[dataharvested]

The Northbridge overheating typically happens on dual socket Mac Pros especially when combined with high watt CPUs, however it is rarely an issue with single socket models. Your temperatures also seem fine.

My guess is that there is a problem with your GPU overheating, either because of a hardware issue like it needs thermal paste or more likely a driver issue. I can’t tell what your GPU is from what you have posted but I would try to see if that could be the problem instead.

I applied new thermal paste to the Northbridge, CPU's, and GPU over the last few days. Yes, temps see to be good overall from what others have said and the apps I have used to monitor and test with. I have tried 2 different graphics cards, the AMD HD 5770 my Mac Pro came with and my Sapphire Pulse RX 580 8GB card with the same results, random locking up of the entire computer while gaming. I have used both the driver that Windows 10 installs automatically for the RX 580 and the driver that comes in the AMD software package (Adrenaline). The only drivers that I installed on the computer since installing Windows 10 the other day was the Realtek audio and Bluetooth drivers. I will investigate those drivers when I get home, but I'm doubting they will make any difference. :/

 

[jscipione]

No that all sounds fine I wouldn’t bother checking the audio and Bluetooth drivers. I have had crashes on RX 580 using custom drivers but the Adrenaline drivers from the AMD website should work fine. If the system is crashing even with the HD 5770 installed that indicates something else is the problem. Could it be a marginal stick of RAM perhaps?

 

[dataharvested]

No that all sounds fine I wouldn’t bother checking the audio and Bluetooth drivers. I have had crashes on RX 580 using custom drivers but the Adrenaline drivers from the AMD website should work fine. If the system is crashing even with the HD 5770 installed that indicates something else is the problem. Could it be a marginal stick of RAM perhaps?

I thought of checking the hardware like you mentioned, such as the RAM. I briefly read something about Apple diagnostic tools yesterday but have not looked into it further yet.

 

[eicca]

Mine did the same thing in both Windows and macOS due to wonky GPU firmware. I have the same GPU as you. Does yours have the BIOS switch? If so, shut down, flip the switch, and see if anything changes.

 

[dataharvested]

Mine did the same thing in both Windows and macOS due to wonky GPU firmware. I have the same GPU as you. Does yours have the BIOS switch? If so, shut down, flip the switch, and see if anything changes.

No, it doesn't have the switch. I believe it was version 2 that had to switch which would make mine version 1 from what I read online. I did have to flash the VBIOS on the card in order to get it to show the OpenCore Legacy Patcher boot selection and startup screens. I used what I believed to be a stock VBIOS ROM for the flash.

 

[Soba]

Since the crash is GPU independent, I suggest running Apple Hardware Test and/or Apple Service Diagnostics on your system to thoroughly test your RAM and many other low-level components.

More info here:

https://forums.macrumors.com/threads/cmp-5-1-aht-wont-run-bad-backplane.2332375/post-30802101

Also, I'll offer the generic advice of making sure Windows, the drivers, and games are all fully updated, but it sounds to me like a hardware problem.

 

[dataharvested]

Since the crash is GPU independent, I suggest running Apple Hardware Test and/or Apple Service Diagnostics on your system to thoroughly test your RAM and many other low-level components.

More info here:

https://forums.macrumors.com/threads/cmp-5-1-aht-wont-run-bad-backplane.2332375/post-30802101

Also, I'll offer the generic advice of making sure Windows, the drivers, and games are all fully updated, but it sounds to me like a hardware problem.

Thank you for the info.

 

[dataharvested]

Apple Service Diagnostics are currently running and I'll report back on the results.

I forgot to mention that I found this floating around in the Mac Pro's case when I received the computer off of eBay. It appears to be a capacitor off of the backplane. I'm pretty sure it's from the location in my pictures.

 

[macguru9999]

So, I just installed Windows 10 on its own SSD for gaming purposes and it will crash eventually mid-game. I did some reading around these forums around temperatures and I believe these temps may not be the issue causing the crashing. I play for about 15-20 minutes and entire system will freeze up with a buzzing sound coming through the speakers and I'm forced to power off the computer manually and restart it.

I did upgrade the CPU's and there is new thermal paste on them. But I did not learn about re-pasting the Northbridge until afterward and have not done that yet.

Just seeing if anyone thinks these temps (pictured) are of any concern. I'm going to have to find the replacement pins for the Northbridge before I can re-paste it.

A radeon 580 can be made to crash the mac pro by sucking too much juice through the backplane .... when i was using one i managed to do it once in a test, but i imagine you are doing it with your game ....

 

[tsialex]

Apple Service Diagnostics are currently running and I'll report back on the results.

I forgot to mention that I found this floating around in the Mac Pro's case when I received the computer off of eBay. It appears to be a capacitor off of the backplane. I'm pretty sure it's from the location in my pictures.

The ripped off cap is part of the PCIe slots PWM circuit, you should replace it right now.

If you can't do it yourself, get a replacement backplane.

 

[dataharvested]

Thanks for the feedback everyone. Looks like I'm going to need to do some repairing.

 

[dataharvested]

I may just go ahead and do the Pixlas mod while I'm at it. Just to be sure there's enough power for that card. Also would be good to have for future card upgrades. I'll do the capacitor repair first and then test to be sure that it's working, then do the power mod.

 

[tsialex]

I may just go ahead and do the Pixlas mod while I'm at it. Just to be sure there's enough power for that card. Also would be good to have for future card upgrades. I'll do the capacitor repair first and then test to be sure that it's working, then do the power mod.

You have to be real careful when desoldering the broken terminals since it will need a lot of heat and you will probably need to pre-heat the whole board, since the cap is connected to a power plane.

I'd pre-heat the board first and then try to heat the backside solder while pushing the terminals from the other side.

 

[dataharvested]

You have to be real careful when desoldering the broken terminals since it will need a lot of heat and you will probably need to pre-heat the whole board, since the cap is connected to a power plane.

I'd pre-heat the board first and then try to heat the backside solder while pushing the terminals from the other side.

Pre-heat the board how exactly? Use a heat gun?

 

[tsialex]

Pre-heat the board how exactly? Use a heat gun?

I usually do it with my kitchen oven, to around 100 to 120ºC, just to not have the power plane sink all the heat from the solder tip. It's a lot easier after you pre-heat and you don't have to set your solder station to crazy temperatures.