[SOLVED] Screen freezes on heavier games

[TurboTukker]

Greetings everyone and thanks in advance for your time!

To start with my system is the following:

C:\ Samsung 850 EVO 250 GB SSD (2015) Holds Windows and a few games
D:\ Samsung Spinpoint 250 GB 7200 rpm HDD (2006) Only used as back-up of photos and documents
E:\ Seagate Expansion Portable 4 TB (2019) Main source of photos, documents and large (so most) games.

Power: Seasonic G Series 550 (2013)
Gigabyte GA-H77M-D3H Motherboard (2013)
Corsair Vengeance 16 GB (2x 8 GB) (2013)
Intel Core i5 3570 / 3.4 Ghz (2013)

Geforce GTX 1660 SUPER (2020)

I built the system myself in 2006, updated the motherboard in 2013 and just yesterday I finally replaced the AMD HD5700 video card (I know I know) with a GTX 1660 SUPER. Now with the replacing I had hoped to solve the problem, however it has only worsened somehow:

When playing games that are heavier on the system Middle Earth Shadow of War, Call of Duty Modern Warfare for example everything will run lovely and smooth and then at some point suddenly the screen freezes completely and I am left with no other option than to restart. The exact characteristics of the freeze differ a bit but the following happens:

• Screen freezes and sound stops entirely
• Keyboard and mouse LED are still on but non responsive
• Fans in the case are still on
• I have left it in that state for about 10 minutes waiting to see if anything would happen but nothing

I can browse, watch movies or do any non-gaming stuff on the computer all night without trouble. I can play Rimworld all night without trouble. But when I ask more of the computer it will start doing the freezes. With the old videocard this would usually be about 3 crashes per evening, but on that old videocard I could obviously not even run the heavier games mentioned.

I have timed the last two crashes I had on Middle Earth Shadow of War on max setting and they happened at 3 minute 30 in actual gamplay (not the menu) and then 5 minutes later it lasted 6 minutes.

I have already checked the following possible solutions to no avail:

• Ran a check of the computer with Malwarebytes
• Ran a check with Avast Antivirus
• Updated all drivers I could think of, even went desperate and installed Driver Booster 7 to check
• Ran a check on the RAM memory with no errors found
• Scanned all hard disks for errors with the windows tool with no errors found
• Checked the CPU temperature using Core Temp while in-game, never goes beyond 75 degrees celsius
• Cleaned the entire inside of the case of dirt, made sure all fans were running

All in all I am really at a loss on what is going wrong here. Just to add, this system without the new videocard and without the external HDD has been running fine for years, but then suddenly the crashes sometimes started happening.

Like I started with, thanks in advance for your time and any help or advice given, it is truly much appreciated!

TurboTukker

 

[Grobe]

Very nice HW presentation, wish everybody could start threads like that.

First off, I'd suspect the PSU since it's old.

I'd like you to run some benchmark software (3Dmark or similar) on the computer, and then see if that makes the computer to stall.

 

[TurboTukker]

Thank you for the swift response Grobe!

I ran two of the "similar" benchmark options on the PC:

UserBenchmarks: Game 64%, Desk 59%, Work 37%

CPU: Intel Core i5-3570 - 65.8%

GPU: Nvidia GTX 1660S (Super) - 70.6%

SSD: Samsung 850 Evo 250GB - 52.4%

HDD: Samsung SP2504C 250GB - 25.5%

HDD: Seagate Expansion 4TB - 51.3%

RAM: Corsair Vengeance LP DDR3 1600 C10 2x8GB - 51.9%

MBD: Gigabyte GA-H77M-D3H

And the second:

Computer ran through them both with no visible trouble at all. I even let the Unigine Valley one run for a good 10 minutes without officially benchmarking it.

To me nothing on either of these benchmarks seems off, but I hope it gives you something to go on!

 

[Grobe]

You can also try to run OCCD, and provide a screenshot. It should provide a 12V measurement that can say something about the voltage is ok or not.

 

[TurboTukker]

I did a lot of googling en think you meant OCCT? I installed that and the power test came up with the following result:

And in idle the voltage stats were as follows:

The only thing I do notice is that in the test the CPU cores were going up to rather high temperatures. Which is odd because with the CPU stress test the max temperature given was about 80 celsius. Would that somehow be the problem after all?

 

[TurboTukker]

I just figured I'd run the CPU OCCT thing again together with Core Temp to double check the CPU temperature values and 38 seconds into the test it actually did the freeze. So it is doable outside of a game as well.
I made a picture of the values of OCCT and Core Temp in the frozen screen, those are the last values measured before the freeze.

And:

 

[Grobe]

The voltage seems to be within limits, only CPU_VCORE seems to drop more (per cent).

Since you've replaced the GPU, I doubt that is the issue. However, the mainboard is aging - and thus I'd assume that the motherboard is the source of the problem. There is already an indicator on this because of the CPU_VCORE - that in turn should be regulated by the motherboard.

 

[TurboTukker]

Oof.. I was sincerely hoping the motherboard would not be a problem, because that of course is the most expensive one to replace. Not the motherboard itself per say, but if I replace the motherboard then the CPU and probably RAM would need replacing as well.

I did check again with both tests running and indeed the CPU temperature during the test is slowly pushed into the high 90 celsius, which is way too hot I think. I was more or less hoping a replacement of the processor cooler would be enough to tide me over a while again.

I think I may start by getting a CPU cooler that is also compatible with an 1151 socket in case the motherboard does need replacing.

 

[TurboTukker]

And a few hours more of thinking brings a relatively rigorous decision. I'm going to replace the motherboard, CPU and RAM on the computer. About time I do that anyhow! Now I only cross my fingers that it's not the PSU after all.
The PSU will of course be first on my list to replace after this, but I'm hoping it will stick with me a while longer.

Replacement thoughts so far is relatively simple:

AMD Ryzen 5 2600 processor
Gigabyte GA-A320M-S2H motherboard
RAM: Hyperx Impact 16 GB 2666 Mhz

Thanks a lot for the thinking along and advise Grobe! It is truly much appreciated!