[SOLVED] PC 'Memory' crashes with 100% working memory?!?

[ddule.srb]

I can't stress it enough, i am really getting annoyed by my system in the last month. I have the next system:

RX 570 Sapphire NitroX+ 4GB ( Bought used)
Ryzen 5 1600 ( New)
HyperX 1x8 GB 2666Mhz ( New)
Asus A320M-K ( New)
Kingston A400 240gb SSD ( New)

The system started causing artifacts that last for one frame every so often that don't create lag or stutter or anything, but come from all shapes and sizes and they are not like the permanent artifacts that you can see on really damaged GPUs like squares, big chunk of pixels with different colors, a grid of cubes, No! Just a '1 frame black/white lines, NO SIGNAL type of color mush' that i even can't describe because i cannot focus on them when they happen. Now that is mostly a GPU issue but what really stresses me is that my applications, games, background tasks are being shut down and crashing mostly without error outputs and some which do give outputs, output "ERROR: OUT OF MEMORY" or "Insufficent memory, app crashed" type of outputs. Opera browser frequently crashes with "Page crashed due to lack of memory. Refresh?". My RAM usage isn't even high, it sits at around 3-3.5gb with a couple of apps used in the background and using a paging file when i game.And all of the apps crash at once which means something causes it inside the system and it all crashes. Sometimes it freezes completely and i have to hard-boot the system. I am tech-savy but i do not have spare components to test all of the components individualy. This does not happen when the GPU is stressed, more so, it doesn't even happen 99% of the time while i'm playing games. The Page file in Win10 is properly adjusted with 8-16gb of SSD storage( I do not have an HDD), i ran Memtest86 on a USB flash drive and it has done 3 tests for 6 hours with 0 ERRORS so RAM is clear. The Processor under 100% load ALSO does not re-create these crashes, i've done 100% stress tests of t he GPU and tried to fully allocate 4096mb of my GPU and while using all of it's VRAM it doesn't do anything, BUT...i do manage to re-create it by playing a game, then alt-tabing OUT of the game and watching a youtube video, safe to say it happens ALOT while browsing the web and having small apps running in the background. I, actually, saw ALOT of artifacts during videos i watch on youtube while DotA2 is finding a match in the background but when i start playing and the browser is still behind, it doesn't happen. Does ANYONE know what the issue is? I am pointing my finger at the GPU but why don't these things happen while gaming? Help me, i'm desperate as the crashes can be at any point in time and i don't know how to avoid them and it straight out becomes annoying.

 

[MJS WARLORD]

How much ram are you using , have you looked in reliability monitor to see error reports , look at them and google the error report.

 

[ddule.srb]

How much ram are you using , have you looked in reliability monitor to see error reports , look at them and google the error report.

I have an 8GB stick and while idle watching videos, it uses about 3-3.5gb but when i game, it's mostly capped at 7.8-8gb and using an additional memory from the SSD paging file at around additional 6-8gb, depends on what i am doing. The SSD is fine, HDD sentinel defragmented and healty status of 90% health with 100% perfromance (it's about a year old). I will check the error reports and come back to you in about a minute.

 

[ddule.srb]

How much ram are you using , have you looked in reliability monitor to see error reports , look at them and google the error report.

I've just looked into the report monitoring and there is not specific critical errors that are suspicious. It just shows Opera crashing and "RuntimeBroker.exe". About a few days back, it was "LogonUI.exe" and "SearchUI.exe", day before that "bad_module_info". I had a problem on that logonui day where my search wasn't automaticly working and i had to re-create the system with command prompt, as if something deleted the search and logon UI. I fixed that, but the Windows itsels is like, a month old as, as soon the artifacts started and apps started crashing, i thought the crashes were to the system and the GPU created artifacts on it's own so i wanted to fix one thing and to admit the gpu died but the crashes continued so...the system is relatively fresh.

 

[mamasan2000]

Paging file shouldn't be on an SSD, they don't like it. Bad for longevity. Second, pagefile is really slow. Not really meant as a replacement for low amounts of RAM. Maybe if you have MS Word running in the background.
I would get 8 gigs more. Or just close all other apps before gaming and don't open many tabs in Opera. Some games jsut require 16 gigs minimum, others even more.

 

[hotaru.hino]

Paging file shouldn't be on an SSD, they don't like it. Bad for longevity.

I don't see how that's true when Microsoft way back in Windows 7 did a study and reported the following (from https://docs.microsoft.com/en-us/archive/blogs/e7/support-and-qa-for-solid-state-drives)

Should the pagefile be placed on SSDs?

Yes. Most pagefile operations are small random reads or larger sequential writes, both of which are types of operations that SSDs handle well.

In looking at telemetry data from thousands of traces and focusing on pagefile reads and writes, we find that

• Pagefile.sys reads outnumber pagefile.sys writes by about 40 to 1,
• Pagefile.sys read sizes are typically quite small, with 67% less than or equal to 4 KB, and 88% less than 16 KB.
• Pagefile.sys writes are relatively large, with 62% greater than or equal to 128 KB and 45% being exactly 1 MB in size.

In fact, given typical pagefile reference patterns and the favorable performance characteristics SSDs have on those patterns, there are few files better than the pagefile to place on an SSD.

Considering in my experience I've kept the pagefile on the SSD in my last three builds and they've had at worst something like 2TBW per year, for modern drives, I think I'm going to die before I exhaust the TBW endurance rating.

 

[Mike2009]

Early Ryzens i heard had little hiccup issues,, my experience is i run 3200mhz 16gb dual channel on my ryzen 3600 you can get this set i run for 63 bucks its corsair vengeance LPX , I can see this system is a low budget system ,, but ryzen really benefits from dual channel high speed memory , plus not every ram stick is compatible with every motherboard. could be a matter of compatibility ,,,, one little thing off with it. but ,,,, it also could still be your graphics card ,, the ram on it c going bad.

 

[ddule.srb]

Early Ryzens i heard had little hiccup issues,, my experience is i run 3200mhz 16gb dual channel on my ryzen 3600 you can get this set i run for 63 bucks its corsair vengeance LPX , I can see this system is a low budget system ,, but ryzen really benefits from dual channel high speed memory , plus not every ram stick is compatible with every motherboard. could be a matter of compatibility ,,,, one little thing off with it. but ,,,, it also could still be your graphics card ,, the ram on it c going bad.

The system is 1 year old but still in warranty therefore i call it "New". Just so i can say that if something is bad, it could be the GPU because it was bought used also a year ago. And the system was working PERFECTLY for that whole year but this started maybe a month ago so it isn't due to compatibility or bottlenecking.

 

[ddule.srb]

in my last three builds and they've had a

I don't see how that's true when Microsoft way back in Windows 7 did a study and reported the following (from https://docs.microsoft.com/en-us/archive/blogs/e7/support-and-qa-for-solid-state-drives)



Considering in my experience I've kept the pagefile on the SSD in my last three builds and they've had at worst something like 2TBW per year, for modern drives, I think I'm going to die before I exhaust the TBW endurance rating.

He meant the longevity of the SSD which is not calculated by hours or days in HDD but actually by Terrabytes written on the SSD which constatly writes more data on the SSD shortening it's lifespan. That's how the SSDs calculate lifespan and yes, it's true what's he saying but it's kinda off topic. I'm looking for a solution or a temporary fix on my crashes during usage.

P.S
The speed thing of the Paging file on the SSD isn't actually true, just the longevity of it is. I've been strictly using for the whole year my SSD as a paging file storage and it NEVER made a stutter or whatever. It mostly depends of what games you're trying to run. It certainly made a hiccup here or there in COD Warzone but just because the game had like 90GB of storage which was mostly being used real-time because of the bloody size of the games textures, animations and details peaking my Page file up to 22-24gb. Just played Resident Evil 3 Remake and total RAM usage + Page file didn't rise over 12-14GB.

 

[hotaru.hino]

He meant the longevity of the SSD which is not calculated by hours or days in HDD but actually by Terrabytes written on the SSD which constatly writes more data on the SSD shortening it's lifespan. That's how the SSDs calculate lifespan and yes, it's true what's he saying but it's kinda off topic. I'm looking for a solution or a temporary fix on my crashes during usage.

The problem with the whole "bad for longevity" is that the fact of even using the SSD normally is "bad for longevity." Things are written to the SSD all the time. If you want the SSD to last forever, you'd treat it like a CD-R: write to it once and be done with it. Worrying about the longevity of an SSD is like worrying if Yellowstone will explode tomorrow.

And while yes, while hitting the page file a lot is a sign of trouble, if something is going to hit the page file, I'd rather have it be on the fastest thing that can service it.

So I'll leave it at that since I don't have anything else to say that's relevant to the original post.

 

[Colif]

Could be a driver memory leak.

you might want to look into poolmon to see what process is asking for memory and not giving it back. Its often drivers but I don't see any common culprits on your motherboard website - https://www.asus.com/au/Motherboards/PRIME-A320M-K/HelpDesk_Download/

I would get more ram as using it all for gaming sure doesn't leave much space for windows. Although win 10 can run in 500mb of ram, its really slow when its mostly running off the page file.

 

[Rogue Leader]

Paging file shouldn't be on an SSD, they don't like it. Bad for longevity.

Need to butt in here in case anyone has read and listened, but this is REALLY bad advice, if there was every a way to kneecap the hell out of your system its to force the pagefile to a hard drive over your SSD.

SSD wear is not a thing to ever worry about, we have mods here with absolutely ancient SSDs that are barely showing wear. The SSD will be obsolete before you ever have to worry about wearing it out.

 

[USAFRet]

Paging file shouldn't be on an SSD, they don't like it. Bad for longevity.

"Longevity" for an SSD, regarding too many write cycles, is a concept that is long past.

Given sufficient RAM, the pagefile is rarely used.

Consider systems that are SSD only. Like mine.

 

[USAFRet]

He meant the longevity of the SSD which is not calculated by hours or days in HDD but actually by Terrabytes written on the SSD which constatly writes more data on the SSD shortening it's lifespan. That's how the SSDs calculate lifespan and yes, it's true what's he saying but it's kinda off topic. I'm looking for a solution or a temporary fix on my crashes during usage.

P.S
The speed thing of the Paging file on the SSD isn't actually true, just the longevity of it is. I've been strictly using for the whole year my SSD as a paging file storage and it NEVER made a stutter or whatever. It mostly depends of what games you're trying to run. It certainly made a hiccup here or there in COD Warzone but just because the game had like 90GB of storage which was mostly being used real-time because of the bloody size of the games textures, animations and details peaking my Page file up to 22-24gb. Just played Resident Evil 3 Remake and total RAM usage + Page file didn't rise over 12-14GB.

Yes, we know exactly what he was referring to.
Completely a non-issue.

 

[Colif]

most ssd will be obsolete before they wear out. Most ssd don't get the amount of usage needed to wear out.

 

[mamasan2000]

https://www.enterprisestorageforum.com/storage-hardware/ssd-lifespan.html "Measuring SSD/Flash Endurance"

Write cycles is going down, not up. Old drives ARE better. Having a pagefile on an SSD means lots of writecycles because Windows is constantly writing something to SSD. Threadstarter doesn't have enough memory to turn off pagefile. So extensive use of pagefile. And Windows acts like a little kid that got stolen of his candy if you disable pagefile. Gotta have some small pagefile.

Another factor is drive space. It is already limited on my SSDs and NVMe SSD, I don't want a useless pagefile taking up even more of it .

You are free to do whatever you want. I choose to do it this way, for the above mentioned reasons.

 

[Rogue Leader]

https://www.enterprisestorageforum.com/storage-hardware/ssd-lifespan.html "Measuring SSD/Flash Endurance"

Write cycles is going down, not up. Old drives ARE better. Having a pagefile on an SSD means lots of writecycles because Windows is constantly writing something to SSD. Threadstarter doesn't have enough memory to turn off pagefile. So extensive use of pagefile. And Windows acts like a little kid that got stolen of his candy if you disable pagefile. Gotta have some small pagefile.

Another factor is drive space. It is already limited on my SSDs and NVMe SSD, I don't want a useless pagefile taking up even more of it .

You are free to do whatever you want. I choose to do it this way, for the above mentioned reasons.

You can feel free to go ahead and keep doing that. Again real world has shown us minimal impact on SSD life. If you want to kneecap your own system, be my guest but we collectively will call it out as its poor advice and a non-issue for any normal use case.

Now lets stop taking this thread off on a tangent as this is not the OP's issue

Side note: the font size of your post was reduced, do not post in large fonts like that, its disruptive to the thread especially that we are now off topic.

 

[USAFRet]

https://www.enterprisestorageforum.com/storage-hardware/ssd-lifespan.html "Measuring SSD/Flash Endurance"

Write cycles is going down, not up. Old drives ARE better. Having a pagefile on an SSD means lots of writecycles because Windows is constantly writing something to SSD. Threadstarter doesn't have enough memory to turn off pagefile. So extensive use of pagefile. And Windows acts like a little kid that got stolen of his candy if you disable pagefile. Gotta have some small pagefile.

Another factor is drive space. It is already limited on my SSDs and NVMe SSD, I don't want a useless pagefile taking up even more of it .

You are free to do whatever you want. I choose to do it this way, for the above mentioned reasons.

"Write cycles is going down, not up."
Your link says nothing of the kind. Or maybe I missed it. Care to parse out the relevant text for this?

As I've asked many people...Have you personally ever had an SSD die from too many write cycles, in normal consumer use?
To date, I've not heard a Yes.
Die from something else, yes. But not from too many write cycles.

Even approaching the warranty TBW number is exceedingly rare.


And you shouldn't turn the page file OFF anyway. There needs to be at least a minimal one. Even just 1GB.

 

[Colif]

Turning page file off wouldn't fix this guys problem, it would only make it happen a lot faster since he is using most of his ram during gaming and then another 6gb of page file.

I have an 8GB stick and while idle watching videos, it uses about 3-3.5gb but when i game, it's mostly capped at 7.8-8gb and using an additional memory from the SSD paging file at around additional 6-8gb, depends on what i am doing.

so where the page file is located is meaningless really as it doesn't solve the problem. They shouldn't be getting oom errors if its managed by windows

Minimum should be - Varies based on page file usage history, amount of RAM (RAM ÷ 8, max 32 GB) and crash dump settings.
Maximum should be - 3 × RAM or 4 GB, whichever is larger. This is then limited to the volume size ÷ 8. However, it can grow to within 1 GB of free space on the volume if required for crash dump settings.
https://docs.microsoft.com/en-us/windows/client-management/determine-appropriate-page-file-size

now i don't know what his is set as but it should be at least 16gb and therefor his usage shouldn't be creating oom errors. Why I suggested he look into poolmon - https://docs.microsoft.com/en-us/windows-hardware/drivers/devtest/poolmon - as it is likely a driver that is asking for memory allocations and not giving it back once its finished, so it grows and grows until your PC runs out of memory to allocate.

 

[mamasan2000]

"Write cycles is going down, not up."
Your link says nothing of the kind. Or maybe I missed it. Care to parse out the relevant text for this?

As I've asked many people...Have you personally ever had an SSD die from too many write cycles, in normal consumer use?
To date, I've not heard a Yes.
Die from something else, yes. But not from too many write cycles.

Even approaching the warranty TBW number is exceedingly rare.


And you shouldn't turn the page file OFF anyway. There needs to be at least a minimal one. Even just 1GB.

• SLC: Single-level cell NAND flash supports 50,000 to 100,000 write cycles.
• MLC: The 2-bit data multi-level cell (MLC) flash generally takes up to 3,000 write cycles. eMLC (enterprise MLC) sustains up to10,000 write cycles, and can reach 35,000 cycles on 3D NAND.
• TLC: Triple-level cells (3-bit) NAND flash is low at 300-1000 write cycles, and can achieve 1500-3000 write cycles with 3D NAND.

I even mentioned the Headline for it.
Pretty sure QLC is even worse.

 

[USAFRet]

• SLC: Single-level cell NAND flash supports 50,000 to 100,000 write cycles.
• MLC: The 2-bit data multi-level cell (MLC) flash generally takes up to 3,000 write cycles. eMLC (enterprise MLC) sustains up to10,000 write cycles, and can reach 35,000 cycles on 3D NAND.
• TLC: Triple-level cells (3-bit) NAND flash is low at 300-1000 write cycles, and can achieve 1500-3000 write cycles with 3D NAND.

I even mentioned the Headline for it.
Pretty sure QLC is even worse.

And drive size is going up, leaving far more spaces to level with.

Samsung warranty TBW over time.

SSD Product Warranty | Support | Samsung Semiconductor Global

Samsung ensures the best quality in all products and services beyond customer expectations. Warranty policy, conditions for Samsung Consumer and Portable SSD.

www.samsung.com

All at the 512GB size:
840 Pro - 73TBW
850 Pro - 300TBW
860 Pro - 600TBW

950 Pro - 400TBW
960 Pro - 400TBW
970 Pro - 600TBW

Samsung seems to think it isn't an issue.

Again...have you personally had or seen any consumer lever SSD that died from too many write cycles, in normal use?


This is like buying nice tires for your car, but never using it because you don't want to wear them out.


Now..if you buy 3rd rate devices, all bets are off.

 

[Rogue Leader]

• SLC: Single-level cell NAND flash supports 50,000 to 100,000 write cycles.
• MLC: The 2-bit data multi-level cell (MLC) flash generally takes up to 3,000 write cycles. eMLC (enterprise MLC) sustains up to10,000 write cycles, and can reach 35,000 cycles on 3D NAND.
• TLC: Triple-level cells (3-bit) NAND flash is low at 300-1000 write cycles, and can achieve 1500-3000 write cycles with 3D NAND.

I even mentioned the Headline for it.
Pretty sure QLC is even worse.

You conveniently ignore the part about how as the Cells have gone down in write cycles the density has significantly increased, so while you get less writes per cell, the amount of times that cell needs to be written during normal operations has significantly decreased because the data is broken up more amongst more cells.

Again, can you point to your own experience or anyone's real life experience with drive wearing out? It doesn't exist because it doesn't happen. Just like everyone wants a PCIe 4 NVME drive for the fastest speed, but other than in a lab during a benchmark real life use doesn't hit those speeds, nor will you wear out a drive.

Now again, time to put on my moderator hat, LETS DROP THIS AS ITS IRRELEVANT TO THE OP. If you'd like to continue to discuss this feel free to start a discussion thread and you can even invite us to it. But knock it off in here.