Something seems wrong with my RAM

El_Syd · 1 Dec 2018 at 01:03

My setup is as follows:

CPU: AMD - Ryzen 7 2700X
Motherboard: ASRock - X470 Taichi
Memory: G.Skill - Trident Z 16 GB F4-3200C16D-16GTZKW
Storage: Samsung - 960 EVO 250 GB M.2-2280 SSD
Storage: Crucial - MX300 1.05 TB 2.5" SSD
Storage: Seagate - Barracuda 2 TB 3.5" HDD
Video Card: GALAX - GeForce GTX 1080 8 GB EX OC SNPR
Power Supply: SeaSonic - PRIME Ultra Platinum 650 W 80+ Platinum Certified Fully-Modular ATX
Operating System: Windows 10 Enterprise LTSC 2019

I built the machine all from new parts back around May this year. All I did was load an XMP profile in the BIOS and fuse the NVME (where OS installed) and SSD using StoreMI. Things have been fine until about 2 weeks ago, where I started experiencing strange issues, especially in The Division.

The first thing that I noticed was how every once in a while the game would suddenly jerk the camera around so that I was looking down towards the ground. At first, I thought that it was my mouse going crazy. Then one night, I had stop code errors:

https://imgur.com/a/k6ChENP

None of them got to 100%.

I didn't see them again, but I did start experiencing issues with booting up, where the PC would turn itself on and off about 5 times before booting up normally. It was around this time that I also started experiencing framerate issues. At 1440p with most settings at the highest in the Division, it seemed to range from about 90FPS to well below 50FPS at times. This wasn't the case before, as it used to be quite consistent at about 80FPS, so I suspected that something was up.

I also had a few occasions where I had to manually press the power button to shut down, as the hardware would still be going despite my monitor not outputting anything. It was at this point, that I opened up and re-seated the graphics card and RAM, as well as the CMOS battery (because I thought that would be the thing to do). That seemed to have sorted out the weird boot up issues, but the framerate was still bad.

I took my machine to a PC store for a check-up and they said that they couldn't find any issues. So I was left thinking that maybe it was a software issue. I decided to reformat everything and start fresh. I did a bit of reading and came across someone mentioning that StoreMI could result in system issues if the Tiered drive contained the OS and something were to happen while it was writing across two drives, causing corruption. I had even more reason to start over, so I did.

It took a while, but I am back to square one now. Disappointingly, I'm not seeing much of an improvement in framerate.

I'm suspecting that it could be the RAM. It's a G.Skill F4-3200C16D-16GTZKW, so it isn't the B-die kind that everybody praises. This is what can be seen in Task Manager:

https://imgur.com/a/XTsH4P2

This is the RAM speed that is also shown in the BIOS even though I loaded an XMP profile that should bring it to 3200 Mhz:

https://imgur.com/a/4O3iWqr

Also, I never noticed it before but, is it normal to have this much reserved for hardware?
Despite all this strangeness, Memtest showed no errors after running for about 9 and a half hours for 6 passes. If I try a single stick of RAM, it powers up but shuts off before it gets to the BIOS. Tried this with both sticks, each one in the number 2 slot. The system doesn't boot up properly until I have both slot 2 and 4 occupied. Is this normal?

I really need to figure out if something is going bad on the hardware side of things, particularly the RAM. If any of you need more details or pictures, let me know.

Jamin280672 · 1 Dec 2018 at 07:58

El_Syd said:
My setup is as follows:

CPU: AMD - Ryzen 7 2700X
Motherboard: ASRock - X470 Taichi
Memory: G.Skill - Trident Z 16 GB F4-3200C16D-16GTZKW
Storage: Samsung - 960 EVO 250 GB M.2-2280 SSD
Storage: Crucial - MX300 1.05 TB 2.5" SSD
Storage: Seagate - Barracuda 2 TB 3.5" HDD
Video Card: GALAX - GeForce GTX 1080 8 GB EX OC SNPR
Power Supply: SeaSonic - PRIME Ultra Platinum 650 W 80+ Platinum Certified Fully-Modular ATX
Operating System: Windows 10 Enterprise LTSC 2019

I built the machine all from new parts back around May this year. All I did was load an XMP profile in the BIOS and fuse the NVME (where OS installed) and SSD using StoreMI. Things have been fine until about 2 weeks ago, where I started experiencing strange issues, especially in The Division.

The first thing that I noticed was how every once in a while the game would suddenly jerk the camera around so that I was looking down towards the ground. At first, I thought that it was my mouse going crazy. Then one night, I had stop code errors:

https://imgur.com/a/k6ChENP

None of them got to 100%.

I didn't see them again, but I did start experiencing issues with booting up, where the PC would turn itself on and off about 5 times before booting up normally. It was around this time that I also started experiencing framerate issues. At 1440p with most settings at the highest in the Division, it seemed to range from about 90FPS to well below 50FPS at times. This wasn't the case before, as it used to be quite consistent at about 80FPS, so I suspected that something was up.

I also had a few occasions where I had to manually press the power button to shut down, as the hardware would still be going despite my monitor not outputting anything. It was at this point, that I opened up and re-seated the graphics card and RAM, as well as the CMOS battery (because I thought that would be the thing to do). That seemed to have sorted out the weird boot up issues, but the framerate was still bad.

I took my machine to a PC store for a check-up and they said that they couldn't find any issues. So I was left thinking that maybe it was a software issue. I decided to reformat everything and start fresh. I did a bit of reading and came across someone mentioning that StoreMI could result in system issues if the Tiered drive contained the OS and something were to happen while it was writing across two drives, causing corruption. I had even more reason to start over, so I did.

It took a while, but I am back to square one now. Disappointingly, I'm not seeing much of an improvement in framerate.

I'm suspecting that it could be the RAM. It's a G.Skill F4-3200C16D-16GTZKW, so it isn't the B-die kind that everybody praises. This is what can be seen in Task Manager:

https://imgur.com/a/XTsH4P2

This is the RAM speed that is also shown in the BIOS even though I loaded an XMP profile that should bring it to 3200 Mhz:

https://imgur.com/a/4O3iWqr

Also, I never noticed it before but, is it normal to have this much reserved for hardware?
Despite all this strangeness, Memtest showed no errors after running for about 9 and a half hours for 6 passes. If I try a single stick of RAM, it powers up but shuts off before it gets to the BIOS. Tried this with both sticks, each one in the number 2 slot. The system doesn't boot up properly until I have both slot 2 and 4 occupied. Is this normal?

I really need to figure out if something is going bad on the hardware side of things, particularly the RAM. If any of you need more details or pictures, let me know.

Forget the Samsung B Die thing, I was running G.Skill Trident Z RGB 3200mhz C16 (HYNIX) from day 1 of RyZen without any issues, brought some Samsung B die kits and don't see any difference, if its working then it doesn't matter that its Hynix based, what matters is its working at its rated speeds.

As for your issue, ive never used StoreMI so I cant comment on it, but one thing I did notice in one of your pictures there is that only 8gb of RAM in total with 62% usage and the other 8gb is hardware reserved, no that's not normal, you said the system wont boot with just one stick (which is strange on its own), you need to use both (16gb), this is pointing me to the Motherboard, sounds like a bad ram slot to me, slot 2 as that's the one you've tried to use with both sticks in single mode and failed to boot on both attempts, when you add the 2nd stick into slot 4 so now using slot 2 & 4 the system is booting on slot 4 only, hence task manager showing 8gb usable (slot 4), and 8gb hardware reserved (slot 2), it cant be read properly.

Try putting your RAM in slot 1 & 3 then booting with that, see what task manager says then, if all 16gb is available then you need to RMA your motherboard, faulty ram slot 2.

So much for the computer shop not being able to find a fault, hope you didn't pay them.

El_Syd · 1 Dec 2018 at 09:57

I sure hope I don't have a faulty motherboard or CPU. It's such a hassle to set everything up again. I'd rather deal with faulty RAM.

I tried slots 1 and 3. I still get the same result as what you see in the second picture here: https://imgur.com/a/4O3iWqr. The difference is that 'None' is listed next to A2 and B2, while A1 and B2 detect something. It's weird because it doesn't seem like the system is only detecting 8GB of RAM because the BIOS still shows 16GB as you can see in the picture. And Windows still says that there's 16GB, but half of it is 'Hardware reserved'.

Luckily the 'diagnostic' service was free. I think to them, just being able to boot up means that nothing is wrong.

WantoN · 2 Dec 2018 at 05:32

I had ram issues similar to this on my old rig. Turned out to be a dodgy motherboard. I’m not necessarily saying it’s the same case for you.

You need to confirm if the ram is the issue. Is there another machine you can try it in?

Jamin280672 · 2 Dec 2018 at 05:57

El_Syd said:
I sure hope I don't have a faulty motherboard or CPU. It's such a hassle to set everything up again. I'd rather deal with faulty RAM.

I tried slots 1 and 3. I still get the same result as what you see in the second picture here: https://imgur.com/a/4O3iWqr. The difference is that 'None' is listed next to A2 and B2, while A1 and B2 detect something. It's weird because it doesn't seem like the system is only detecting 8GB of RAM because the BIOS still shows 16GB as you can see in the picture. And Windows still says that there's 16GB, but half of it is 'Hardware reserved'.

Luckily the 'diagnostic' service was free. I think to them, just being able to boot up means that nothing is wrong.

The bios will detect it all, its just a detection, when it comes down to windows, its a different story, its not just about detection, its about if it can use it or not, in this case, Windows can still see it, it just cant use it so its dumped that 8gb into reserve, you say you've done memtest overnight for hours without any issues, the trouble with that is if you have both sticks installed, memtest might not of been able to read the other stick that windows is reserving too and just tested the one 8gb stick that it can read all night long. It might be worth doing that again but as WantoN said, in a different machine, test 1 stick at a time, this will rule out RAM.