Is there an issue with my SSD?

Bageland2000

Limp Gawd
Joined
Jan 19, 2013
Messages
331
I’m having an issue with things loading really slowly all of the sudden on my Dell XPS 15 (9550). I wish I could say what changed right before this started, but I can’t. Seems like everything I try to do is slow, like it’s struggling to read from the NVMe (maybe? Still trying to figure out the problem). One thing that happened for instance recently was that I tried to install the latest 960M driver. The package extraction process took like three minutes. Then, the installation process took about ten minutes. Also, for instance, when I click on “This PC” It takes a few seconds to come up, and the progress bar in the folder path area just continues to progress slowly.

xRPO2cg.png


I ran Crystal Bench, from what I can tell, the speeds are fine, but I don’t totally know what I’m looking for:

4yxq0Yw.png


CPU usage during is around 30%, memory utilization is around 50% during some loads (like installing the driver package.) Can anyone help me figure out what the issue is? Drivers/BIOS are all the latest, windows is fully updated. Plenty of SSD capacity unused. I can’t think of the next troubleshooting step…
 
Don't think so. The mapped drives are super stable, and I don't think network issues would cause the behavior I saw in the compressed install package taking forever to extracts. This just happened too BTW...

SNS282F.jpg


I had another recent BSOD for memory management...
 
I keep opening "This PC" and it is taking forever to show all the drives. Is there any way to "ask" the computer, "which drive are you looking for that you can't find?" I feel like that's the keep to the lack of responsiveness on my PC.
 
Did you check top see if windows is working on downloading updates in the background?
 
Go into the bios setup and slow your memory down.
See if that helps.
 
resource monitor?
I'll open that up and get back to you on this.

Did you check top see if windows is working on downloading updates in the background?
It's not. Irt's fully up-to-date and not working on new ones

Go into the bios setup and slow your memory down.
See if that helps.
I guess I can try this, but I don't think the memory freqs have changed recently, so not sure why this would be the culprit.
 
Have you checked temps? Sounds like it may be overheating and throttling -- like an internal fan stopped working or something.
 
Totally agree that you should check the smart of the drive first. Smart should show you if the drive has any issues. In addition, you can try to boot from Live USB and check the SSD.
 
Have you checked temps? Sounds like it may be overheating and throttling -- like an internal fan stopped working or something.

I ran AIDA64 stress test for 20 minutes. The temps barely budget and didn't come close to causing throttling.

Totally agree that you should check the smart of the drive first. Smart should show you if the drive has any issues. In addition, you can try to boot from Live USB and check the SSD.
I used crystal disk to check, looks like no SMART issues at all.

resource monitor?
OK, so I opened Chrome and four game launchers (Blizz, Steam, GOG, and Epic). I noticed some serious program hangs, so I opened resource manager and saw I was at 80% RAM utilization and seeing about 50-800 memory hard faults/sec. Could it simply be that 8Gb is not enough RAM? I think maybe those slow downs were hard disk thrashing?
https://appuals.com/hard-faults-per...faults are a normal,the physical memory (RAM).

IDK, it doesn't seem like it would explain why extracting a driver package took forever when I didn't have any other programs open... Still trying to find out what might be going on.
 
I ran AIDA64 stress test for 20 minutes. The temps barely budget and didn't come close to causing throttling.


I used crystal disk to check, looks like no SMART issues at all.


OK, so I opened Chrome and four game launchers (Blizz, Steam, GOG, and Epic). I noticed some serious program hangs, so I opened resource manager and saw I was at 80% RAM utilization and seeing about 50-800 memory hard faults/sec. Could it simply be that 8Gb is not enough RAM? I think maybe those slow downs were hard disk thrashing?
https://appuals.com/hard-faults-per-second/#targetText=Hard faults are a normal,the physical memory (RAM).

IDK, it doesn't seem like it would explain why extracting a driver package took forever when I didn't have any other programs open... Still trying to find out what might be going on.

chek ressoruce monitor you can direclty see if system is acdessing pagefile alot under the hdd section. but it does look like a lot of swapping
 
What drive is it? Your sequential write speed seems to be outside the SLC cache and your 4K results are generally quite poor for NVMe. Give the sequential read speed, I'm guessing it's a 4-channel controller or otherwise limited to x2 PCIe 3.0 or x4 PCIe 2.0, or both (some Dells limit to GT2/x2 but that's mostly the 13", but could be related). Post the CrystalDiskInfo please.
 
What drive is it? Your sequential write speed seems to be outside the SLC cache and your 4K results are generally quite poor for NVMe. Give the sequential read speed, I'm guessing it's a 4-channel controller or otherwise limited to x2 PCIe 3.0 or x4 PCIe 2.0, or both (some Dells limit to GT2/x2 but that's mostly the 13", but could be related). Post the CrystalDiskInfo please.
EK0Lsa4.png
 
Thanks. Samsung's OEM drives start with "PM" if they're TLC-based ("3-bit MLC") and "SM" if they're MLC-based ("2-bit MLC"). The numbers following tell you the generation; the PM951 comes prior to the 960 EVO (PM961) and so utilizes 2D/planar TLC and the older ARM Cortex-R4 based UBX controller. These technical details likely don't mean much to you and that's fine.

It is possible to use the Samsung NVMe driver for these drives. I have a post on Reddit that includes a link to the driver as well as a pre-extracted version if necessary (depends on OS). Further, the lower section of my post details how to make a bootable Linux drive if you want to check error logs and such outside of the Windows environment.

Despite its age, the drive supports x4 PCIe 3.0; this is not listed under "Transfer Mode" in your CDI screenshot. You can try the program Hard Disk Sentinel and see if it has information on this. However, your performance results don't seem too far off for that drive, and I don't see any other issues with the SMART data, although you might want to check the error logs as there appears to be a lot (this is not atypical for OEM Samsung drives).
 
Thanks. Samsung's OEM drives start with "PM" if they're TLC-based ("3-bit MLC") and "SM" if they're MLC-based ("2-bit MLC"). The numbers following tell you the generation; the PM951 comes prior to the 960 EVO (PM961) and so utilizes 2D/planar TLC and the older ARM Cortex-R4 based UBX controller. These technical details likely don't mean much to you and that's fine.

It is possible to use the Samsung NVMe driver for these drives. I have a post on Reddit that includes a link to the driver as well as a pre-extracted version if necessary (depends on OS). Further, the lower section of my post details how to make a bootable Linux drive if you want to check error logs and such outside of the Windows environment.

Despite its age, the drive supports x4 PCIe 3.0; this is not listed under "Transfer Mode" in your CDI screenshot. You can try the program Hard Disk Sentinel and see if it has information on this. However, your performance results don't seem too far off for that drive, and I don't see any other issues with the SMART data, although you might want to check the error logs as there appears to be a lot (this is not atypical for OEM Samsung drives).
Thank you! I understand the basics of what you said regarding the drives specs. I will definitely look into using the Samsung NVMe driver as you outlines there.
 
I'm going to assume you aren't trying to run two antivirus programs simultaneously. This is always my FIRST thought when I see performance issues like this, but usually it's on machines belonging to people that don't know a damn thing about computers. You seem more than competent, but double check installed programs anyway, something might have sneaked in on you during an install of another program.

Also, your RAM usage does seem a bit high. I also assume you've tried shutting it down, leaving it completely off for a minute or so, then restarting it. A memory leak could be the culprit, and is hard to narrow down because it won't show in processes.

If it isn't doubled-up antivirus, and doesn't fix itself on cold start, this honestly sounds like a Windows issue to me, either a weird services issue, or god knows what (Thanks Windows 10!). Try running in Safe Mode to see if you still have the same problem, though it could still be Windows even if the problem persists in Safe Mode. It's just a good way to limit the number of running services. Seems Windows 10 devs like to use us as Beta Testers.

A fresh install might be in order if you can't narrow it down though. I would try that first before buying a new drive. Then if the new install doesn't work, you aren't out much time reinstalling again on a new drive. Or clone the drive if you have another NVMe slot or external adapter you can use. I would recommend going ahead and updating the RAM at this point too while you have the machine torn apart. We're right at that point where 8GB just isn't enough IMHO.
 
Yeah, for a main drive, that's not a ton if it's a few years old. I have similar sized drives with well over 100TB writes, who cares about reads though?
 
I'm going to assume you aren't trying to run two antivirus programs simultaneously. This is always my FIRST thought when I see performance issues like this, but usually it's on machines belonging to people that don't know a damn thing about computers. You seem more than competent, but double check installed programs anyway, something might have sneaked in on you during an install of another program.
That's correct. I run Malwarebytes alone. I assume that having Windows Defender enabled in its default status would not be causing issues.

Also, your RAM usage does seem a bit high. I also assume you've tried shutting it down, leaving it completely off for a minute or so, then restarting it. A memory leak could be the culprit, and is hard to narrow down because it won't show in processes.
Yes, I've done this several times over the last few weeks

If it isn't doubled-up antivirus, and doesn't fix itself on cold start, this honestly sounds like a Windows issue to me, either a weird services issue, or god knows what (Thanks Windows 10!). Try running in Safe Mode to see if you still have the same problem, though it could still be Windows even if the problem persists in Safe Mode. It's just a good way to limit the number of running services. Seems Windows 10 devs like to use us as Beta Testers.
That's what I've been thinking with the BSODs and such. Trying to keep all the possibilities open though.

A fresh install might be in order if you can't narrow it down though. I would try that first before buying a new drive. Then if the new install doesn't work, you aren't out much time reinstalling again on a new drive. Or clone the drive if you have another NVMe slot or external adapter you can use. I would recommend going ahead and updating the RAM at this point too while you have the machine torn apart. We're right at that point where 8GB just isn't enough IMHO.
Obviously trying to avoid that, but it's clearly an option. I think that RAM upgrade is my best option. Probably would be a benefit even if there's a separate root cause.

22TB of reads, 11TB of writes? Did you buy this drive used?
I did indeed.

kek, my PM SSD has almost 100 tb of read/writes on it, under 25TB is pretty low.
Do we need to have a Zoolander walk-off now?
 
OK, I think it's likely that the Hard Faults are a cause (or at least symptom?) of the problem. When things are crawling along, it seems to coincide with high HF/s/ See the two snapshots below. I ran Memtest extensively with 0 errors.
qJJSUxl.png

jHZtDEY.png
 
Now THIS is strange... I just ran an AIDA 64 stress test while I have CPU-Z open. The multiplier field shows a variable range between 8-35. But from what I can tell, it isn't increasing past the minimum 8x multiplier! No thermal issues at all, not throttling. It just seems like the multiplier isn't increasing. If this what's actually happening, I think I can safely say I figured out the issue. What the heck is going on!?
 
OK, so I did some research, and someone suggested unchecking "BD PROCHOT" on Throttlestop. That fixed it! But... someone said that if this fixes it, it's an indicator that there's a broken sensor on the mother board, and that this will require a motherboard swap out. Can someone who knows this for certain confirm if this is accurate?
 
Back
Top