First, a basic overview of the system would be helpful. inxi -bz would provide that. -b is brief. -z removes S/N information and other things that could be tracked. Always helpful to post that with any HW issue, so people have a place to start.
Use the nvme tool, that's the name, rather than smartctl, for NVMe storage information and health.
There have been UEFI viruses found that can infect an OS. My understanding is that they are OS specific and write the virus boot code into the firmware of the motherboard. Of course, MS-Windows is the main target, but there have been Linux-specific variants as well. Both are blocked by using UEFI + SecureBoot. You might check that is you are booting with UEFI, that SecureBoot is also enabled. Being in a dualboot setup with a supported version of MS-Windows restricts your easy ability to use Legacy/BIOS booting in Linux, but it is possible with some hassle (have to toggle the BIOS boot type every time).
When you say "things are slow", that's extremely broad. Hard for anyone to get specific enough to really help. If you could use some specific programs that are commonly used, perhaps someone who also uses those tools would be able to speak to issues?
As a comparison,
Code:
$ sudo hdparm -Ttv /dev/nvme0n1p4
/dev/nvme0n1p4:
HDIO_DRIVE_CMD(identify) failed: Inappropriate ioctl for device
readonly = 0 (off)
readahead = 512 (on)
HDIO_DRIVE_CMD(identify) failed: Inappropriate ioctl for device
geometry = 953117/64/32, sectors = 1951985039, start = 1540096
Timing cached reads: 41144 MB in 1.99 seconds = 20676.33 MB/sec
HDIO_DRIVE_CMD(identify) failed: Inappropriate ioctl for device
Timing buffered disk reads: 5072 MB in 3.00 seconds = 1690.32 MB/sec
and
Code:
$ sudo hdparm -Tt --direct /dev/nvme0n1p4
/dev/nvme0n1p4:
Timing O_DIRECT cached reads: 4546 MB in 2.00 seconds = 2274.29 MB/sec
HDIO_DRIVE_CMD(identify) failed: Inappropriate ioctl for device
Timing O_DIRECT disk reads: 5370 MB in 3.00 seconds = 1789.69 MB/sec
So, it seems that the performance of your SSD is massively slower than mine. No NVMe SSD should be that slow (yours, not mine). Here's the SMART information for my NVMe drive:
Code:
$ sudo nvme smart-log /dev/nvme0n1
Smart Log for NVME device:nvme0n1 namespace-id:ffffffff
critical_warning : 0
temperature : 37 C
available_spare : 100%
available_spare_threshold : 10%
percentage_used : 2%
data_units_read : 15,149,741
data_units_written : 32,251,324
host_read_commands : 144,685,574
host_write_commands : 590,073,169
controller_busy_time : 536
power_cycles : 27
power_on_hours : 1,508
unsafe_shutdowns : 13
media_errors : 0
num_err_log_entries : 0
Warning Temperature Time : 0
Critical Composite Temperature Time : 0
Temperature Sensor 1 : 37 C
Temperature Sensor 2 : 46 C
Thermal Management T1 Trans Count : 0
Thermal Management T2 Trans Count : 0
Thermal Management T1 Total Time : 0
Thermal Management T2 Total Time : 0
It is a Samsung 980 using an m.2 slot on the B450 motherboard. Less than 50% of the storage is allocated for use, so there are plenty of cells available for load balancing. As SSD cells get older (more writes to the same cells), they get slower. I've had 2 SSDs go bad, but they were 1st generation SSDs, so nothing like what we all have today. They got slower and slower and slower, would turn "read-only" and cause OS crashes. Also, this system is running Ubuntu Server 20.04, not 22.04.
I have a KIOXIA SSD in a Dell laptop (product: KXG60ZNV512G NVMe KIOXIA 512GB). It's not powered on now, so I can't easily get the data from that device. Screw it ... I'll go turn it on. I remember being a little disappointed in the SSD when the laptop arrived. Of course, I wiped it and put a Linux on it immediately. I don't dual boot. Ok, it is booted and unlocked. For laptops, I always use LUKS encryption. Also, it has LM 22 on it, not Ubuntu 22.04.
Code:
$ sudo hdparm -Ttv /dev/nvme0n1p3
/dev/nvme0n1p3:
readonly = 0 (off)
readahead = 256 (on)
geometry = 486164/64/32, sectors = 995663872, start = 4550656
Timing cached reads: 36182 MB in 2.00 seconds = 18126.49 MB/sec
Timing buffered disk reads: 4438 MB in 3.00 seconds = 1479.29 MB/sec
and
Code:
$ sudo hdparm -Tt --direct /dev/nvme0n1p3
/dev/nvme0n1p3:
Timing O_DIRECT cached reads: 4018 MB in 2.00 seconds = 2008.54 MB/sec
Timing O_DIRECT disk reads: 6878 MB in 3.00 seconds = 2292.27 MB/sec
and FWIW,
Code:
$ sudo nvme smart-log /dev/nvme0n1
Smart Log for NVME device:nvme0n1 namespace-id:ffffffff
critical_warning : 0
temperature : 25 °C (298 K)
available_spare : 100%
available_spare_threshold : 10%
percentage_used : 13%
endurance group critical warning summary: 0
Data Units Read : 41069646 (21.03 TB)
Data Units Written : 58587882 (30.00 TB)
host_read_commands : 890019342
host_write_commands : 2323965034
controller_busy_time : 2779
power_cycles : 828
power_on_hours : 9714
unsafe_shutdowns : 23
media_errors : 0
num_err_log_entries : 3
Warning Temperature Time : 0
Critical Composite Temperature Time : 0
Temperature Sensor 1 : 25 °C (298 K)
Thermal Management T1 Trans Count : 0
Thermal Management T2 Trans Count : 0
Thermal Management T1 Total Time : 0
Thermal Management T2 Total Time : 0
And FWIW, the laptop HW overview:
Code:
$ inxi -bz
System:
Kernel: 6.8.0-48-generic arch: x86_64 bits: 64
Console: pty pts/3 Distro: Linux Mint 22 Wilma
Machine:
Type: Laptop System: Dell product: Latitude 5320 v: N/A serial: <superuser required>
Mobo: Dell model: 0Y7GXY v: A00 serial: <superuser required> UEFI: Dell v: 1.36.0
date: 03/22/2024
Battery:
ID-1: BAT0 charge: 49.5 Wh (100.0%) condition: 49.5/61.8 Wh (80.1%) volts: 15.6 min: 15.2
CPU:
Info: quad core 11th Gen Intel Core i5-1145G7 [MT MCP] speed (MHz): avg: 579 min/max: 400/4400
Graphics:
Device-1: Intel TigerLake-LP GT2 [Iris Xe Graphics] driver: i915 v: kernel
Device-2: Realtek Integrated_Webcam_HD driver: uvcvideo type: USB
Display: x11 server: X.org v: 1.21.1.11 with: Xwayland v: 23.2.6 driver: X:
loaded: modesetting unloaded: fbdev,vesa dri: iris gpu: i915 tty: 115x31 resolution: 1920x1080
API: OpenGL v: 4.6 compat-v: 4.5 vendor: mesa v: 24.0.9-0ubuntu0.2 note: console (EGL sourced)
renderer: Mesa Intel Xe Graphics (TGL GT2), llvmpipe (LLVM 17.0.6 256 bits)
Network:
Device-1: Intel Wi-Fi 6 AX201 driver: iwlwifi
Device-2: ASIX AX88179 Gigabit Ethernet driver: ax88179_178a type: USB
Drives:
Local Storage: total: 476.94 GiB used: 10.43 GiB (2.2%)
Info:
Memory: total: 16 GiB note: est. available: 15.35 GiB used: 3.26 GiB (21.2%)
Processes: 324 Uptime: 15d 23h 38m Init: systemd target: graphical (5) Shell: Bash
inxi: 3.3.34
If is a very stable system. The uptime is really mostly standby time. It goes into and out of standby without any issues. I don't do Wayland and avoid using wifi. Plenty of CPU and RAM for a laptop. Only 1 internal SSD, barely used. I don't keep anything important on laptops.
If I wanted to see real storage performance, I'd use fio. I know there are GUI performance testing tools in KDE and I think Gnome-Disks has something, but neither of those tools are on my systems. Here's an article about running fio tests with large reads, large writes, random reads and random writes and mixed workloads. https://arstechnica.com/gadgets/2020...-way-with-fio/ explains how to use fio.
Most of my "servers" run 5-10 virtual machines, so the storage is mostly random reads+writes 24/7. Also, the servers have NFS storage, so consistent streaming to other devices of larger files happens. All devices are wired GigE connected and there's a caching DNS on my LAN to prevent slow responses. Slow DNS can make a system feel slow.
RAM and CPU are things we can't do anything about, unless we tweak the kernel to disable known CPU protections against attacks. I've never done that. The default protections are enabled. I do see that you have mixed speed in your RAM. I had issues with RAM in my first Ryzen 2600 after adding 16GB (2x8GB) more to the system. In theory, it was all the same part number from the same manufacturer and all certified for 3200Mhz, but in practice, that caused an unstable system when all 4 slots were filled. If I used either of the two pairs at 3200Mhz, it worked fine, but with both pairs in the system, I had to slow down to 2666 Mhz for stability reasons. I fixed that later, by getting 32GB of RAM in 2x16GB sticks. Automatic OC using the built-in MB OC tool provides 7% more performance - or so it claims. I didn't try to manually achieve higher OC. The systems the Ryzen was replacing were all significantly slower - Pentium or 1st-gen Core i5 systems, so 8-10x slower. I didn't bother with bench marking before/after the RAM changes. I sold the used RAM to friends for cheap. With Ryzen, I found that only "matched" RAM worked for me. I spent far too much time tweaking speeds to get the fastest I could while still being stable. I understand that Intel CPUs/MBs aren't picky at all.
I would concentrate efforts on the SSD performance. That is clearly an issue.
Anyway, some ideas for your consideration. These forums are about to go read-only. Beware of that.
