Is there a legitimate software that can check the "health" of physical system components?

[ClaustrophobicGames]

Recently I have been noticing that there is something off about my system. Things such as much lower performance in gaming than I would expect and general tasks in windows taking far too long to respond. For example I opened the file browser yesterday and actually had to wait for it to launch. Within the past year it has had a brand new SSD that received a fresh copy of Windows 11 Pro. I have checked every conceivable driver update, did some googling for answers, and closely monitored system telemetry through Hardware Monitor (CPUID). I haven't been able to figure anything out. 

 

This issue has made me wish that there is a software that could help me narrow down the possible issue. But we all know that there are a million "pc health checkers" that are either malware completely or include malware or malicious intent. So I figured the best thing to do would be to reach out to the community and see if anyone else has found a solution.

 

Below are the PC specs:

 

Ryzen 5800x (AIO water cooled)

Asus 3070 OC

MSI MPG B-550

32 GB of Corsair DDR4 @ 3200 MHz

2 TB 980 EVO PRO

(no manual overclocking or mining has been done on this system)

[DreamCat04]

7 minutes ago, ClaustrophobicGames said:

Recently I have been noticing that there is something off about my system. Things such as much lower performance in gaming than I would expect and general tasks in windows taking far too long to respond. For example I opened the file browser yesterday and actually had to wait for it to launch. Within the past year it has had a brand new SSD that received a fresh copy of Windows 11 Pro. I have checked every conceivable driver update, did some googling for answers, and closely monitored system telemetry through Hardware Monitor (CPUID). I haven't been able to figure anything out. 

 

This issue has made me wish that there is a software that could help me narrow down the possible issue. But we all know that there are a million "pc health checkers" that are either malware completely or include malware or malicious intent. So I figured the best thing to do would be to reach out to the community and see if anyone else has found a solution.

 

Below are the PC specs:

 

Ryzen 5800x (AIO water cooled)

Asus 3070 OC

MSI MPG B-550

Corsair DDR4 @ 3200 MHz

2 TB 980 EVO PRO

(no manual overclocking or mining has been done on this system)

Since you have an AIO, what are your temperatures like when these issues occur? It may be your pump that's failing but still working enough for your system to not shut down

[Brooksie359]

There are quite a few reasons why a pc can slowdown but I find that the biggest one is when your ssd is close to being full it's speed decreases noticeably and this can make your system feel slower as it actually is. The second reason is if you are running into ram limitations where you are running out of system ram and end up pagefiling which simply means you have to go to the ssd for the data rather than the data being stored on the ram which again ram is much faster than ssd so you would feel it. 

[manikyath]

storage self-reports 'smart' data to indicate it's own health, but past that your best bet is using something like hwinfo and looking at sensor data that seems off.

 

if all that seems fine, it's probably a software issue. (i.e. windows being brokendows.)

[ClaustrophobicGames]

4 minutes ago, DreamCat04 said:

Since you have an AIO, what are your temperatures like when these issues occur? It may be your pump that's failing but still working enough for your system to not shut down

My temps run around 26C-28C at idle and max out around 50C under gaming load. Running something like a 15 minute benchmark can push the temps as high as 65C-70C. But they are consistent between when I built the system and now. Also no noise from the pump.

[ClaustrophobicGames]

2 minutes ago, Brooksie359 said:

There are quite a few reasons why a pc can slowdown but I find that the biggest one is when your ssd is close to being full it's speed decreases noticeably and this can make your system feel slower as it actually is. The second reason is if you are running into ram limitations where you are running out of system ram and end up pagefiling which simply means you have to go to the ssd for the data rather than the data being stored on the ram which again ram is much faster than ssd so you would feel it. 

Good point. I don't know how much is too much though I have 550 GB of 2TB left. That has been very consistent over the past 2 months.

[DreamCat04]

Just now, ClaustrophobicGames said:

My temps run around 26C-28C at idle and max out around 50C under gaming load. Running something like a 15 minute benchmark can push the temps as high as 65C-70C. But they are consistent between when I built the system and now. Also no noise from the pump.

Then it might not be a hardware issue. That was my first thought, but what Brooksie359 said could be an issue, so can you do what they said?

[ClaustrophobicGames]

2 minutes ago, manikyath said:

storage self-reports 'smart' data to indicate it's own health, but past that your best bet is using something like hwinfo and looking at sensor data that seems off.

 

if all that seems fine, it's probably a software issue. (i.e. windows being brokendows.)

Windows is a little on the not super great side of things lol

[manikyath]

Just now, ClaustrophobicGames said:

Windows is a little on the not super great side of things lol

windows is like a golden retriever puppy.. it will play fetch just fine, but sometimes it just takes a big ol' dump on the carpet.

[ClaustrophobicGames]

5 minutes ago, DreamCat04 said:

Then it might not be a hardware issue. That was my first thought, but what Brooksie359 said could be an issue, so can you do what they said?

 

10 minutes ago, Brooksie359 said:

There are quite a few reasons why a pc can slowdown but I find that the biggest one is when your ssd is close to being full it's speed decreases noticeably and this can make your system feel slower as it actually is. The second reason is if you are running into ram limitations where you are running out of system ram and end up pagefiling which simply means you have to go to the ssd for the data rather than the data being stored on the ram which again ram is much faster than ssd so you would feel it. 

I will fire something up and keep an eye on ram usage.

[Mark Kaine]

38 minutes ago, ClaustrophobicGames said:

This issue has made me wish that there is a software that could help me narrow down the possible issue. But we all know that there are a million "pc health checkers" that are either malware completely or include malware or malicious intent. So I figured the best thing to do would be to reach out to the community and see if anyone else has found a solution.

 

Below are the PC specs:

 

Ryzen 5800x (AIO water cooled)

Asus 3070 OC

MSI MPG B-550

32 GB of Corsair DDR4 @ 3200 MHz

2 TB 980 EVO PRO

(no manual overclocking or mining has been done on this system)

userbenchmark actually does that... its not 100% reliable ( and you definitely need to do several runs, like 5-10 ideally) but its not "malware" and it can definitely show you if there are certain issues with your system... but it cant detect *all* issues for sure, and its therfore also not 100% reliable as said.

 

its still the best thing in that regard and the closest to what you're asking for (and again it definitely "can" help)

 

https://www.userbenchmark.com/Software

 

as said, let it run a couple of times and link the results here.

 

 

[Mark Kaine]

34 minutes ago, manikyath said:

storage self-reports 'smart' data to indicate it's own health,

yeah, but honestly something is off with that... like my infamous Kingston A400 drive... smart data says "good" no apparent errors etc, yet the speed is down to like 2.5mb/s max (doesn't matter if the drive is full or empty, new fw, old fw etc...)

 

And you know how i actually figured out what's slowing down my whole system (the Kingston "A400")? Yeap, userbench, told me right on the head the drive is faulty (probably simply by checking actual read/write speeds instead of "smart") 

 

Removed the drive and my pc immediately became "responsive" again. 

[manikyath]

5 minutes ago, Mark Kaine said:

userbenchmark actually does that... its not 100% reliable ( and you definitely need to do several runs, like 5-10 ideally) but its not "malware" and it can definitely show you if there are certain issues with your system... but it cant detect *all* issues for sure, and its therfore also not 100% reliable as said.

 

its still the best thing in that regard and the closest to what you're asking for (and again it definitely "can" help)

 

https://www.userbenchmark.com/Software

 

as said, let it run a couple of times and link the results here.

 

 

eh.. even if we ignore their 'editorial freedom' tending towards a specific brand more often than not, their methodology to determine performance is flawed at best if taken in a sample size of one. for example, it runs far too short to make potential thermal issues show up.

1 minute ago, Mark Kaine said:

yeah, but honestly something is off with that... like my infamous Kingston A400 drive... smart data says "good" no apparent errors etc, yet the speed is down to like 2.5mb/s max (doesn't matter if the drive is full or empty, new fw, old fw etc...)

 

And you know how i actually figured out what's slowing down my whole system (the Kingston "A400")? Yeap, userbench, told me right on the head the drive is faulty (probably simply by checking actual read/write speeds instead of "smart") 

 

Removed the drive and my pc immediately because "responsive" again. 

if they say your SSD is faulty despite *ALL* smart data being good.. that's one HELL of a stab in the dark. 

[Mark Kaine]

5 minutes ago, manikyath said:

if they say your SSD is faulty despite *ALL* smart data being good.. that's one HELL of a stab in the dark. 

but they were right...?

 

7 minutes ago, Mark Kaine said:

yet the speed is down to like 2.5mb/s max

you know how slow that actually is... lol

 

7 minutes ago, Mark Kaine said:

Removed the drive and my pc immediately became "responsive" again. 

(i still have the drive, just not anywhere near my pc, and yes its still slow as hell , tested on my laptop)  ¯\_(ツ)_/¯ 

 

 

5 minutes ago, manikyath said:

if we ignore their 'editorial freedom

which, yes we should... the results definitely need interpretation,  especially in regards to performance... but it can still give you hints (like with my A400)or lets say the gpu /cpu is close to the bottom,  basically everything thats red could give you a hint that there might be something wrong,  which is the whole point... its literally better than nothing especially for inexperienced users, who, admittedly will have issues with the correct interpretation,  but that shouldn't be an issue for more experienced users, who can use these results to help...

 

 

Thats also why i said its not 100% reliable... it might not show anything of substance,  then it obviously cant help, but in other cases it can indeed show faulty / sus parts (ie again better then nothing or just guessing in many cases)

[manikyath]

Just now, Mark Kaine said:

but they were right...?

were they? was it a hardware issue *in* the SSD? did you try to reformat it? will they be right every time?

 

even a broken clock is right twice a day, i'm not gonna delve into it deeper to not derail the topic, but i do not want to take what appears as a lucky shot in the dark, to determine that a benchmarking tool that was never an accurate thing in the first place would suddenly be a reliable troubleshooting tool.

[Mark Kaine]

12 minutes ago, manikyath said:

even a broken clock is right twice a day, i'm not gonna delve into it deeper to not derail the topic, 

its actually the topic,  but you might better not, its clear that you have ideological issues with userbench and not really the test itself... its a good "troubleshooting tool" actually, thats all it is though,  you definitely  have to ignore  the "editorial freedom" as you called it, its just not 100% reliable,  and needs some interpretation,  which i said from the getgo, its still *far better* than simply guessing in the dark in many cases.

 

i agree the test is actually too short, but that still fits in the "better than nothing" approach...

 

 

[manikyath]

1 minute ago, Mark Kaine said:

 

i agree the test is actually too short, but that still fits in the "better than nothing" approach...

yes, absolutely, i agree.

 

but for everything it does, there is an alternative that is better than userbench. *that* is my problem.

- disk: crystaldiskinfo reads out the smart data, crystaldiskmark gives you a clear image of the speed of the drive in different circumstances (despite what you seem to claim.. 2.5MB/s or Mb/s depending on which one you meant.. is still faster than a normally operating mechanical drive in some circumstances)

- cpu: prime95 will bring up just about any instability you may have, cinebench will do a very realistic load to do longer term thermal load tests.

- memory: 7-zip's benchmark will make memory errors show up faster than memtest in some cases. besides that it's also a very realistic CPU load that can run overnight.

- gpu: the unigine benchmarks, and well.. hundreds of other options.

- other sensors: hwinfo64

 

oh.. and if you want something that'll do all tests at once, and poop out a report... AIDA64 trial. i'm pretty sure the trial will still poop out a report.

[Mark Kaine]

35 minutes ago, manikyath said:

yes, absolutely, i agree.

 

but for everything it does, there is an alternative that is better than userbench. *that* is my problem.

- disk: crystaldiskinfo reads out the smart data, crystaldiskmark gives you a clear image of the speed of the drive in different circumstances (despite what you seem to claim.. 2.5MB/s or Mb/s depending on which one you meant.. is still faster than a normally operating mechanical drive in some circumstances)

- cpu: prime95 will bring up just about any instability you may have, cinebench will do a very realistic load to do longer term thermal load tests.

- memory: 7-zip's benchmark will make memory errors show up faster than memtest in some cases. besides that it's also a very realistic CPU load that can run overnight.

- gpu: the unigine benchmarks, and well.. hundreds of other options.

- other sensors: hwinfo64

 

oh.. and if you want something that'll do all tests at once, and poop out a report... AIDA64 trial. i'm pretty sure the trial will still poop out a report.

but you have to see it from an inexperienced users perspective too... all these tests you mentioned are even more difficult to run (ie have them in the first place) and  interpret... ub is really just a "first look" kinda thing and more or less easy to understand... if you're lucky it tells you something (like with my A400),  if not then the journey continues lol...

 

btw... i have Aida64 non trial, its near useless for me, it can show me my ram latency (yay?) and i can use it as plugin for afterburner,  but that's literally it how that program is useful to me (just as an example that these programs often really aren't easy to use and interpret) 

 

ps: theres a report that tells you if something is wrong?  (didn't know that, might be useful)

 

 

35 minutes ago, manikyath said:

2.5MB/s or Mb/s

eh, i don't actually know,  but the point is my other drives are around 500mb/s, even my conventional hdds are around 130mb (or better), so the Kingston definitely *is* that slow lol... is that like a floppy drive speed or even slower?

 

^this is with crystaldiskmark (since i had to fact check ub, obviously,  and yeap....)

 

edit: apparently it's MB/s 

[manikyath]

1 minute ago, Mark Kaine said:

 

btw... i have Aida64 non trial, its near useless for me, it can show me my ram latency (yay?) and i can use it as plugin for afterburner,  but that's literally it that program is useful to me (just as an example that these programs often really aren't easy to use and interpret) 

 

ps: theres a report that tells you if something is wrong?  (didn't know that, might be useful)

it has a BUNCH of builtin benchmarks, hoovers up a lot of system info, etc.

 

it doesnt do a "this component is broken" thing because that's always a stab in the dark if not going straight off SMART data. but if something is broken to the point it's causing OP's behavior, it's usually obvious. case in point.. i have a broken SSD laying around that'll just dunk out when you try to do a sequential benchmark on it. in fact.. i have two of those.

 

you're still going on about your *one* time it just happened to point towards the problem for you, but if you go back to the time userbenchmark was actually popular, there were PLENTY of threads where userbench poops out that your hardware is underperforming.. but it's just a huge error margin that they blindly stuck some tresholds on.

 

if it were as easy as a piece of software telling you "this component is broken", there wouldnt be a need for experienced techs in repair shops anymore. And up until a year ago i had the pleasure of working in the sort of place where people come after a repair shop that blindly shoots the parts cannon didnt solve their thing.. so i've seen plenty of evidence this isnt your solution.

[Mark Kaine]

55 minutes ago, manikyath said:

there were PLENTY of threads where userbench poops out that your hardware is underperforming..

it still does that... that's where the interpretation thing comes in... like for example is it "underperforming" at 96% (it really does say that sometimes lol) or is it "underperforming" at 11%?  <-- there's definitely a difference in severity here...

 

55 minutes ago, manikyath said:

if it were as easy as a piece of software telling you "this component is broken", there wouldnt be a need for experienced techs

yes, true, there isnt a software that can do that reliably, but the ub thing is the closest thing to do that, as i said you really have to forget about it as a yes/no/good/bad* thing primarily,  more like "where is something that doesn't perform as expected" and then you still need to interpret it...

 

for example it kept telling me my ram latency is bad , yes, but only by ~10 ns... its really not something I care much about,  but it was still shown in *red* hence i needed to look it up, compare with other data, etc.

 

 

*the main reason they do things as they do is to sell you new parts through their affiliate links, hence you really need to see through that and concentrate on the actual results and if they could really be an issue and not *their* interpretation thereof (so in no way its a perfect test or anything,  it can just give some hints really) 

 

 

55 minutes ago, manikyath said:

And up until a year ago i had the pleasure of working in the sort of place where people come after a repair shop

i did kinda too actually,  for 3 months  as a "volunteer" it was actually fun and i learned a lot -- but that actually funny enough didn't happen often... i just recall 1 case now where someone had a "prebuilt" with a way too weak power supply as an example where another shop messed it up...

 

mostly was: 

- it doesn't work

- its slow

- i think i may have a gazillion viruses installed "by accident"

- i forgot my password  / my ex changed my password  (lol)

 

also lots of half dying hdds and power supplies,  sometimes mobos...

 

ps: my absolute favorite:  "my wifi doesn't work and i did not turn it off!"

- every single time they turned off the wifi, sometimes even with a physical button (on older laptops)

 

Also sidenote, there were a lot of long-term returning customers,  you just knew its a good repair shop, and it was cheap as well, and also we had 1 guy (unofficial boss) that really knew 99% what's wrong immediately even with the lousiest of error descriptions... 

and yes, we definitely had some customers who came back the next day just to say that its really fixed!  : p (like dead mobos etc)

 

oh, and thats also where i built my first "gaming pc" ... no instructions and it worked perfectly... the only thing i didn't use the standoffs (because i never seen that before,  definitely looked like junk to me lol) but *that* the boss told me right away,  i better install those. : D

 

^ the standoffs also were clear plastic... ive never seen anything like that that again lol

 

[ClaustrophobicGames]

I figured it out! I was unfamiliar with the MSI Bios and didn't realize that the ai overclocking and XMP profiles were turned off.