Adata SX8200 pro , running slow

[CoSmA]

Hello everyone,

So, recently because of my Masters degree final , I needed a nvme drive to be able to pass-trough to a vm.
The one I choose after some research was a Adata sx8200 pro 512gb , brand new .
First thing i did when I got it was to mount it in the bottom m2 slot of my x570 tuf , at the time reaching the top m2 meant removing the first gpu, which meant removing my d15.....and the bottom slot had the nice heatsink .

All fine and dandy, booted windows started crystal diskmark and bam ....reads of 3000 ....not bad ..but the package says 3500 .
Tried as ssd 2999 something, ok this looks ok as other reviews had the same value there.

Getting annoyed, removed the cpu fan, the gpu , applied the stock heatsink that came with the ssd and mounted on the top m2 slot .
Woala , 3500 on crystaldisk .

In the meantime the project is done and i can turn back to gaming , so i upgrade the bios , configure my ram again and .....quck ..2700 -3010 read on crystaldisk ?

Ok , maybe the bios is bad, revert to the previous one, same, cmos reset , battery out of the mobo, same .

I'm kinda annoyed by this, the ssd is only 6% filled, it's les than a month old and it only had a ubunu install on it which later turned in to a windows install and thats it , no intensive read and write on the ssd.


What can i do? or should i just let it be ?

EDIT: just as I ran the benches again to screenshot, the speed of crystaldisk got back to normal .


Is my mobo acting up or is the ssd faking the speed, I'm very new to nvme stuff so any info is greatly appreciated.

https://imgur.com/a/xqYkc79

 

[EarthDog]

Run atto... what does it say? Crystaldiskmark?

A quick search shows pretty similar results in as ssd, no?
https://www.legitreviews.com/adata-xpg-sx8200-pro-ssd-1tb-review_215091/5

 

[trents]

Package claims are theoretical and seldom realized in real life. Does it seem slow to you in actual usage?

Have you checked temps? Those things start throttling at about 70c. This can affect overall test scores in longer, more rigorous tests.

 

[EarthDog]

Package claims are theoretical and seldom realized in real life.

Not really theoretical so much as it is basing the results on the best-case scenario, typically (sequential R/W). This is why I suggested using ATTO. Some MFG use this to rate their drives...

 

[trents]

That's what I meant. Real life doesn't consist of best case scenarios.

 

[EarthDog]

That's what I meant. Real life doesn't consist of best case scenarios.

Ahh, I see now.




Anyway, I don't see anything wrong with the screenshot you (OP) posted as it matches up well (if a bit slower) than your results.

Run ATTO and you should see somewhere around 3500.

 

[CJustin]

The promises that are made by the manufacturers are not real.
Practically, the performance that is promised is not attained.
If it is attained then it deals a huge blow to the longevity of the card.
So keep away from such false claims.


I advise you run ATTO and see the actual performance.

 

[EarthDog]

The promises that are made by the manufacturers are not real.
Practically, the performance that is promised is not attained.
If it is attained then it deals a huge blow to the longevity of the card.
So keep away from such false claims.

1. They are real.
2. It is attained in specific situations.
3. When it is attained it is not in any way a 'huge blow' to the longevity of the card.

 

[GunnerMan]

ADATA provides numbers for the specific sizes and benchmarks (ATTO, Crystal, AS SSD) in their data sheet.

Notice that they use a queue depth of 32 for their CDM test. This is different from the default of 8 or 1 in CDM. Different QD's can significantly change the performance of an SSD. SSD's can often make IO optimizations and operate on data in parallel which allows them to produce faster speeds at certain QD's (at the expense of latency). There are a couple of good explanations to this on serverfault.