[SOLVED] Help. Is my new ADATA XPG SX8200 Pro not the original version?

[QwerkyPengwen]

So it's simple, I purchased the ADATA XPG SX8200 Pro on Black Friday from Best Buy website, because it was discounted to $100.
Furthermore, it was rated in an SSD tier list as a Tier B SSD (high end) and had great review from Toms Hardware.

I got it in the mail yesterday and have installed it.

However, I have just now learned that this is one of the drives (along with others) that some SSD companies have been doing a bait and switch on, where they basically change out everything from controller to NAND, etc.

If anyone who has an original drive, has benchmark data, knows a bit about this drive, etc. can help me with looking into whether or not I got ripped off with a poor performing drive or if I ended up getting one that was from older stock with original hardware inside that would be great thank you.

I need to know what to start looking at, what tools I should be using to run diagnostics and tests, etc.

 

[jojesa]

What performance are you getting out it?
I have the non pro version

CrystalDiskMark
Real Word Performance selected
I could get higher score, if Peak Performance is selected

 

[fzabkar]

Compare your own benchmarks against Adata's published specs:

https://www.adata.com/upload/downloadfile/Datasheet_XPG SX8200 Pro_EN_20181017.pdf
https://www.adata.com/upload/downloadfile/Datasheet_XPG SX8200_EN_20180316.pdf

 

[QwerkyPengwen]

Compare your own benchmarks against Adata's published specs:

https://www.adata.com/upload/downloadfile/Datasheet_XPG SX8200 Pro_EN_20181017.pdf
https://www.adata.com/upload/downloadfile/Datasheet_XPG SX8200_EN_20180316.pdf

not sure how comparing my own to the data sheets that ADATA provides is going to help me.
You almost never get the same results as what they try to market at you. ever.

 

[QwerkyPengwen]

I just did an 8GB test for all of you here on the forums using CDM.

If there are other tools I should use (that are free) or different test sizes and test run amounts I should use let me know, thanks.

 

[OrlyP]

not sure how comparing my own to the data sheets that ADATA provides is going to help me.
You almost never get the same results as what they try to market at you. ever.

I think the gist was, if your results pretty much match up with the figures provided by the manufacturer (within a small margin of error), you should be ok.... regardless if parts were replaced.

But if you're getting consistently poor performance as detailed in the expose article, then your fears are correct. Though, it's not to say that what you have is a fake. It's likely still a real ADATA NVMe, but using lower-tier pars.

 

[QwerkyPengwen]

just now discovered the NVME option in CDM settings.
Switched to that and here are results.
Looks like the only thing it changed was the second and third test options

 

[fzabkar]

ISTM that your results exceed Adata's published spec.

 

[QwerkyPengwen]

ISTM that your results exceed Adata's published spec.

their published specs say 3500/3000 and I'm getting average 3350/2600
Read data being within margin as far as I can guess from their "peak" performance specs, but writes being quite low by more than a small margin.

The Random 4K after doing some math is as follows:
ADATA published specs (I'm assuming their results are for Q32 btw)
390k/380k
My results:
315k/330k

My reads seem to be fine while my writes are lack luster.
And my R4K reads are quite low by comparison to published specs, while my R4K writes are a bit better but still quite low by comparison still.

So I wouldn't say I'm exceeding their published specs at all.

Perhaps you were comparing my results to either that of the non pro version, or a lower capacity drive.
I have 1TB Pro version.

 

[Maxxify]

There are many factors that can influence performance. For example, using a chipset/PCH M.2 socket on the X570 platform with a SMI drive will have much worse results.

 

[jojesa]

There are many factors that can influence performance. For example, using a chipset/PCH M.2 socket on the X570 platform with a SMI drive will have much worse results.

Adding to that, some apps (e.g. Antivirus) and background process could cause lower scores.
My score goes up by 210, just by disabling security apps while running CDM.

 

[Prince_Porter]

I ordered 2 of these drives on Black Friday, from Amazon. I just recently got my PC up and running (was waiting on the 5900x, and had no CPU!), and just tested both of my drives. It seems VERY clear to me that my C drive is faster than my G drive. I'm using a B550 MPG Gaming Edge motherboard, with the C drive in the Gen 4 m.2 slot (not that it should matter), and the G drive in the other gen 3 M.2 slot. I ran the test quite a few times, at 1GB, 4GB, and 8GB, "normal" and "nvme" modes, everything else default, and the results are consistently skewed in that my C drive has significantly faster write speeds. I'm definitely happy my C drive appears to be the faster one.

Is there any way to see the controller name, or any deeper info that can confirm if the drives are indeed V2 and V3? Any other thoughts?

EDIT - I ended up using AIDA64 to check the controller, and they are BOTH ending in ENG, which is the good one. I read on Reddit that AIDA64 was reporting ENG, despite the chip saying otherwise on it, so I checked, and the "bad" chip is indeed ENG. I didn't check the good one, because it's under my GPU, and it's performing great. That said, what on earth is causing the "bad" drive to perform 20% slower in writes? Is there ANY logic to the 5900x powered M.2 slot being more potent than the B550 Gaming Edge M.2 down below? I really just want to solve the puzzle of it at this point, so curious.

 

[Maxxify]

SM2262G is the "bad" one, SM2262EN/ENG is good. Although the former comes with multiple flash variants with different performance characteristics.

Drives will always perform worse over chipset/PCH. SMI-based drives specifically have issues with Zen and Zen platforms. I documented this a long time ago on Reddit, contacted SMI and AMD, etc, but they never got back to me. Specifically it tanks in sequentials.

 

[QwerkyPengwen]

in case none of you could read my signature, I have a 6700K on a Z170 mobo.
suffice it to say, I am just going to say whatever and stick with the drive as the performance is better than what I would get from a cheaper $80 drive that uses TLC and does work around magic to not have dedicated cache on board.

This turned into some kind of discussion about said drives on AMD platform which I don't have, and has derailed from being about helping me answer my question which still hasn't actually been answered. but thanks for your time anyways folks.

 

[jojesa]

in case none of you could read my signature, I have a 6700K on a Z170 mobo.
suffice it to say, I am just going to say whatever and stick with the drive as the performance is better than what I would get from a cheaper $80 drive that uses TLC and does work around magic to not have dedicated cache on board.

This turned into some kind of discussion about said drives on AMD platform which I don't have, and has derailed from being about helping me answer my question which still hasn't actually been answered. but thanks for your time anyways folks.

That's the beauty of forums, we might give you answers that might not have anything to do with your issue.
You might need to do a bit of leg work to sort the useful posts.

The lower than expected performance of your SSD could be due to the fact that some manufactures release different versions of SSDs with the same model number but with diferente components.
To start they release a SSD version with top notch controller, cash and flash memory, and get rave online reviews.
People sees the reviews and purchase the item...
Then, after the first batch, manufactures, keep the SSDs with the same model number but with lesser components. Most users will never know. manufacturers spend less and could charge even higher prices depending on supply and demand.
This is nothing new...if you check back, 12 years ago, they use to do the same with flash drives...
They got away with that then, and of course they are getting away with it now.

Check this resent article about it.
https://www.tomshardware.com/news/adata-and-other-ssd-makers-swapping-parts

 

[Maxxify]

in case none of you could read my signature, I have a 6700K on a Z170 mobo.
suffice it to say, I am just going to say whatever and stick with the drive as the performance is better than what I would get from a cheaper $80 drive that uses TLC and does work around magic to not have dedicated cache on board.

This turned into some kind of discussion about said drives on AMD platform which I don't have, and has derailed from being about helping me answer my question which still hasn't actually been answered. but thanks for your time anyways folks.

Well, I answered one question first, then someone else interjected and I answered them. However, in both cases, it's true that going over chipset/PCH can have worse performance - Intel's consumer boards have all M.2 over chipset. However to address you original question - if you're looking for tools to ID the hardware you'll have to check my Reddit post.