News Adata Switches to Slower Controller, NAND on SX 8200 Pro

[Co BIY]

Has this article gotten a softer rewrite ? I thought the original had a very justified "bait and switch" subtitle.

Providing reviewers with one set of hardware and then selling units with lower specified hardware is worse than just sending known good review units that went though extra QC.

 

[seanwebster]

This tool says Controller : SM2262 [SM2262AB], do you know what AB stands for?

The SM2262EN usually reads as something like "Controller : SM2262EN [SM2262BA]" or "Controller : SM2262EN [SM2262AA]." Your results point to your model having the SM2262/SM2262G.

 

[Oshyan]

Anyone know if they've done anything to the Gammix S50 Lite as well? 😬

 

[Hlsgsz]

I would really like to see a comparative performance analysis, especially in the 4k Q1 metric, which this SSD excelled in. To be honest, while this practice is definitely not cool, if the new drives approximately retain their performance in that crucial metric, they are still a great deal. The EX950 is a good replacement, but, the way the market works, it's good to have two to choose from when you need an SSD for a build.

 

[VonLee]

Hi, first time here in this forum.

I just bought this drive 2 days ago, and while searching for NVME drive, i stumbled upon this article.

my drive is ADATA SX8200PNP 512 GB. While using the smi nvme tool, the controller listed as SM2262AB . But when i use AIDA64 Extreme, the controller listed as SM2262ENG.

So which one is correct ?

Because as seanwebster said, SM2262AB = SM2262G

 

[Gussv]

Yeah... it should read sth like:

Controller : SM2262EN [SM2262BA]

 

[seanwebster]

Hi, first time here in this forum.

I just bought this drive 2 days ago, and while searching for NVME drive, i stumbled upon this article.

my drive is ADATA SX8200PNP 512 GB. While using the smi nvme tool, the controller listed as SM2262AB . But when i use AIDA64 Extreme, the controller listed as SM2262ENG.

So which one is correct ?

Because as seanwebster said, SM2262AB = SM2262G

The controller has what model it is etched/laser engraved into it. What does the controller say on it physically? You don’t need software to determine that.

I really don’t use Aida all that much, but suggested it because some others had better luck with it. I guess it isn’t always correct. That other tool says the controller is the vanilla SM2262, not the EN like Aida reported.

 

[techscrub20]

and i was about to buy one too

 

[razholio]

If the issue were just a modest decrease in performance, I wouldn't have such an issue with it, but I saw a staggering drop in random I/O performance as well as some strange behavior. After happily using 2x 1TB sx8200pro's with the 'G' controller in a RAID1, I started to notice long read times on a large dir. Testing confirmed that IOPs on 4k random fio test had gone from around 130k/50k r/w, to averages as low as 20k/5k with some spikes approaching normal. Yes, trim was enabled and I manually trimmed prior to testing. The odd behavior is that while it was occasionally spike up to full performance, the lows were so incredibly low as to bring the average way down. I'm not sure exactly how much data had been written but it was at least a few TB, so it's something that may not show up until it's seen some use. I can't help but wonder if the firmware was doing a bad job at defrag.
I called amazon and told them the problem, and they were able to work out a return for full purchase price even though I'd had them for 6mos.
I replaced them with 2x 1TB SKHynix gold p31's that perform better when new, and hopefully will hold that performance for longer.

 

[javiindo]

If the issue were just a modest decrease in performance, I wouldn't have such an issue with it, but I saw a staggering drop in random I/O performance as well as some strange behavior. After happily using 2x 1TB sx8200pro's with the 'G' controller in a RAID1, I started to notice long read times on a large dir. Testing confirmed that IOPs on 4k random fio test had gone from around 130k/50k r/w, to averages as low as 20k/5k with some spikes approaching normal. Yes, trim was enabled and I manually trimmed prior to testing. The odd behavior is that while it was occasionally spike up to full performance, the lows were so incredibly low as to bring the average way down. I'm not sure exactly how much data had been written but it was at least a few TB, so it's something that may not show up until it's seen some use. I can't help but wonder if the firmware was doing a bad job at defrag.
I called amazon and told them the problem, and they were able to work out a return for full purchase price even though I'd had them for 6mos.
I replaced them with 2x 1TB SKHynix gold p31's that perform better when new, and hopefully will hold that performance for longer.

No way you can notice the difference without benchmark tools.

 

[TMehanna]

AIDA64 Extreme does the job, indeed.

Checked my drives yesterday and the results coincide with the dates purchased as well as the money they costed me!

Aida does not actually read the drive information, it displays what it has in its own database for a particular model. It will display the same information for any XPG8200 Pro regardless of the actual hardware.

 

[Pleiades]

I guess a lot of people do it albeit in different ways of thier own. GTX 970 = 3.5 GB. WD and Seagate, crappy slower models included sliently among faster/better drives. AMD lawsuite on number of cores etc etc. The list goes on. Unfortunately for people who dont check tech sites or benchmark drives, they never know what happened.

Personally wouldn't include the old AMD Bulldozer thing in with this, it's not really comparable IMHO. Think the plaintiffs only got away with it in California, IIRC? Wouldn't wash in the rest of the world AFAIK (don't think it did in the EU anyhow)...

This OTOH is much more egregious and underhand...

Ha, just realised my over-use of txt spk, sorry!

Edit: Oops, bit necro too, sheesh.