Samsung Is the Latest SSD Manufacturer Caught Cheating Its Customers (extremetech.com) 53
Crucial and Western Digital have recently been caught swapping the TLC NAND used for certain products with inferior QLC NAND without updating product SKUs or informing reviewers of this change. Now, Samsung was caught doing something similar. Samsung is "swapping the drive controller + TLC for a different, inferior drive controller," according to ExtremeTech. "The net effect is still a steep performance decline in certain tests." From the report: The other beats of this story are familiar. Computerbase.de reports on a YouTube Channel, which compared two different versions of the Samsung 970 Plus. Both drives are labeled with the same sticker declaring them to be a 970EVO Plus, but the part numbers are different. One drive is labeled the MZVLB1T0HBLR (older, good) and one is the MZVL21T0HBLU (newer, inferior). Peel the sticker back, and the chips underneath are rather different. The Phoenix drive (top) is older than the Elpis drive on the bottom. Computerbase claims a production date of April 2021 for the Phoenix, but if the 2110 and 2123 codes are production dates, this would seem to indicate March and June. It's possible that Samsung uses specific numerical codes for each month. Either way, the Phoenix drive is older and faster and the Elpis drive is newer and slower.
And -- just as we've seen from Crucial and Western Digital -- performance in some benchmarks after the swap is just fine, while other benchmarks crater. [...] The original 970 Plus starts with solid performance and holds it for the entire 200GB test. The right-hand SSD is even faster than the OG 970 Plus until we hit the 120GB mark, at which point performance drops to 50 percent of what it was. Real-world file copies also bear this out, with one drive holding 1.58GB/s and one at 830MB/s. TLC hasn't been swapped for QLC, but the 50 percent performance hit in some tests is as bad as what we see when it has been.
And -- just as we've seen from Crucial and Western Digital -- performance in some benchmarks after the swap is just fine, while other benchmarks crater. [...] The original 970 Plus starts with solid performance and holds it for the entire 200GB test. The right-hand SSD is even faster than the OG 970 Plus until we hit the 120GB mark, at which point performance drops to 50 percent of what it was. Real-world file copies also bear this out, with one drive holding 1.58GB/s and one at 830MB/s. TLC hasn't been swapped for QLC, but the 50 percent performance hit in some tests is as bad as what we see when it has been.
Just the latest cases (Score:5, Interesting)
Re: (Score:2)
I suspect what's new here is that extremetech.com only recently discovered Slashdot - all of these stories are coming from the same place.
Re: (Score:2)
I suspect what's new here is that extremetech.com only recently discovered Slashdot - all of these stories are coming from the same place.
I suspect extremetech.com use of Slashdot predates even your quite low UID. There have been 560 Slashdot stories posted from extremetech.com dating back to 2001 https://slashdot.org/story/01/... [slashdot.org]
Incidentally all of the comments have UIDs lower than 300000 so in the eyes of extremetech.com... you must be new here.
Re: (Score:2)
They changed the part number which is good because it's made out of different parts, but they didn't change the product name which is bad because "average Joe buyer" doesn't look at part numbers - which is exactly what these companies are betting on.
Re: (Score:3)
About three years ago, I have decided to upgrade my RAM. Had a kit of 2x 16 GB Corsair Vengeance LPX (3000 MHz) in my PC at the time, and I bought another identical (or so I assumed) kit. Same SKU, same code, same package, same everything. After plugging the new kit in the PC, next to the old kit, I could not make the RAM work in quad channel at the advertised speed. Crashes upon boot, failures to boot, etc.
Troubleshooting revealed the memory dies were from different manufacturers across kits. One kit had H
Re: (Score:2)
Or incompatible parts...
Manufacturers will frequently change the chipset used on things like ethernet or wifi cards, where the changed chipsets not only have varying performance but require entirely different drivers (which may be buggy or not available for some platforms) and different features (master mode, monitor mode etc).
Re: Just the latest cases (Score:1)
Re: (Score:2)
It's a ASUS X99 A II, which does indeed support quad channel. Still using it, by the way.
Re: (Score:2)
This is why the distance selling regulations in the UK are such a wonderful thing. Basically if you buy something remotely, e.g. online, you have 14 days from when you receive it to return it, no questions asked. You don't even have to keep the packaging, just return the item undamaged.
Of course if it's not as advertised then the seller pays the return shipping. I'm sure Samsung were careful to add an asterisk to their performance figures though, but because of distance selling regs the worse you are out is
Re: (Score:3)
I guess the idea is that you replace all the bulbs in a chandelier when a single one goes. More sales that way.
Re: (Score:2)
I've done this many times. Sad but true.
So What We're Hearing... (Score:2)
...is that we're in the final decline of Late Stage Capitalism, [theatlantic.com] preceeding a major crash in both the economy and society, and the major corporations are shovelling their crap out the door to make their last few pennies before scurrying to the bunkers.
Re:So What We're Hearing... (Score:4, Informative)
...is that we're in the final decline of Late Stage Capitalism, [theatlantic.com]
There is nothing new about misrepresented products. The first instance was a serpent with an apple in the Garden of Eden, ... and God blamed the customer.
Re: (Score:2)
Re: (Score:2)
Without delving into fiction or myth, you can point back to at least 1750 BC with the Code of Hammurabi [wikipedia.org], which dealt with many aspects of life, including commercial law and fraud. And it's not the earliest known text, just one of the more famous. Capitalism has always had to be paired with proper regulations in order for it to work well. Human nature doesn't change all that much, and we still require a stick to keep some people from the temptation of defrauding their customers.
Re: (Score:2)
Rule of Acquisition #239: Never be afraid to mislabel a product.
Re: (Score:2, Funny)
--
www.fark.com/politics
Re: So What We're Hearing... (Score:3)
This is hilariously close to what they normally post too.
Re: So What We're Hearing... (Score:2)
Clearly this is a commie slam humor peice, but look at the awsome games to come from behind the iron curtain!
https://youtube.com/watch?v=Yd... [youtube.com]
People in the US would've been emptying their bank accounts just to play thrilling games such as "Merkspiel"!
Re: So What We're Hearing... (Score:1)
Re: (Score:3)
wE'Ve NEVEr hAd rEaL capItaliSm.
Re: (Score:1)
"Late Stage Capitalism" is the leftist's version of end-times. It'll all collapse Real Soon Now which means the glorious revolution can occur. Until then, the definition will continue to shift to mean whatever intolerable conditions exist at the time.
Re: (Score:3)
Just let me know (Score:2)
Re:Just let me know (Score:5, Insightful)
Will we all get 75 cents after a lengthy class action suit
Class action lawsuits rarely make the customers happy, but they do serve as a deterrent. A business is less likely to cheat or cut corners if they fear an expensive lawsuit.
Basically, we outsource consumer protection to the private sector.
Re:Just let me know (Score:4, Interesting)
Re: (Score:3)
Re: (Score:2)
First we have to turn this into a proper scandal and add the -gate suffix to it. Like Dieselgate. Maybe NAND-gate?
I see what you did there. You should be ashamed of yourself...
Re: (Score:2)
Re: (Score:1)
I only skimmed the article, but does the new drive *not* match the specs? The specifications that Samsung publish, I mean - not what some dipshit "tech journalist" posts on their blog.
When Intel is the honest company... (Score:5, Interesting)
Maybe it comes out that Intel is pulling the same nonsense but now we have Micron/Crucial, WD/SanDisk and now Samsung which as I believe it are 3 out of probably the top 5 largest manufacturers of SSD drives. This is starting to get on the level of the RAM price fixing scheme.
I think Intel is certainly capable of and has engaged in some fairly shitty behavior but I don't think(hope) they would lower themselves in such an obvious bait and switch scheme like this.
I expect very little honestly out of any of these corps but even I am a little surprised these companies would engage in such a petty greedy scheme. They either have very little credence in their own brand status or likely very little faith in the memory and discretion of their consumers.
Indeed the bar is close to the floor at this point.
warranty (Score:3)
Re: (Score:2)
Re: (Score:2)
If it is like most warranties it will be replaced with refurbished shit.
That's only on an integrated level device level such as a laptop. On a component level almost nothing is refurbished. Your SSD ends up in an e-waste pile melted down by some 10 year old who will die of cancer before he's 30, and you'll get a factory new.
Re: (Score:2)
Complain all you like, nothing will ever be as sketchy as this thing. [wikipedia.org]
Nope (Score:5, Informative)
This is being heavily mis-reported by numerous outlets, even by ars.
In smaller writes that you experience under normal usage, there is no performance impact. And for large writes, the first about 115GB is faster than before, not slower. And then after that it is slower.
They switched to the controller from their 980 Pro model. It just has different tradeoffs. Even for most write-heavy loads, this new controller is faster, not slower. But it is true that when you're doing benchmarks, and set the write set size to be really huge, it benchmarks slower.
Re: (Score:2)
Nobody's answering the real question that everyone is asking: will they finally release a 40-pin IDE version of it?
Re: (Score:2)
will they finally release a 40-pin IDE version
Or at least an adapter to connect to the interface that matters: ST-412.
Re: (Score:3)
People choose different products based on those tradeoffs. Someone may have specifically bought a 970 because it has higher sustained write speeds for large files, which is a perfectly valid use case. If they wanted the higher write speed for smaller files, they could have gone with the 980 to start with.
Products using different parts should be sold under different model numbers because they are fundamentally DIFFERENT PRODUCTS.
Re: (Score:1)
That's not a "use case," that's a abstract benchmark type of thing. What is the use case that encounters "sustained write speeds for large files" above 120GB?
Normal use cases that are write-heavy aren't copying files of that size. For example for data collection, you'd fill the disk really fast if you were writing that fast. Most write-heavy workloads, like video recording, do not write large enough data blocks to fill that cache.
Products using different parts should be sold under different model numbers because they are fundamentally DIFFERENT PRODUCTS.
In the current environment of shortages, that would destroy supply chains. Ins
Re: (Score:2)
> But it is true that when you're doing benchmarks, and set the write set size to be really huge, it benchmarks slower.
TFS has specific claims about 50% slower file copy speeds, outside of benchmarks. Are you saying that's a feature of the 980 Pro controller or that the claim is false?
Re: (Score:2)
Right. They made that claim. On a huge file copy. But their benchmarks show it was faster, not slower during the first 115GB of the write operation.
The broad claim, without caveat, that it is slower isn't just false, it is a blatant lie.
Re: (Score:2)
This is being heavily mis-reported by numerous outlets, even by ars.
In smaller writes that you experience under normal usage, there is no performance impact. And for large writes, the first about 115GB is faster than before, not slower. And then after that it is slower.
They switched to the controller from their 980 Pro model. It just has different tradeoffs. Even for most write-heavy loads, this new controller is faster, not slower. But it is true that when you're doing benchmarks, and set the write set size to be really huge, it benchmarks slower.
When I was selecting SSDs last year, my primary criteria was write endurance and QLC drives have about 1/3rd that of TLC drives. I would be pissed off if the drive I bought to last 3 years only lasted 1 because the specifications lied.
Now people are wishing they had TLC? (Score:2)