11 ms·
Silent changes to Western Digital’s budget SSD may lower speeds by up to 50%
- Const-me 5y agoWe only have 4 manufacturers of SSD drives who also making their own chips: Intel, Micron, Samsung, and SK Hynix. Here’s the table: https://en.wikipedia.org/wiki/List_of_solid-state_drive_manufacturers https://en.wikipedia.org/wiki/List_of_solid-state_drive_manu... I tend to buy Samsung and Intel, both working well in my devices.
- kogepathic 5y agoI thought Western Digital had vertically integrated their SSDs after the purchase of SanDisk? According to that Wikipedia list, Western Digital does manufacture their own flash, through a joint venture with Kioxia.
- andruby 5y agoAlso, Western Digital is looking at completely buying that Kioxia https://www.reuters.com/technology/western-digital-advanced-talks-merge-with-japans-kioxia-holdings-wsj-2021-08-25/ https://www.reuters.com/technology/western-digital-advanced-...
- justinclift 5y ago> I tend to buy Samsung and Intel, both working well in my devices. It's a generalisation, but be careful of using new Samsung consumer SATA drives (860, 870) if you're using Linux. The Linux kernel has workarounds in place for the older ones (840, 850 series). But the workarounds for the newer models are only starting to be looked at this week: https://bugzilla.kernel.org/show_bug.cgi?id=203475 https://bugzilla.kernel.org/show_bug.cgi?id=203475 Without those workarounds (eg disabling NCQ trim), there are data loss problems.
- Const-me 5y agoThanks for the warning. Fortunately, I’m unaffected because running Windows. I’m running Linux on top in WSL1 and VMWare. Also running Linux on ARM devices which have neither SATA nor PCIx. When I need fast disks in them to compile my C++ code, I simply use a network-mounted drive to build, gigabit LAN is good.
- wtallis 5y ago> disabling NCQ trim Fixed support for queued trim was part of the marketing claims for the 860 series. Whatever's going on with the 860 and newer drives requires deeper investigation, because there's plenty of reason to believe that simply applying the old workaround is not accurately targeting the real problem.
- justinclift 5y agoThe old workaround was to turn off support for NCQ (queued) trim, instead just using standard trim. It doesn't seem to have any real downsides. :)
- wtallis 5y ago> It doesn't seem to have any real downsides That depends greatly on what your software's policy for issuing Trim commands is. Partly because of the dearth of drives properly supporting queued trim, it is common for Linux distro to only run trim commands via cron job, rather than issuing them in real time during a workload.
- alias_neo 5y agoThanks, I had no idea about this. I've been using Samsung drives in all of my machines and servers and I run Linux exclusively. With a mixture of 850, 860 and 870's I must be at least partially affected.
- aidenn0 5y agoAre 860s new? They are what's in my system and while I don't remember when I upgraded it was definitely pre-COVID... [edit] smart says they have 631 days of power-on time, so maybe 2019? [edit] I seem to remember having issues a while back but I haven't seen any since I upgraded my firmware to RVT04B6Q
- scns 5y agoSadly my goto Crucial started doing it too, but they paired it with a huge price drop. Was the P2 NVME IIRC. ~70€ for 1TB.
- scns 5y agoJust read that Samsung changed the BOM on 970 evo plus. Switched to the controller of the 980 pro. Test reults are mixed. 4K QD1 is up ~50->82MB/s, serial writes after exhausting SLC cache (115GB) are nearly halved 1500->800MB/s.
- wtallis 5y agoThat's not what happens when you switch to a newer, faster controller. Those performance changes are what happens when you switch to newer NAND flash that's manufactured with twice the per-die capacity, so you have half as many dies for a given drive capacity. (Assuming the newer NAND isn't split into twice as many planes per die to compensate.) Edit: Looking into this further, TechPowerUp, Computerbase and others have mistaken K90UGY8J5B for K9DUGY8J5B. It's an easy mistake to make, and I've done exactly this before. But it completely explains the performance difference. The D that changed to a 0 signifies a switch from 16 dies per package down to 8 dies per package. The digits signifying the capacity of the package have stayed the same. The "B" at the end signifying the generation has also stayed the same, but Samsung didn't introduce 512Gbit TLC dies until a generation after they introduced 256Gbit TLC dies, so on the smaller dies the "B" means 92L and on the larger dies the "B" means 128L.
- kasabali 5y ago> That's not what happens when you switch to a newer, faster controller. Wasn't also caching strategy changed with the newer controller? Larger dynamic slc, slower folding speed?
- wtallis 5y agoI haven't done the math, but my gut feeling is that the relatively modest increase in SLC cache size could not on its own lead to that big a drop in post-cache sustained write speed except possibly where QLC is concerned, and especially not when mitigated by a generational increase in per-die performance. Samsung's drives are still operating with SLC caches that are smaller than the theoretical limit, which means that even after the cache is full, the drive should be able to accept some amount of writes that bypass the cache and have near native TLC performance; flushing the cache can be deferred a bit longer until the drive is actually starting to run out of free blocks, as opposed to typical QLC drives with maximally-sized SLC caches that run out of free blocks at the same point where the cache runs out.
- randerson 5y agoAnd Intel just sold their SSD business to SK Hynix, so soon there will only be 3!
- fomine3 5y agoThis issue happened on WD (acquired SanDisk). So the "just buy NAND manufacturer's SSD" strategy don't work. Also you miss Kioxia!
- userbinator 5y agoProducts that don't behave the same should really have different part numbers. In this example, perhaps they could call them SN550 and SN550A (B, C, etc.) or even SN551, SN552, ... if the majority of the design hasn't changed.
- mirker 5y agoYes, a great current example is consumer motherboards with Intel’s I225V network controller. The controller is unable to run at 2.5Gbps without a fix. The controller was fixed in revision 3, but a huge number of boards are shipping with revision 2. This is two years after the issue was identified, and there is no way to see which one you will get. Even boards manufactured this year have the old NIC, so I assume vendors are being shady to liquidate old NIC stock. So you get sold a motherboard with a 2.5X NIC slowdown compared to advertised (it gets downgraded to 1Gbps). Update the figure on the box if you’re going to do that!
- sennight 5y agoThe most recent example of a "soft recall" would be Gigabyte's exploding PSUs. It makes you wonder how much retail outlets knew, because it is clear that Gigabyte was fully aware of the issue - and silently fixed their design.
- xyzzy21 5y agoThis happens with products all the time. Strictly "shrink-flation" of consumer goods is the same thing. One industrial vendor we once partnered with took a product P that had speed S and accuracy A, discontinued that product and replaced with product P1 which was speed S/10 and accuracy A paired with product P2 which was speed S and accuracy A/10. The original P products were precious if you could find them.
- belorn 5y agoSilent change in this concept means there is no way to identify P2 from P, and both are sold under the same product name, product code, product specifications, and same product reviews. The later is the key aspect as selling products with the intent to trick the consumer is a defining aspect of fraud when done intentionally for profit purpose. While shrinkflation often also intend to trick the consumer, the details are usually defined. The packet weight or volume number is changed, the weight cost is change (a list requirement for many products where I live), the number of items has changed. With silent changes however there is no such information available and the consumer just will have to hope that their P is not P2.
- ta988 5y agoI'm super happy with my SN850 except those pesky leds that you cant turn off from Linux.
- MR4D 5y agoWhy is this not fraud? If you as a company source a device for reviewers and then change the specs in a way that you do not communicate, that is intent to defraud. If you bought a corvette because all the car reviews said it went 0-60 in 5 seconds, and then you buy one and it takes 10 seconds to do the same thing, you’d have a pretty good argument - especially if everyone else experienced your same situation. I’d be surprised if a class action suit or FTC investigation doesn’t happens within the next year.
- tyingq 5y agoThey seem emboldened by the event where they silently introduced SMR into NAS drives. That is, they didn't suffer any real consequences there, so now it's a free-for-all.
- alias_neo 5y agoThey must have lost a fair amount of sales? I previously would never have considered buying Seagate, but since the debacle, when it came to replacing a failed drive in a ZFS array of WD Red CMR drives, I couldn't trust I'd get a drive that would actually resilver so I just replaced them all with Seagate IronWolf NAS drives. The ops team at my work also avoids WD now, and I no longer buy WD SSDs or external drives because of this distrust.
- kaladin-jasnah 5y ago> They must have lost a fair amount of sales? (Anecdotally) Not from me. SMR or not, their easystore drives that you can shuck are cheap. Between buying a cheap drive and buying something else, I like to spend less money, so I'd choose the easystores. Granted, I don't know how much has changed with the easystores since the SMR debacle since I've not really been in the hard drive market for a while.
- gruez 5y agoAFAIK people who were pissed about the reds getting switched to SMR are disjoint from the people who shuck stuff. The latter group usually verifies the drives that they get, because you're not guaranteed to get a drive when shucking.
- cableshaft 5y agoThis happens to video games also. Publishers give review copies with microtransactions not included, the reviewers review it, and then shortly after release they add their free-to-play garbage so they get reviews that don't take the microtransactions into account and then they add it later. The also do it during ESRB submission to get around having to have a label that mentions in-game purchases. I'm surprised ESRB lets them get away with it, actually. Jimquisition on it: https://youtu.be/9LeZ89_u2Gc https://youtu.be/9LeZ89_u2Gc Article on it: https://www.tweaktown.com/news/66889/publishers-delaying-microtransactions-bypass-reviews/index.html https://www.tweaktown.com/news/66889/publishers-delaying-mic...
- some_random 5y agoESRB is industry controlled iirc, and microtransations/lootboxes are a huge cash flow for the largest games companies
- wiz21c 5y agoEt voilà. Conflict of interests all the way...
- plorkyeran 5y agoYeah, it's voluntary self-regulation that was establish to stave off laws being passed in the early 90s during the panic over Mortal Kombat, Doom, etc. Presumably if legislators start looking into microtransactions and rating tricks then the ESRB will start taking them much more seriously.
- pulse7 5y agoWhat are those microtransactions doing? Crypto mining?
- michaelt 5y agoThe reviewers get sent an account with everything enabled, and have a great time playing as Darth Vader Or Han Solo Or Whatever. The general public buy the game, and find if you want to play as Darth Vader or Han Solo unlocking them costs $20 each.
- aurizon 5y agoWanna buy a one legged duck? People who will want this are low end system builders who can add the "Has an SSD" who market to people have no idea how fast a 2 legged duck can run....
- zucker42 5y agoThis was a problem with the Adata XPG SX8200 Pro, which LTT covered in this video[1]. [1] https://youtu.be/K07sEM6y4Uc https://youtu.be/K07sEM6y4Uc
- post_break 5y agoShip the top tier hardware NVME drives to reviewers, wait a few months then start swapping out slower memory chips, controllers, go through multiple revisions while keeping the same name so when gamers look for reviews they see blinding speeds only to be duped. It's incredible what drive manufacturers are getting away with.
- HauntingPin 5y agoI've also seen this happen with laptops. They'll have a decent Samsung NVMe SSD when reviewers get their hands on the devices, but then after release, they switch it out for a cheaper one that performs much worse, particularly in longer workloads.
- p_j_w 5y ago>It's incredible what drive manufacturers are getting away with. I think it's more incredible that we continue to elect governments that let them get away with it.
- exporectomy 5y agoThis doesn't need more regulation. Just stop trusting professional reviewers who are not bound by the existing false advertising laws that the manufacturers/retails already are. If you want to know how something performs, read the specs published by the manufacturer. If they're too shy to say how slow it is, then there's a reason for that!
- olao99 5y agoYes it does need more regulation. It is fraud and as a manufacturer you shouldn't be allowed to get away with deceptive shit like this.
- exporectomy 5y agoIt might be deceptive, but only because consumers somehow believe product names correspond to unchanging products. They obviously don't, across many industries. Cars keep the same names but change design every few years. Food keeps changing package size and recipe. It's everywhere! They didn't promise not to change it, and the products still conform to the published specs. If you're too lazy to read that from the horse's mouth and rely on outdated 3rd party reviewers, that's on you.
- bitL 5y agoI bought a 1TB WD Blue M.2 SSD like the one on the picture and my computer has issues booting from it when all USB devices are plugged in (x299/10980XE). It looks like more people have issues booting from it as well, based on the feedback I saw on the seller site. How can one botch SSD that way?
- y04nn 5y agoUnbelievable, but is there a datasheet available and does the datasheet have read/write speed figures? Edit: [1] The datasheet shows "up to" speeds, which means that it is the disk cache speed and not the NAND speed? Quite disappointing. [1] https://documents.westerndigital.com/content/dam/doc-library/en_us/assets/public/western-digital/product/internal-drives/wd-blue-nvme-ssd/data-sheet-wd-blue-sn550-nvme-ssd-idk.pdf https://documents.westerndigital.com/content/dam/doc-library...
- deelowe 5y agoDon't expect useful specs to be available for consumer drives. The datasheets have slowly been turned into marketing materials over the past 10 years. Only the devices targeted for business use have useful specs and even then, key details such as MTBF or similar can be missing.
- 0-_-0 5y agoQuick summary: The drive is 1TB with a 12 GB SLC cache that's capable of 2160 MBps write speeds. When you write more than 12GB in a short time to the drive you can run out of this cache and speeds drop to 390 MBps. The dive was previously tested as capable of 610 MBps in this scenario. (Read speeds are fast in every scenario.)
- burnished 5y agoThis is kind of a hyperbolic re-write of the linked article at Tom's Hardware[1], which is I guess originally sourced (by both articles) at a Chinese site called expreview. The linked article at Tom's Hardware has more information and more context that is, confusingly, left out of the top level article. [1]: https://www.tomshardware.com/news/wd-blue-sn550-ssd-performance-cut-in-half-slc-runs-out https://www.tomshardware.com/news/wd-blue-sn550-ssd-performa... p.s is it possible to change what the link is to? it seems sort of pointless to have a discussion on an article that appears to be clickbait, in that it adds nothing substantive and in fact neglects to include a lot of information.
- aidenn0 5y agoNote that it's not just SSDs that can have completely different components on the same part. I've seen many USB wifi dongles with the same model number use completely different wifi chips internally.
- theropost 5y agoI think I'm done with WD from now on. The whole WD MyBooklive remote factory result fiasco annoyed me as well - for those impacted they offered a 40% discount on one of their new MyBookLive devices, as long as you send the old drive back (shipping paid by user).
- reviews-chat 5y agopeople trust reviews, that's why
- kstrauser 5y agoWD is on my personal blacklist for now. Don't forget last year when they lied and marketed SMR drives as NAS-compatible: https://www.tomshardware.com/news/wd-fesses-up-some-red-hdds-use-slow-smr-tech https://www.tomshardware.com/news/wd-fesses-up-some-red-hdds... I can forgive hardware failures. Packing a few trillion bits into a box and making them perfectly reliable is a hard job. What I absolutely cannot and will not forgive is Western Digital lying about their specifications. If I can't trust their words, damned if I'm gonna trust their products.
- andrewstuart2 5y agoWe ought to have the equivalent of crowdsourced semantic versioning and CVEs for products and reviews. As a consumer, I should be able to know _easily_ if any breaking changes have happened that could render a specific set of reviews or benchmarks obsolete, and just as easily if a specific issue is resolved in future revisions. Too often I've e.g. bought an Amazon product I've bought before, only to receive Revision 2 that doesn't fit my case any more, but that the company silently rewrote descriptions for just to keep their review score. I understand why people want to keep their review scores, etc, and to some extent they _should_ keep them, but not at the product level, when the product changes.
- cwizou 5y agoThis happens a lot with flash in general, mostly because those who manufacture the end user product don't usually make the NAND (or don't always use their own). I had the issue with "high end" USB sticks a while back during a review. The speeds given are completely irrelevant and always based on best case scenarios. And while a few manufacturers stick to the same formula for controller + NAND for the life of the product (say Sandisk), others just don't care and will silently switch both the controller and NAND with 0 indication on packaging (no rev., no nothing). On SSDs I can't remember seeing that though, but then again, this is Western Digital we're talking about. A company best known for, let's say, being a very late entrant to the SSD market. Edit : At least they (kinda) owned up to it after being caught, quoting the updated article : "For greater transparency going forward, if we make a change to an existing internal SSD, we commit to introducing a new model number whenever any related published specifications are impacted,"
- BearOso 5y agoAlso, a warning on Western Digital”s SN750 SE series. These are PCIe 4.0 protocol ones that claim to outperform the regular SN750 drives. In reality they’re slower because they’re DRAM-less. Last I checked they were the same price, so be careful to not get the SE model.
- LinAGKar 5y agoHow much of a difference does the DRAM cache make anyway, given that there is already a significantly larger disk cache in the system RAM, which is also closer to the CPU?
- fomine3 5y agoOn SSD market, silent change in same model and performance degrade is continuously happen. Any maker should change model number.
- reviews-chat 5y agoNot fair