User Posts: Jesper Berg
Now this isn't the most exciting SSD you can get your hands on for testing, and it certainly isn't among the fastest. What makes the Kingston A400 series ...
One of the somewhat confusing aspects of the SSD market is that 'newer' technology does not necessarily equal 'better'. This is most pronounced in the move ...
High-end SSDs have been saturating the aging SATA 6 Gbps interface for quite some time. Consequently, the only alternative for those who require faster storage ...
There are several alternatives available that will move your OS to a new location, e.g. from a mechanical hard drive to an SSD. Several manufacturers include ...
Samsung has just announced a new line of external SSDs with excellent transfer rates, but one of the most interesting aspects is the tiny form factor. The T1 ...
PC builders will recognize Lite-On as one of the most prominent manufacturers of affordable and capable optical drives. Now the company is apparently looking ...
Today's SSDs use NAND flash. In a few years the situation might be different. Western Digital subsidiary HGST is apparently making progress in the field of ...
Intel continues to expand its line of consumer SSDs with a SATA 6 Gb/s mSATA model. It's intended for thin and light laptops (Ultrabooks) and other systems ...
While the next SATA standard, SATA Express, is on the horizon, it is not ready for prime time just yet. Meanwhile, the SATA 6 GBps interface is largely ...

Samsung Acquires Nvelo

Samsung has just announced that it's acquiring California-based Nvelo, a company developing SSD caching algorithms (or at least one). Nvelo software has been ...
User Deals: Jesper Berg
Samsung 860 Evo SATA SSD
1 week ago

Samsung 860 Evo SATA SSD

$109.99 $72.00 -35%
All drives in Samsung's EVO series use TLC (triple-level cell) NAND – a more affordable type of memory chips compared to the MLC NAND used in the Pro ...
Best deal at:
Samsung 970 Evo NVMe M.2 SSD
2 weeks ago

Samsung 970 Evo NVMe M.2 SSD

$149.99 $119.99 -20%
Endurance rating varies by capacity.
Best deal at:
Samsung 970 Pro NVMe M.2 SSD
3 weeks ago

Samsung 970 Pro NVMe M.2 SSD

$199.99 $177.99 -11%
Best deal at:
Browsing All Comments By: Jesper Berg
  1. JJ,
    Good point. First of all, I haven’t gotten my hands on an x25-e unit yet 🙁 On the other hand, the x25-e is explicitly a server/enterprise drive based on SLC NAND, so it wouldn’t be a fair apples-to-apples comparison.

  2. Thanks a lot for the interesting info, Richard! Sounds like a great setup. Those stripe sizes are always an interesting experiment, but most users seem to get the best results when skipping the small sizes and choosing 128k or larger with SSDs (although i have no idea what usually works best for CAD).

  3. Thanks Daniel for the comment. And I agree it’s weird. I’ve also seen other tests and noticed the discrepancy. The new SandForce controller is nearly always better than Intel’s, but for some reason this particular system didn’t see it that way. For what it’s worth, I’m personally convinced that a long-term test would knock out both the Intel and Crucial drives in terms of “enduring performance” (an aspect that the RealSSD in particular has had problems with.)

  4. vt: Thanks for yout input and sorry that you got stuck in moderation. That issue should be fixed now. I agree with your views on the “enterprise” moniker and there’s no doubt that HDDs are the cheaper alternative even if you build a failsafe RAID 1+0 array vs. no RAID for the SSD. As for effective lifetime, MLC drives are rarely if ever used in constant 24/7 write load environments, but even then a 3-4 month lifespan is extremely pessimistic considering the average MTBF values of today’s drives.

  5. Thanks for the insightful comment, vt. What are the obstacle then, that need to be overcome for TLC (as well as MLC) drives to increase durability? Surely wear leveling, trim and other controller improvements have been beneficial for MLC drives. At any rate, if it’s good enough for faster and cheaper removable media than it’s definitely useful.

  6. Thanks vt, those are VERY interesting calculations. On the other hand 99% of first time SSD buyers will no doubt end up with an MLC drive anyway. Price tags tend to speak for themselves and for a laptop of simple workstation capacity is also a major factor (especially in laptops).

    For servers it’s of course a whole other matter. But wouldn’t you say that MLC drives can still be an alternative, considering the entire (relatively short) life cycle of the average PC?

  7. Great categorization. It would be really interesting to delve a little deeper into that subject, and try to find the scale tips in favor of SLC/MLC on the cost/endurance scale. Are you by any chance available for a guest post? 🙂

  8. MitchFL: As vt mentions it’s the manufacturer, meaning it’s a different PCB and firmware. They both have the same controller though, so there shouldn’t be a noticeable difference in performance. There’s a rumor that OCZ has an exclusive arrangement with SandForce that gives them the option to “unlock” some extra potential in the drives, but Corsair is supposedly on top of this by circumventing any such lockdowns in the firmware.

    vt: You are of course right, lifetime is determined by the NAND. By reliable I meant relative to older controllers. Particularly early JMicron ones that I’ve had the distinct displeasure of experiencing first hand (loss of data, stuttering, performance deterioration).

  9. Many thanks Les! I have to admit though, that PCMark Vantage is a bit unpredictable in my experience. It seems overly dependent on other factors than the drive even on the HDD specific score.

  10. True, thanks for pointing it out. I removed the value. It’s odd that the manufacturers keep using it.

  11. Very interesting.
    The MTBF value was obviously intended for hard drives, but it’s pretty useless there, too. After about 20 years of owning and building computers, hard drives are without a doubt the #1 least reliable part in my experience. I believe the longest lasting drive was the 20MB one in my first 80286-based PC. 🙂

  12. Enigma,
    Thanks for your comment, let’s clear this up. The statement is that the SandForce SF-1200 controller (SSD processor) is killing the competition right now, which I stand by.

    Moreover, this is a 2.5″ MLC drive, the Z-drive is an enterprise PCI Express SSD, so it’s a big difference.

    On a side note, the new OCZ RevoDrive uses four Sandforce controllers in each PCI-Express SSD.

  13. Thanks for sharing Jack! It’s pretty amazing what a good ssd can do even for lower-end systems. Watching your desktop load up almost instantly for the first time is one of those moments that makes you never, ever want to go back to hard drives for anything besides backup.

  14. @vt: I totally agree. It’s ridiculous to tout them as a new generation when in fact they will perform worse (in the long run) than older devices! We have gotten used to that a smaller production process is always beneficial, but not so for SSDs, apparently. This whole marketing strategy is very misleading.

  15. Hmm.. I thought they “solved” the problem with overprovisioning? At any rate, the move to 20nm fab does not fee like a step forward however it’s justified.

  16. Thanks a lot for the reply. It will surely be interesting to see what happens with the 25nm drives in the long run.

  17. Perhaps the manufacturers are simply counting on the 25nm generation to provide “enough” endurance as not to be noticed by most users? At any rate, the hype and marketing tactics surrounding 25nm is a bit odd to say the least.

  18. Thanks for the correction. I’ve updated the piece.

  19. Suddenly that ugly avatar I keep getting suits me perfectly 🙂

  20. Thanks vt! Actually a ton of research has gone into this from my colleague Glenn who wrote the piece and we plan to update it regularly. Hopefully it stays fresh for a couple of weeks… It’s a bit opportunistically based on “top speed” only, but that seemed logical considering this site’s name.

    And yes, something has to be done about those avatars 🙂

  21. Yes, I’ve also seen some web hosting providers offering “the latest” MLC SSD in their dedicated boxes (for a hefty premium, of course). Angry consumers is one thing, pissed-off companies with legal departments another. It will be interesting to see what happens down the line.

  22. Right now things seem to be going in the exact opposite direction, with manufacturers bringing more MLC into the enterprise market…

  23. I’m not entirely sure what you mean. As far as I know there’s no additional costs involved for the consumer when using a rebate like this, but if you are counting the entire marketing infrastructure involved we could make it a long discussion indeed.

    For SSDs specifically I know for a fact that some retailers are selling them at cost or even at a loss just to catch customers who hopefully buy other stuff as well. This has been a reasonably successful business model for consoles like the PS3/Xbox360 (both sold at a loss, with expensive games on the side), but I can’t see how it will work in the long run for SSDs.

    But perhaps the shrink to cheaper 25nm NAND has helped in this regard(?) In that case I have no idea whether it’s the manufacturers or retailers who are better off. Not that I mind either of them making money for their efforts as long as the marketing is honest, but as you know there have been a few question marks there lately, without naming names.

    On a side note I have one of these little 25nm wonders on the way (but the 256GB version of the m4/C400). Will post a review when it gets here.

  24. @CMW: Yes, the RevoDrive X2 could probably be listed as drive #11 with 740/720MB/s reads/writes

  25. Hehe, I can imagine. I have not seen the actual terms of the warranty, but my wild guess is that putting it in a server with high load and then returning it (when it inevitably wears out before the warranty expires) is covered by a separate clause 😉

  26. On the other hand this is a consumer drive and hopefully most buyers recognize it as such. I say “most” because I know that everyone doesn’t, sadly.

  27. @Simon: That’s because the “get go” was a couple of months ago. It’s been updated.

  28. OK, that clarifies things 🙂

  29. That’s very interesting. They are careful to point out that this a 1st-gen prototype though, so a lot of things will probably happen before/if it’s universally adopted.

  30. An interesting and somewhat discouraging calculation. It’s too bad that the “universal solution” MRAM is so far off.

  31. If i remember correctly, they are calling the 3000000 cycle SLC “Enterprise” SLC, but I have no idea either whether it’s actually used in any product today.

  32. Good idea. The HyperDrive would definitely be interesting to take for a spin. It’s too bad about the SATA II interface, it looks like a RAID card and a dual drive setup is practically a must.

  33. I could also add that I’ve never owned a computer that I haven’t had to do a hard reset on at some point.

  34. Thanks Albatorsk for the Linux info! I’ve updated the post to reflect this.

  35. True, but correct me if I’m wrong here, doesn’t this erase cycle “even out” since it would have to go through this cycle on subsequent writes anyway?

  36. Thanks for the clarification! Good to have you around 🙂

  37. Unfortunately you have to burn the firmware ISO file to a CD first and then boot from it. On a Mac that would mean holding down the option key.

  38. @maggou,
    Thanks for the links. It certainly looks like a revision of the list is needed. I would personally not interpret the numbers as fake though, they are probably correct in synthetic benchmarks under optimal conditions. On the other hand, the Vertex 3 has had its share of firmware issues, but it is starting to look more and more solid and mature.

  39. So, those numbers are actually DRAM cache figures? To be perfectly honest I have no idea how they manage to reach those numbers…

    Nice to see you again BTW 🙂

  40. No I haven’t watched the ticker price (until now), but it’s sad to see yet another dud from AMD. Meanwhile Intel is taking it real slow and pushes Ivy Bridge forward. Makes you wonder… There’s no end to the misery either, from looking at AMD’s “roadmap” to “Excavator” in 2014. Still no big changes in sight. This is just bad for everyone except Intel. The only place where AMD at least has some potential is in the mainstream/budget notebook segment. If they manage to create proper drivers for their dual graphics setup it might entice some gamers to the platform. Still, the CPU itself is no match even for a Core i5.

  41. I couldn’t agree more. Of course, OCZ could claim any number of years without telling an outright lie. It might last for 10 years or 15 as long as you don’t insist on pushing the “on” button too often. I don’t see how a 30% price reduction (if all of it is passed on to the consumer, which is unlikely) would justify such a huge drop in durability in any situation, save for perhaps a memory card or USB stick.

  42. @Bubu: Thanks for the tips! However, the RAM drives you link to are volatile and belong in their own separate category IMO.

  43. That’s clearly a disappointment, but hardly surprising from Intel at this point. With AMD so far behind it looks like they are trying to squeeze as much as they can out of the market by rebranding old crap instead of launching new products that are already ready for prime time.

  44. Thanks for the correction. They have done pretty well in the marketing departments though. A lot of tech writers that should know better still seem to equate new production processes with “better”.

  45. Thanks for the info Kenneth, looks like you’re right about the freezing part, but at least there’s a fix (originally intended for the C300 but also works with the m4):
    That doesn’t help Mac users unfortunately.

  46. Yes it does sound pretty incredible. Lots of over-provisioning?

  47. But it has “adaptive flash algorithms and advanced techniques for signal processing”! Seriously though, it does sound a little too good to be true. It wouldn’t hurt if they were a little more forthcoming about exactly how this is accomplished.

  48. @Alan: The difference between the SF-2281 drives is marginal, and the quoted speeds are not to be taken as gospel. In real-world tests the Vertex 3 Max IOPS generally beat most of the competition (MLC in the same form factor). But you have a point, this listing is due for an update.

  49. vt: Interesting thread. It’s hard to defend the indefensible…

  50. Thanks vt. I hope this paper gets a lot of attention and that it puts puts some perspective on the current tradition of marketing of flaws as advantages (unlikely, but who knows).

  51. LOL. Yes, neon lights is probably just what the kids need. I also average gigabytes per day on my main PC, but I also have a couple of laptops that average way less. None at all in servers, since I don’t need the throughput and SLC drives are hideously expensive. I’ve seen a couple of web hosts that will “upsell” you a single Intel 300-series (or similar low-end drive) though. For a hefty fee of course 🙂

  52. I’m curious as to what data exactly these drives store. Wouldn’t it make more sense to put the entire OS on the drive and keep it there?

  53. Hmm, yes, I suppose that makes sense. I would still very much like to know how that caching algorithm works. Everything that can’t be controlled or checked manually makes me nervous.

  54. I didn’t know that about the LCD situation, very interesting. Well it all boils down to milking your cash cows I guess (i.e. us).

  55. If it costs an entire $10 I wouldn’t bet on them adding it…

  56. @odiebugs: About the difference between the Vertex 3 and 4, yes, but it’s not quite that simple:

  57. Thanks Bruce! An update to the enterprise section is long overdue. Thanks for the reminder.

  58. Thanks for the catch Mark, this has to be changed ASAP.

  59. Hi Timothy, yes it’s a very slow process apparently, with few updates. Current status of SATA Express is “in ratification”…

  60. Marx; you are aware that ATTO results often exceed specs for SSDs? Are you suggesting that the write buffer be disabled, why? Please elaborate. Naturally, default settings are (and should) be used unless otherwise noted.

  61. Hi Ashkan, sorry for the late reply. The interface is important because it is may cap the speed of an SSDs (not
    mechanical drives, which are too slow). For example, most recent SATA III (6 Gbps) SSDs are capped by the interface, which is why you need to move to PCI Express (PCIe) to see an improvement. To confuse things a bit further, PCIe can be in regular PCIe card form factor or the smaller, on-board M.2 (the latter of which may also use a SATA interface, not to be confused with mSATA).

  62. Thanks for the info! Post updated. I was totally unaware of their Dell connection, but Lite-On drives seem to be used quite extensively in Dell systems.

  63. Thanks for pointing that out, the numbers certainly look very good!