New Intel 520 Series SSD Taps SandForce Controller

crookedvulture writes "Intel continues to partner with third-party controller makers for high-end SSDs. Its new 520 Series drives pair the latest SandForce controller with Intel's own firmware and 25-nm NAND. HotHardware, Tech Report, and PC Perspective all have reviews of the drive, and the verdict is pretty consistent. While the Intel 520 Series offers slightly better performance than competing SandForce solutions, it also costs 30-40% more. That's a steep margin even considering the Intel SSD's five-year warranty."

Comments

Re:super fast

[Anonymous Coward]

Congratulations, you win 1 (one) ASCII boner.

8====D

Sometimes

[Gr8Apes]

Ad yet my X25M, which came at a steep price, is still performing well 1.5 years later while several others I know have gone through several competing drives in the same period. Anecdotal, I know, but when I decided on a drive, the only one without the kind of failure stories my colleagues went through was the Intel drive. Sometimes you do get what you pay for.

Re:Sometimes

[sexconker]

Ad yet my X25M, which came at a steep price, is still performing well 1.5 years later while several others I know have gone through several competing drives in the same period. Anecdotal, I know, but when I decided on a drive, the only one without the kind of failure stories my colleagues went through was the Intel drive.

Sometimes you do get what you pay for.

I went for the Crucial M4 (2x256GB in RAID 0) instead of the Vertex 3 because I just didn't want to deal with the bullshit.
No idea why Intel went with a SandForce controller, though. They have a good reputation with their SSDs - why risk it?

Re:Sometimes

[spectral7]

Anandtech mentions that the 520 went through a year of testing, so it should be much more reliable than other SF-2200 SSDs. Also fixed a BSOD issue in one of their systems, which was also using SF-2281.

Re:Sometimes

[Cimexus]

Yeah the x25 is a great drive, no doubt about it. It was just so expensive compared to the alternatives back then... I could have bought three competing drives for the same cost. I keep backups so drive failures aren't a huge problem (though still annoying).

Luckily though I haven't had issues with my SSD at all. A Corsair Force drive well over a year old now and still performing great. I'd heard the stories about how most SSDs seem to die quickly so I was paranoid at first, but maybe I just got one out of the good batch or something ;)

Re:Sometimes

[blackraven14250]

I was paranoid for a while too. I got two SuperTalent drives, hooked them up in RAID0, and have been doing weekly images because of the track record they have obtained. A year and a half later, they're still surviving.

Re:Sometimes

[hairyfeet]

The problem with SSDs was nailed by the guys at coding horror and that is SSDs are still new enough tech it works on a "Hot/crazy" scale. Smoking hot speeds, crazy failure rates. A couple of my gamer customers have gone through like 4 of the things each in the past two years, and these ain't the cheap ones either, its whatever scored highest on the benches which is usually the top dollar stuff. They've already learned don't put anything on an SSD you give a shit about unless you have it backed up because unlike a HDD which will give you some warning the SSDs just go "poof" and that's it. No getting your data back, no tricks, just one day it works and the next its toast.

Re:Sometimes

[Anonymous Coward]

If you were really paranoid, you would not use RAID 0. Or are you a lazy paranoid?

Re:Sometimes

[DigiShaman]

It's what I call the holy SSD Trinity. NAND, Controller, and the Firmware.

The quality of the hardware is important, but these days it's more or less commoditized across the industry anyways. So that leaves the firmware. Given the level of voodoo nessary to pull of the LBA remapping magic and garbage collection across all the cells, I would rather spend my money on a reputable company such as Intel. Especially on a new form of storage as complicated as this under the hood.

Re:Sometimes

[fnj]

It's BECAUSE he's tempting fate by using RAID0 that he is paranoid. RAID0 is like throwing the middle finger at fate while doing a jig.

If you were really paranoid, you would not use RAID 0. Or are you a lazy paranoid?

Re:Sometimes

[Anonymous Coward]

I'm glad you went with an SSD that induces data loss after 5200 power-on hours. See their (Crucial's) site for a firmware update.

Re:Sometimes

[AHuxley]

crazy failure rates?
3-5 year support was often on offer for consumer units, 7 years for enterprise $ now too (28% over-provisionin).

Re:Sometimes

[smash]

Support to get a replacement drive doesn't mean dick if your data is gone. And warranty won't recover that.

Re:Sometimes

[Anonymous Coward]

I completely agree. I put one in my work laptop and after 14 months it just decided to crap out. No warning at all.

Re:Sometimes

[hairyfeet]

Exactly, thanks smash. I mean who gives a flying fuck about the drive? but if you are working on some important term paper or business proposal and the SSD decides NOW is the time its gonna crap itself? That's when you realize that hot speed isn't worth the crazy failure rates. I mean either you have every. single. change of any note backed up or you are literally rolling the dice. try contacting them and asking about your data, they'll tell you tough luck, it isn't covered.

One of my gamers learned the hard way on the first failure as it was on his gaming laptop and he had uploaded his recent vacation pics to it and hadn't had a chance to backup before it crapped out. With HDDs I can usually get something back, even with one with a dying controller i can usually put it in the freezer and then use spinrite and get it to function long enough to get at least some of the data back, but when that SSD went there was nothing, the BIOS couldn't even detect a drive, it was just gone. You can't even do the old controller board swap trick, when they go.....its toast.

So who cares about the drive? Drives are a dime a dozen, its only the data that anyone cares about. Honestly after seeing the failure rates I decided instead to max my machines out on RAM and use Readyboost and with 8Gb each in my desktop and netbook I'd say the only place an SSD is gonna beat me is boot which I almost never do. With everything else superfetch has it already loaded into RAM based on what times I use it and even the best SSD still can't compete with RAM for speed. So I think I'll pass, and wait until i can actually get at least 5 years out of a drive, not just have it replaced for free if it goes tits up. Hell I just used up the last of my 40gb and 80Gb IDEs on some old offlease boxes to dump on CL, every single one of them worked just fine even after sitting in a drawer for years.

Re:Sometimes

[tlhIngan]

A couple of my gamer customers have gone through like 4 of the things each in the past two years, and these ain't the cheap ones either, its whatever scored highest on the benches which is usually the top dollar stuff.

That's the problem.

If you want the new hotness in SATA3 OMGWTFBBQ speeds, you're going to risk crappy firmware. It's because the firmware writers are under pressure to deliver software so the SSD manufacturers can deliver SSDs a month earlier than the competition with their new fast SSD.

So yeah, there will be lots of bugs if you're constantly going cutting edge.

On the other hand, OEMs like Dell, Lenovo, and notably, Apple, are shipping millions of SSD-equipped computers every year. If there was a massive failure rate, we'd hear about it - especially Apple, which shipped millions of SSD-equipped MacBook Airs and Macbook Pros and others. Hell, you'd hear of petitions asking to recall every SSD-equipped Mac if they were dropping like flies. Hell, if even 1,000 of them failed, you'd probably hear of a recall petition

The difference? The OEMs are using SSDs that are more conservative - sure they're not going to push 250+MB/sec (probably 150MB/sec or so), but they're far more reliable. The controllers have long matured and the firmware is stable and major data loss bugs are now squashed.

Perhaps rather than caring about 150/200/250/300 MB/sec transfer rate (you're talking about seconds in difference at that point), your friends should've bought what was last-generation and stable. Perhaps even see what's in stores like Best Buy because they take so long to get product that all the first adopters would've been burned that by the time retail stores have 'em, they're all stable.

The OEMs usually buy Samsung or Toshiba (with Samsung or Toshiba controllesr, respectively) SSDs. They'd probably use Intel if they could, if it wasn't for the fact every generation, Intel has some SSD problem in the early firmware. The original G1s had a speed degradation issue. The G2's had a massive data loss bug. The G3's also had a dataloss bug. These new ones? Probably a data loss bug as well. Sure they get fixed, but buying a last-gen will probably be a good idea. Or wait 6 months to see what issues arise.

Re:Sometimes

[Anonymous Coward]

If you were really paranoid, you would not use RAID 0. Or are you a lazy paranoid?

Same chipset, same size, same manufacturing date. Probably only one digit differs in the S/N.
Using RAID 1 during those circumstances is like building two houses next to a volcano instead of just one. Sure, you might get lucky and only one of your houses catches fire but I wouldn't count on it.
He is obviously using RAID 0 for speed, the real protection is the weekly images and that is a lot better than any RAID setup can hope for.
The only reason I see to go for any other RAID setup than 0 is if you want uninterrupted function through a disk failure and it is a service that doesn't lend itself to multiple servers. For everything else backup is superior.

Re:Sometimes

[MemoryDragon]

I have been running flawlessly now for a year on a non intel drive.

Re:Sometimes

[cheekyboy]

unless the SSD delivers false 0x00000000000s or 0xffffffffffffffffs. it should never blue screen, normal HDs will never blue screen, even with lots of read errors, they just get super slow.

Yes, the drivers at the OS level are probably crap.

To deliver fast speed from one single file, is trivial, its delivering fast cached, in order, small reads queued at the rate of 50000 IOPS that can push the controller chip to its limits, and some stupid buffer overflows happen.

This is what you get with code that has few range checks, input arg validation, even in loop sections that assume a value is kosher might get 'corrupted' and point to random disk offsets.

Better code takes more IFs , which is slow, so either dodgy code is made now, or we wait years for faster controller chips mhz.

Re:Sometimes

[TheDarkMaster]

I use two Samsung HDs on RAID0 for more than four years, no errors even since. My two cents to "ZOMG, RAID0 will destroy all your data and kill your cat!!!!"

Re:Sometimes

[makomk]

I think some of the SSD brands intentionally disable all access to user data once they detect a single failure, which doesn't really help. Probably very profitable for data recovery firms though.

Re:Sometimes

[smash]

And this is where ZFS (or some other intelligent filesystem that is aware of the notion of heterogeneous storage) is going to be a win. Use spinning rust discs for the storage, and an SSD for accelerating read/write. Unfortunately neither Windows nor OS X have the facility to make use of this in most machines.

Superfetch is all well and good, but USB flash is still slow, and RAM is limited.

I'm keen to see the performance of the current/next generation of seagate hybrid drives. If they can get to something like 2tb with 40gb of SSD cache with intelligent enough caching they should be fairly attractive.

And yes of course everyone should have backups. But as you mention, A: your backups may not be up to the second and B: even if you have the data backed up, you're still without a drive to work with until the RMA comes back. Which can be weeks.

There's an oracle DB tuning document floating around that has been relevant since forever ago, the theory being "SAME" - Stripe And Mirror Everything.

I'm building RAID 1 or RAID 10 or ZFS into all my file storage boxes to protect against drive failure irrespective of storage medium. Disk is cheap enough now that running a single disk, or cheaping out with RAID5 just isn't worth it any more.

Re:Sometimes

[smash]

By "protect against drive failure" i mean keep running. I still back the stuff up manually and run snapshots for the "oh shit i deleted my stuff" event(s). The RAID is to keep running while waiting for RMA or replacement disk without needing to reinstall/restore from backup.

Re:Sometimes

[Anonymous Coward]

Take backups and enjoy your performance (if you have demanding usage).

At work I put together a bang for buck analysis PC with an overclocked i7-2600k and 4 spinning drives in RAID0 using Intel fakeraid. Results and OS are backed up weekly on a 5th drive.
The analysis in question thrives on sequential disk performance when running large models.

Re:Sometimes

[TheLink]

Yeah seems to me that Intel is actually helping Sandforce fix the bugs in Sandforce's product. Not for free of course, but one wonders about the long term strategic reasons for this. They could have let Sandforce keep churning out SSDs with buggy firmware.

Maybe in a world with more fast, cheap and reliable drives, Intel can sell more high power CPUs? Or this is a move against Samsung?

Re:Sometimes

[Anonymous Coward]

RAID0 just increases the chance of data loss, as any failure in either drive is a failure for the entire array. Now, RAID0 + two drives that are well into the start of the second slope of the bathtub curve and probably have already depleted a lot of their spare cells, that's just foolhardy.

Unless, of course, you simply don't care and it is no major hassle to reinstall everything. You did not give us enough data to gague that.

Re:Sometimes

[Miamicanes]

> Using RAID 1 during those circumstances is like building two houses next to a volcano instead of just one.

Yep. Just ask anybody who bought a pair of OCZ Velocity2 Sandforce-based drives and mirrored them in the hope of avoiding data loss... and had both of them simultaneously commit suicide thanks to the wonderful bug that triggers the "3 minutes to deathcrash" condition (never fear, though... the drive itself is "fine" -- just Securely Erase (tm) it, and it'll be good for a few more weeks until it kills all the data on it again).

Sandforce is the worst steaming pile of shit that has ever existed in a mainstream product. Their controllers have fucked so many users, they can't even get ENTHUSIASTS (who lose sleep at night worrying that their controller might be .05% slower than somebody else's) to care about their benchmarks anymore. The only reason I won't be chucking mine is because I might need it for the class action suit against OCZ and/or Sandforce someday.

Re:Sometimes

[JDG1980]

Yeah seems to me that Intel is actually helping Sandforce fix the bugs in Sandforce's product. Not for free of course, but one wonders about the long term strategic reasons for this. They could have let Sandforce keep churning out SSDs with buggy firmware.

The Anandtech article indicates that the fixed firmware will be exclusive to Intel for a certain (unspecified) period of time, before being made available to other SandForce vendors. So what Intel gets for their time and effort is a period of exclusivity.

Re:Sometimes

[JDG1980]

Exactly, thanks smash. I mean who gives a flying fuck about the drive? but if you are working on some important term paper or business proposal and the SSD decides NOW is the time its gonna crap itself? That's when you realize that hot speed isn't worth the crazy failure rates. I mean either you have every. single. change of any note backed up or you are literally rolling the dice. try contacting them and asking about your data, they'll tell you tough luck, it isn't covered.

That's why you store your actual data on magnetic drives - preferably on a NAS with ZFS software RAID. The SSD should only have the OS, installed software, and miscellaneous stuff like cache and swapfile. All this can easily be replaced from a drive image; at most you might need to reinstall a couple of software updates. If you keep a lot of important data on a SSD, you're doing it wrong. Putting the OS and software on SSD gives you all the speed advantage while minimizing downside risk.

Re:Sometimes

[TheDarkMaster]

I do not understand your apparent anger against RAID0 ... For me it works very well, why it annoys you? I have a third hard drive for backup, the HDDs used in RAID are reliable (not SSDs, still too experimental), and according to reports from the S.M.A.R.T. they still have long useful life ahead. Oh, the boot disk is in a SSD (yep, four drives rig, the RAID0 is a gaming/general data drive), but I do not mind having to reinstall if there is a problem here, since it is still experimental and therefore I do not put anything critical on the SSD.

In short: RAID0 increases the chances of data loss? Yep. But is not the panacea that everyone writes as if he was going to explode when they were powered on, or kill your cat.

Re:Sometimes

[TheLink]

That period of exclusivity is what Intel gets out of it in the short-term and not long-term.

Intel generally thinks longer term than most US IT companies.

Re:Sometimes

[hairyfeet]

Well as far as RAM goes with it being so cheap frankly maxing out the board is usually the best advice anyway. With my 8Gb of DDR 2 800 (which I bought when DDR 2 was as cheap as DDR 3 is now) in my desktop all but the largest games are prefetched based on my usage, for example Comodo Dragon is pretty much kept in RAM 24/7 while WMP 12 is loaded every evening. With my netbook which I strictly use for media and web that 8Gb of DDR 3 is total overkill and all the apps I use are always prefetched into RAM Most would say 8gb was overboard for an AMD E350 netbook but again by buying when it was dirt cheap the difference between 4gb and 8Gb was a whole $6 so it really wasn't a big deal to max out.

Now with readyboost the trick is while it WILL run on just about any stick frankly its better to buy a 4Gb stick that is crazy fast than a 16gb that is slow. Look for something that gets 5Mbs or even better 10Mbs and since readyboost uses it for the small read/writes it really is superfast, IMHO it gets pretty close to the speed of a hybrid with the only penalty being it adds a few seconds to shutdown. With a netbook/notebook with a card reader its even easier to get crazy speeds because the card reader is usually hooked straight to a PCIe lane so using a 10x or better SDHC makes for a hell of a kick in the pants.

Finally you may want to look into ReFS which is gonna be coming out with winServer 8 as that is the next gen filesystem for Windows and will address many of the things you are asking for. But in many cases RAID simply isn't an option, you can't RAID a laptop for example and in many desktop the spinning rust discs are takin g up too much space to RAID SSDs, not to mention the cost of RAID SSDs. Also I'm not sure that having RAID 1 on SSDs would even be helpful as I have seen a bad batch of discs have multiple failures and from the looks of things SSDs are the same only worse. Ultimately this is why at the current time i simply can't recommend SSDs to my customers as the risk VS the reward is too great. while its all well and good to back up often having the time and keeping everything current can quickly become impossible, especially with mobile devices. Then finally add in the costs of the backup spinning rust at a time the HDD OEMs are gouging because of the flood and you can see why people can easily get a week or two behind. I have my backups scheduled for the day after patch Tuesday and i'm just too busy at the shop to do multiple backups more than once a month and with 3 main machines doing a triple disc image can turn into an all afternoon job VERY quick.

Hell here we are talking about NAND flash failing and just this morning i plugged my barely 6 months old 16gb flash stick in the PC and found its gone, a practically new TDK stick just poof! no files, no nothing. Good thing that didn't have anything on it I cared about but its a good example of how flash can just "go" without the slightest bit of warning.

Re:Sometimes

[smash]

Yup, my strategy for laptops is store anything important on a server with RAID1. :)

Re:Sometimes

[Gr8Apes]

I've had a lot of experience with RAID over a long time, both in the enterprise and at home. I prefer software RAID 0 for home systems. I'm currently running a software RAID 0 stripe across 4 drives for roughly a TB of fast workspace space. The remainder of those 4 drives is used as normal storage, with a fifth for backing up that RAID 0 stripe. I no longer use any other RAID for the home. I don't have uptime requirements, which is all redundant arrays are good for anyways. I use another set of external drives to keep backups of the non RAID data.(Yes, that's a lot of drives and a lot of space, I haven't lost anything in over 15 years and a number of drive failures, not to mention that drives are cheap) My home system also uses that X25M for it's boot disk. It was a test, and it's passed. It's fast enough that while I may try a RAID 0 stripe set of SSDs for the system drive, I don't see the need at the moment.

In the enterprise, at this point, I wouldn't use anything other than hardware RAID10 where I needed performance, and RAID 1 elsewhere. Uptime is usually quite important. A good controller will actually use both disks in RAID 1 for reads and the striping can be either software or hardware, depending upon your layout and needs. These aren't the crappy cheap home controllers, and there usually is no performance hit for using hardware based RAID.

Re:Sometimes

[Gr8Apes]

Actually, it's a combination of controller and firmware that seems important. And super fast but exploding vs a little slower but reliable wins in the long run everytime. It's why I got the Intel, it was the only "stable" one at the time.

Re:Sometimes

[hairyfeet]

But you see there is your problem smash, you are thinking like one of us geeks instead of like a normal. hell a normal if you mentioned RAID would go in their kitchen and hand you the bug spray!

You have to remember they have gotten pretty damned good with the failure rates of spinning rust, even in mobile. All the drives now have sensors that park the heads when its moving or jostled, sensors for heat, SMART, they are pretty damned good now. Hell the only way i was able to get folks doing regular backups was with butt simple USB externals and an even simpler software like Paragon drive backup.

That is why I warn folks away from the SSDs now, because they don't know about hot/crazy scales or insane failure rates, they just know "Wow that is a fast laptop!" and then when that SSD takes a dump they are hurting. You don't know how heartbreaking it is to be sitting in the shop with some girl crying her eyes out because there is a good chance the pictures of her late mom she forgot to backup could be gone forever, but at least with HDDs there are usually tricks i can do to get a good chunk of the stuff off, but not with SSDs, its all or nothing.

So I'd say anybody selling SSDs to average folks is frankly doing them a disservice. you and I know "Backup backup and backup some more" and for the important things like pics of my late sis and other important items I have onsite, offsite AND cloud based backups, but most folks frankly don't have the skills nor the dedication, they have stressed out lives and shit happens you know? So I consider it the job of us geeks to try to make their computing experience as safe as we can while still having it useful and I'd say that SSDs just aren't there yet.

Re:Sometimes

[smash]

Time machine for me generally "just works". Set and forget - it will do wireless backups of all my stuff to my time machine data store if it is in range (or plugged in, if i have it on an external drive). If my backup destination is not available, it will do hourly snapshots locally, and then roll them up to the backup store when it becomes available. Previous versions may have been a little flaky, but so far in Lion, it Just Works.

Windows really needs something similar to time machine. Windows backup in 7 is almost there, but it needs to be more intelligent, so you can have it automatically just work without popping up dialog boxes, etc. when the destination becomes available or if it is not available.

I value warranties

[timeOday]

With the concerns often expressed over SSD reliability, it's interesting that Intel's 5 year warranty now bests the other SSD's 3 year warranty, whereas hard drives are moving to 1 year warranty(!)

.

I get really worried when the warranty is so short, I would be extremely frustrated if the thing conked out the day after. Whereas after 5 years, I would probably be ready to move on.

Re:I value warranties

[AHuxley]

http://eshop.macsales.com/shop/SSD/OWC/Mercury_6G/ they had 3 or 5 years for some time. 7 years on their enterprise ssd's

Re:I value warranties

[jamesh]

I think i've posted this before, but in Australia the laws have changed wrt warranty. If you purchased something of "significant value" and it broke after 18 months when you would reasonably expect it to last for longer (eg you bought a laptop for $500 and it failed after 18 months) then the fact that it only came with a 1 year warranty doesn't amount to anything from a legal point of view - the place you bought it from has to provide a repair or replacement.

Hopefully this will stop stores selling crappy merchandise that they know isn't going to last long, or at least they'll be more likely to sell stuff that also has a manufacturer warranty (which is in theory higher quality).

price comparisons depend on the capacity

[Tumbleweed]

I noticed the price comparisons with other SF drives differ GREATLY depending on the capacity of the SSD you're looking at. When checking the 120GB version, the price difference isn't that bad, and if you're using it as a boot drive (as I hope you are), then the peace of mind seems well worth the extra money. Amazon has the 120GB version for around $230 or so.

Hopefully it's more reliable

[AaronW]

I just returned an OCZ Agility 3 due to it dying after 2 weeks of light use as my Linux root partition running XFS. Since their office was local I drove it there to save shipping charges and had to laugh in their lobby since they're touting reliability. Sadly Sandforce has gained a reputation for not being the most reliable controller out there. I have never had a hard drive fail except after years of use until this one died. Fortunately it's being replaced under warranty.

Also, unlike other drives I've had fail, this one died suddenly and completely. Suddenly Linux couldn't access the root filesystem and after rebooting the drive could not be found by the SATA controller. Fortunately I had my old Western Digital Velociraptor drive I had replaced so I swapped that back in.

I have another Sandforce SSD that would often not show up when my computer woke up from standby until after a firmware update. Intel might be better, having written their own firmware, but at this point I think SSDs have a way to go in terms of reliability. There's no excuse for the issues that have been reported with the various Sandforce drives. The problem is either insufficient Q/A testing of the firmware or management pushing out the firmware before it's ready. It's also possible that there might be some hardware glitch causing these drives to fail.

-Aaron

Re:Hopefully it's more reliable

[Cimexus]

Forget OCZ, they might look like the best performers on paper but they have woeful reliabilitiy rates. Even the non-intel competitors such as Crucial, Corsair, GSkill etc. are much better than OCZ in reliability stakes.

Re:Hopefully it's more reliable

[Anonymous Coward]

I have this problem with a Crucial Force 3 drive that I'm currently using. Still happens even on the current firmware level, so I've just been turning off the machine rather than using it in standby mode. Luckily the machine boots very quickly and I don't generally need it on for long periods of idle activity, so it's a minor annoyance. I previously had a Kingston V100+ drive in the same laptop and that didn't have any problems with going in and out of standby.

Re:Hopefully it's more reliable

[Anonymous Coward]

Awsome! Anecdata!

I have an OCZ Vertex LE which claims ~376 power-on-days (~9015 power-on-hours). After I updated from the *terrible* v1.0 firmware, I've had no trouble at all while using it as a primary drive containing EXT4/BTRFS/Swap partitions.

Re:Hopefully it's more reliable

[TheCouchPotatoFamine]

ocz agility 3 - mac osx - 5 mo. use without problem and even though i haven't passed the early failure window im tossing in the datapoint. *IF* it fails (i feel i trust it, but hey) i'll be here to bitch about it in the future :-) Rarely hear from the happy ones, i've found... i wonder what the failure rate really is?

Re:Hopefully it's more reliable

[incer]

Same as above, Agility 3, OS X on a 13" MBP, 5 months.
Heavy usage, we're talking about 12 hours or more monday-friday. A couple of hours on Saturdays and Sundays.
It's not my only drive though, data is stored on an HDD which replaced the DVD unit.

Samsung SSD 830 is also a good choice.

[teh31337one]

When I bought my 2nd SSD a few weeks ago (first was an M4 for my laptop, this being for my desktop), I opted for a 128GB Samsung SSD 830. It's a great little drive, and was ~30% cheaper than Intel SSD 510.

I went for the 830 over a Sandforce-based drive because of their reliability. There's a reason why Apple use Samsung SSDs in their laptops.

Re:Samsung SSD 830 is also a good choice.

[EjectButton]

When I bought my 2nd SSD a few weeks ago (first was an M4 for my laptop, this being for my desktop), I opted for a 128GB Samsung SSD 830. It's a great little drive, and was ~30% cheaper than Intel SSD 510.

I went for the 830 over a Sandforce-based drive because of their reliability. There's a reason why Apple use Samsung SSDs in their laptops.

Intel care more about reliability than anyone else in the SSD game and they are now using the exact same sandforce controller (SF-2200) in this new drive that others have been using for a while now (OCZ Vertex 3 for example). The problems in the past were more with the firmware than the controllers.

Re:Samsung SSD 830 is also a good choice.

[teh31337one]

That's great. But when I was in the market for an SSD, their were three options: M4, Samsung or a Sandforce. Intel was more expensive, and the SandForce SSDs had reliability issues. Samsung SSDs have a great track record for reliability too

Anandtech:

Samsung is a dangerous competitor in the SSD space. Not only does it make its own controller, DRAM and NAND, but it also has an incredible track record in terms of reliability It's also worth pointing out that Samsung SSDs are also one of the two options Apple rebrands and delivers in its Mac lineup. To continue to hold on to Apple's business for this long is an impressive feat on Samsung's part.

It's amazing to see Samsung come so far in the enthusiast space. From a drive that I simply wouldn't recommend to building a downright competitive solution backed by a near flawless track record.

Re:Samsung SSD 830 is also a good choice.

[Spoke]

Implying that SSDs are equal because they use the same controller is like saying a Apple iMac and a Dell are the same because they use the same CPU.

The firmware that runs on the SSD is highly unlikely to be the same on Intel branded drives compared to OCZ branded drives. And firmware is what is a leading cause of reliability issues on SSDs.

Pfft Sandforce

[Vegemeister]

The Sandforce controllers use compression to increase speeds and reduce write amplifications. Thus, they choke on encrypted (incompressible) data. They also have a horrible record of reliability. No thanks.

Re:Pfft Sandforce

[haruchai]

The reliability is fixable but the problem with incompressible data would need a redesign unless they have a way of quickly detecting such data and transferring immediately without compression. Looking at the files I have stored on spinning disks, I estimate it to be 75% incompressible data.

Re:Pfft Sandforce

[Anonymous Coward]

The cool thing about the Intel 520 SSD is that, as they place it as the successor to the Intel 320 SSD, they actually advertise in-drive AES-256 encryption support. This means you set an ATA password and they use it to encrypt the key for encrypting the drive (at least that's how it worked for the 320).

This is a controller feature you usually only get in the (too expensive) business versions of drives. In consumer drives they usually don't let you set the keys and all you get from the always present encryption is secure erase (throwing away the key for the current data and picking a new one at random).

Re:Pfft Sandforce

[Vegemeister]

Which is entirely useless for someone who doesn't trust closed-source encryption schemes.

320 Series Bug

[Kunedog]

I have bought six X25Ms in total; all are tremendous performers and give me no problems. The are all G2s and range in age from 1 to 2 years (except for a used 40GB one I bought a month ago).

Unfortunately the Intel 320 series (really the X25-M G3) has had its own reliability issues with a nasty firmware bug that causes it to suddenly report its capacity as 8MB (causing complete data loss).

http://www.pcworld.com/businesscenter/article/236468/intel_acknowledges_ssd_320_bug_working_on_firmware_upgrade.html

Intel on Sunday acknowledged that a bug could cause its SSD 320 solid-state drives to fail, and said a firmware upgrade is on its way to address the problem.

In some instances, a power loss may cause Intel's SSD 320 drives to crash and lose data. On rebooting the system, the system BIOS could report the SSD as having only 8MB of storage capacity. Intel two weeks ago said the error was possibly a bug, and that the issue was being investigated.

"Intel has reproduced 'Bad Context 13x Error' utilizing strenuous testing methods. This 'Bad Context 13x Error' can be addressed via a firmware update and Intel is in the process of validating the firmware update. A future update will define the schedule to deliver the firmware fix," an Intel spokeswoman said in an e-mail statement.

It's ironic that a power failure triggers this problem, since Intel had marketed the 320 as especially resilient to them:

http://www.anandtech.com/show/4244/intel-ssd-320-review

Intel always prided itself on not storing any user data in its DRAM cache. The external DRAM is only used to cache mapping tables and serve as the controller's scratchpad. In the event of a sudden loss of power, Intel only has to commit whatever data it has in its SRAM to NAND. To minimize the amount of data loss in the event of a sudden power failure, Intel outfitted the SSD 320 with an array of six 470F capacitors in parallel.

Some posters say it can happen without a power failure:

http://communities.intel.com/message/133499

Intel said they found the cause and released a firmware update, but applying it seems to have actually triggered the bug in previously problem-free drives for many posters:

http://communities.intel.com/thread/24121?start=0&tstart=0

Intel has not acknowledged any problems with the fix, nor told anyone which serial numbers were affected. Nobody has reported on the bug since Intel said they fixed it, including Anand.

This issue was enough to convince me to buy something else (even though the 320 series would otherwise have been my first choice) when I had to shop for an SSD last month. I found a used "like new" (according to the SMART data, at least) X25-M G2 on amazon instead.

Interestingly, X25-M G2 prices have held steady ($2/GB or so) and only gone up over the last year. Yeah it's probably because of dwindling supply, but I can't help but suspect that lack of confidence in the 320 series may have contributed to an increase in demand for the G2.

Re:320 Series Bug

[Bryan Ischo]

It sounds like you are trying to convince everyone, including yourself, that you were better off to choose a G2 over a G3.

There is no merit to your argument. The G3 bug you mentioned was real but it was fixed by Intel's firmware update, which is why you haven't heard anything about it.

There is nothing wrong with the G3 that would suggest that the G2 is a more reliable option. There is little to recommend the G3 over the G2 either, except price and availability.

I personally own two G2's, one G3, a real old skool PATA 32 GB SLC SSD from Mtron (perhaps the very first performant SSD), a cheap-o Kinston value series SSD, and a super duper cheap-o "SSDFactory" 32 GB PATA SSD direct from China that I bought because it was the only thing that would work in my ancient Panasonic Y2 laptop.

Not one of these drives has experienced any problem of any kind, and I've had some of them upwards of 2 years.

I would never buy an OCZ drive though. They are terrible.

Re:320 Series Bug

[makomk]

Intel has not acknowledged any problems with the fix, nor told anyone which serial numbers were affected. Nobody has reported on the bug since Intel said they fixed it, including Anand.

Anandtech has a decidedly pro-Intel leaning in general. If memory serves me correctly they only reported on the problem at all once every other site had done so, and even then only because their readers were giving them hell for failing to report on the fact that drives Anandtech had recommended were failing en-masse and eating people's data.

warranty math

[NemoinSpace]

3 extra years warranty 2/5. So you would expect to pay 40% more.
2 extra years warranty 3/5. So you would expect to pay 60% more.
So if you were just buying a warranty alone the price would be worth it especially since the summary states you're only paying 30%-40% more.
Add the fact that your buying from a top quality manufacturer is just icing on the cake.
Note the fact that HDD manfs have cut their standard warranty to 1 year, should show you what crap they are producing these days. Swapping hardware takes minutes, restoring data takes lots of hours.

Re:warranty math

[NemoinSpace]

ok so my math sucks- bit you know what i meant.

The Reason I Buy Intel SSDs . . .

[Kunedog]

The reason I always bought Intel SSDs was because they always had Intel controllers, and Intel controllers were always stellar. Intel for years prioritized random access and did it better than anyone else. The 510 series, the first to be released with a non-Intel (Marvell) controller, seemed like such a pointless drive, with worse random performance than the 2 year old X25-M.

The problem is that sequential numbers are always faster and therefore easier to market to people who don't know any better. The 510 served a marketing need, not an engineering one.

I don't follow the market as closely anymore so I don't know if the 520 series makes any more sense, but if we know that Intel was willing to sacrifice random access performance to use a Marvell controller. What will they sacrifice for a Sandforce controller? Reliability?

Re:The Reason I Buy Intel SSDs . . .

[Bryan Ischo]

You really should read something, anything before bothering to spend the time to post.

The 520 is faster at every metric (random read/write and sequential read/write) than the Intel controller based drives.

It also had a full year of vetting by Intel before being released, and they are putting the same 5 year warranty as their other drives; there is no reason to believe that it will not be as reliable as Intel controller based drives.

The only thing that doesn't compare favorably with this drive is the price.

standby / wakup issues

[bored]

Besides, the well known reliably issues with SF controllers, i wonder if they fixed the resume problems that seem to crop up on a regular basis. The forums are full of people (including one of my machines) reporting their machines simply blue screen with some fairly high frequency when resuming with these drives. The problems seem to cross vendors too...

Re:standby / wakup issues

[Anonymous Coward]

No, that problem has not been fixed, and won't be fixed until Sandforce learns how to design semiconductors.

The problem is that the controller is latching up, evidenced by the fact that the supply current goes through the roof on resume.

My 16G Toshiba Key says 8MB

[cheekyboy]

'Same bug' happened to my 16G Toshiba - generic - no fancy name. After 3 weeks of 12/24 usage, fail, no data, no files, says 8MB.

POS CRUD

Same Story with me

[abigsmurf]

Had a Vertex 2. After 3 months it died overnight without warning, No PC I hooked it up to would even recognise the drive. Data was totally inaccessible, first time I've ever had this with a drive. Didn't appreciate having to RMA a drive with all my data left on it, appreciated even less that I had to send it recorded delivery to the Netherlands out of my own pocket.

I love how they state on their website that the mean time for failure is something like 130 years.

Re:Same Story with me

[Miamicanes]

> I love how they state on their website that the mean time for failure is something like 130 years.

Oh, that's because the hardware ITSELF will last for 130 years. The half-life of any data you STORE on it is about 7 weeks. If the offices of both Sandforce and OCZ were replaced by smoldering nuclear craters tomorrow morning, I'd smile and say, "at least they can't screw anybody else now" (well, once the remaining inventory stocked by Amazon & Newegg was gone).

besides fake GIG sizes advertized

[cheekyboy]

ie the 120G is really a 115G as reported by windows
formated its down to 106 Gb

yeah 100% false fake advertising.

How can they get away with it, a 120 is really a 115, if its 120 in hardware, but appears as 115, its still 115.

Im sure real 128G SSDs have a few hidden Gig as reserve, but its not advertised as a 135G.

Lucky I only used 60 out of the 120 (115) ((106))

Great, now it's just a commodity

[Anonymous Coward]

One reason to buy Intel SSDs is the Intel controller and more-reliable 34nm Flash.

Now that it's sub-2k PEC 25nm Flash and the same buggy controller running buggy firmware everyone else is using, there is absolutely no reason to pay one red cent more for an Intel SSD.

Numonyx

[unixisc]

Question - Intel had spun of their flash memory unit to merge w/ STM's flash memory unit (Intel's unit had only NOR flash, while STM had both NOR & NAND) into a company called Numonyx. That company was later acquired by Micron, which now makes all types of flash memory.

So I understand Intel making SSD controllers, but Intel makes NAND flash and SSD as well? Those things are commodity to begin w/, even more so than Numonyx's StrataFlash was, so if Intel had spun them off, why are they playing in this market? Sooner or later, the various manufacturers of SSD - be it WD, Seagate, Micron, Samsung, Toshiba, et al would make the price so low that it just doesn't make sense for Intel to compete there. And it's not like Intel's fabs would see more utilization, since they are processor i.e. logic fabs, and less optimized for memory than memory fabs are.

Also, is Intel's NAND MLC NAND, or SLC?