Linked by Thom Holwerda on Thu 27th Sep 2012 19:36 UTC
Apple I bought a brand new iMac on Tuesday. I'm pretty sure this will come as a surprise to some, so I figured I might as well offer some background information about this choice - maybe it'll help other people who are also pondering what to buy as their next computer.
Thread beginning with comment 536843
To view parent comment, click here.
To read all comments associated with this story, please click here.
christian
Member since:
2005-07-06

I agree, this is the way to go if you happen to have an SSD. SSD for the system, hard drive for data. But I have to admit that I don't trust SSDs for the long term--even when used as a system drive. For example, if you regularly do a fresh install of your OS while keeping your /home directory intact, and you use a distro that has a release cycle of under a year, how many OS installs will you be able to do before bits start to go bad? Never mind the typical package updates throughout the lifespan of a distro; there are usually lots of those.

The fact that these things have to even be considered makes hard drives a more attractive choice in my opinion. Also, if you're a distro hopper you may run into problems sooner. What about swap space... should you put your swap partition on a hard drive, and then suffer the slowdown of using a hard drive anyway (eliminating one of SSD's biggest advantages) as soon you start running out of RAM--all because the SSD cannot be trusted for frequent rewrites?



Do the maths. 3000 p/e cycles per block. So for a 128GB drive with ideal wear levelling, you can write ~3000 * 128GB. If you write 64GB a day (unlikely) you'd get 6000 days worth of writes. That's > 16 years.

My guess is your SSD will outlive your machine, and may well die a http://en.wikipedia.org/wiki/Whisker_(metallurgy) related death before the FLASH dies.

Reply Parent Score: 2

UltraZelda64 Member since:
2006-12-05

Do the maths. 3000 p/e cycles per block. So for a 128GB drive with ideal wear levelling, you can write ~3000 * 128GB. If you write 64GB a day (unlikely) you'd get 6000 days worth of writes. That's > 16 years.

I believe they "did the math" as well as controlled-climate testing when they came up with the conclusion that the Compact Disc would last up to 100 years, did they not? Also, I've had a few hard drives outlive the computer itself, so that's not saying a whole lot about the SSD. Especially when it's all theoretical anyway.

And in the previously mentioned case of swap use on an SSD, that same math will work out of your favor, speeding up the loss of bits on that device. Unless these drives actually swap bits from different partitions across the entire device. Do they?

Once these things have been in decent use in the real world for about 15-20 years and there is real-world evidence to back it up, I'll believe. ;) Of course, by then the newest SSDs of the time will undoubtedly be far more reliable than what there are now.

Reply Parent Score: 2

Alfman Member since:
2011-01-28

christian,

"Do the maths. 3000 p/e cycles per block. So for a 128GB drive with ideal wear levelling, you can write ~3000 * 128GB. If you write 64GB a day (unlikely) you'd get 6000 days worth of writes. That's > 16 years."


I've got a few issues with your calculation. Obviously "ideal wear levelling" doesn't exist generically: what's ideal for one pattern is non-ideal for other patterns. And in fact a 128GB SSD is likely to be comprised of at least 8 NAND chips, which for performance reasons are running in parallel and may not take part in distributed wear leveling. So removing your assumptions might decrease your calculation by at least a factor of 8.

We are also talking about reliability on it's own, but in real devices reliability is one of many conflicting goals: performance, capacity, cost, dimensions. etc. The point I'm trying to make is that it's not safe to make assumptions. Even if the MTBF was 100% accurate, it only describes a curve with a multitude of failure points. Even with 3-5 years MTBF, you can still fail in a few months time. I'm just recommending those with write-heavy data loads take extra precaution against data loss with flash drives.

There was a study someone did correlating the jitter in flash performance to it's remaining data longevity. The older flash cells are, the more time it takes to re-program them. I'll try to find a link to it. It could offer a way of getting feedback about how much life is remaining on one's SSD.

Reply Parent Score: 2

christian Member since:
2005-07-06


I've got a few issues with your calculation. Obviously "ideal wear levelling" doesn't exist generically: what's ideal for one pattern is non-ideal for other patterns. And in fact a 128GB SSD is likely to be comprised of at least 8 NAND chips, which for performance reasons are running in parallel and may not take part in distributed wear leveling. So removing your assumptions might decrease your calculation by at least a factor of 8.


Of course, there's also write amplification as well to factor in.

But given how modern SSDs work (basically a log structured device) write amplification should be quite low (approaching 1) and wear leveling will actually be close to ideal once static wear leveling is employed.


We are also talking about reliability on it's own, but in real devices reliability is one of many conflicting goals: performance, capacity, cost, dimensions. etc. The point I'm trying to make is that it's not safe to make assumptions. Even if the MTBF was 100% accurate, it only describes a curve with a multitude of failure points. Even with 3-5 years MTBF, you can still fail in a few months time. I'm just recommending those with write-heavy data loads take extra precaution against data loss with flash drives.


My point was that for most people, you're unlikely to hit the FLASH p/e limit even with 3000 p/e cycles. Firmware issues are more likely to toast your data than physical FLASH errors, which I admit has been a problem with the early generations of drives. But firmware is getting better and the market more mature.

Reply Parent Score: 1