New Laptop

[Armando]

Re: New Laptop

I don't even think you know how the hard drives are rated. They are rated at "best guess" ... because no manufacturer would be that stupid to actually test something for years before selling a product.

Obviously. Still irrelevant, because:
* If the manufacturer is wrong and their products last less than they predicted, they're the ones that lose money when tons of units fail within warranty.
* The main contents of the SSD should be OS and software until capacity increases.
* You have a responsibility to back up your stuff no matter what. Even if a drive that could function forever existed, what if it gets stolen?

[Aeni]

Re: New Laptop

Obviously. Still irrelevant, because:
* If the manufacturer is wrong and their products last less than they predicted, they're the ones that lose money when tons of units fail within warranty.
* The main contents of the SSD should be OS and software until capacity increases.
* You have a responsibility to back up your stuff no matter what. Even if a drive that could function forever existed, what if it gets stolen?

I'm not disagreeing with you here. However, you need to balance off a discussion with good and bad. To just ignore a downside because "it's trivial" is doing a disservice to someone you're giving advice. What happens if poor Mezlo ends up getting 3 straight bad SSDs? What is he to think? As long as someone is well aware of the pros and cons, then they can make an informed choice.

To play devil's advocate, I know of several people that have the opposite experience with an SSD. Again, your mileage will vary, depending on circumstances, what you intend to do with the PC, and what your expectations are. Having a fast 15s bootup is nice ... but what's the point of it if you're needing large capacity and/or turning it into a part-time media center?

[Raydeus]

Re: New Laptop

/facepalm back to you for poor reading comprehension.

Logic for you:

1 - defrag is not recommended because as you say
2 - this is bad because of the nature of these cells
3 - continual flash/reflash of said cells wear it out faster

/facepalm

Ah, forget it.

[Mezlo]

Re: New Laptop

There is limited benefit to reading data sequentially (beyond typical FS block sizes, say 4KB), making fragmentation negligible for SSDs.[75] Defragmentation would cause wear by making additional writes of the NAND flash cells, which have a limited cycle life.[76][77]

[Raydeus]

Re: New Laptop

Then again, all storage media has a limited cycle, even *gasp* HDDs!

The thing is Aeni keeps trying to make it sound like your SSD will fail after just a few hundred writes, which is ridiculous. Even temp files aren't overwritten that much in a consumer/enthusiast PC and Aeni even goes as far as assuming the SSD will always write the same file in the exact same cells every single time, completely ignoring all the optimization tools SSDs and OS with SSD support have.

We are talking about at least 10000 writes of the same cell in SSDs from 5 years ago, which even then was difficult to reach on testing (much less in normal use). Check out this article on the subject from 2009. Moderns SSDs are made to last 1 million+ writes, which, I don't know how are you exactly going to reach even if you (for some unknown reason) wanted to defrag a few times a day.

So yeah... I mean it is still more expensive than an HDD for the storage capacity, but cell wear due to writing only applies to server use (thousands of times more intensive than any regular PC would ever see) nowadays.

[Aeni]

Re: New Laptop

Moderns SSDs are made to last 1 million+ writes

Bullshit. Cite me your sources. According to the article regarding the new process for future NAND cells, current best is roughly 10K. This is why smart phone batteries (YES, NAND cells are used there, too) do not have a long life. This is why the new tech is very promising as it will increase integrity by roughly 5-folds and longevity by multiple factors. You don't have to look far than the 3DS ... the battery guarantee Nintendo puts out is only 300 charges. There's a reason why many warranties are all over the place on this ... the tech just isn't as good as you're making it out to be.

[Armando]

Re: New Laptop

However, you need to balance off a discussion with good and bad. To just ignore a downside because "it's trivial" is doing a disservice to someone you're giving advice.

You're not balancing off anything. You're just saying "Don't use it."

What happens if poor Mezlo ends up getting 3 straight bad SSDs? What is he to think?

Sure. What if he's struck by lightning on his way to Best Buy? Is he to think he shouldn't buy the laptop?

I own two SSDs. I have two friends that have SSDs as well. I believe Raydeus and DakAttack also have an SSD. Either the odds of getting 3 bad SSDs in a row are stupidly low, or it's likely and everyone on this forum is stupidly lucky. Which sounds more plausible?

Having a fast 15s bootup is nice ... but what's the point of it if you're needing large capacity and/or turning it into a part-time media center?

We already know that's not what Mezlo's gonna do with it.

There is limited benefit to reading data sequentially (beyond typical FS block sizes, say 4KB), making fragmentation negligible for SSDs.[75] Defragmentation would cause wear by making additional writes of the NAND flash cells, which have a limited cycle life.[76][77]

Ray knows defragmenting is bad. But the real reason not to defragment has more to do with it being a complete waste of time than any long-term wear on the SSD. Defragmenting moves data around on the disk; each part of the SSD will get written to at most twice in the process. Anyone who thinks a defrag will cause significant long-term damage must also be terrified of hitting the Save button on Microsoft Word.

Bullshit. Cite me your sources.

Look at the Tom's Hardware article I posted on the previous page. Page 7 shows data from a data center using 5000 SSDs. They had an average failure rate of 2% in 2 years of use. "Based on the company's usage patterns, we know that none of the failures have to do with write exhaustion."

Page two of the same article also points out that SSD return rates for the first year aren't significantly different than HDD return rates, although that doesn't really say much about failure rates.

This is why smart phone batteries (YES, NAND cells are used there, too) do not have a long life.

I'd like your sources on this. Batteries have circuitry because they can't be charged/discharged naively; full discharge kills it, overcharging could cause an explosion. Show me evidence that what wears out is the circuitry and not the battery itself.

[Aeni]

Re: New Laptop

An article has been published in the December 2012 edition of IEEE Spectrum that highlights an interesting and potentially useful discovery by ROM manufacturer Macronix. Researchers there have discovered that applying heat to NAND flash cells can drastically extend their life, thus overcoming one of the biggest problems with the solid state storage technology.

NAND flash is used everywhere, from smartphones to SSDs to thumb drives, and we've written extensively before on how it works. The technology's biggest failing is that NAND flash only lives so long. Every time the flash cells are erased, they retain some residual charge; eventually, they get to where it takes so long to make them change their charge level that they stop being useful as a storage medium.

As NAND flash grows denser, it gets more delicate; in our discussion on the future of flash, we discuss the decreasing lifetimes of NAND flash and the potential alternatives. SSDs rely on complex mathematical gymnastics at the controller level to reduce writes and hence lengthen the life of their flash cells, but the need for those kinds of workarounds could be substantially lessened by the Macronix discovery.

I seriously recommend you read the article from Ars I posted in the other thread as well as all other related articles to the discussion of Solid State AND new nanoarchitecture based technology (that I can't help ... but great resources on Google)

I'd like your sources on this. Batteries have circuitry because they can't be charged/discharged naively; full discharge kills it, overcharging could cause an explosion. Show me evidence that what wears out is the circuitry and not the battery itself.

Nanoarchitectures for lithium-ion batteries - Wikipedia, the free encyclopedia

^ Read up on Solid State Batteries ... I'm talking about the same process, not the exact same thing from your flash memory.

[Firewind]

Re: New Laptop

Well thanks for the help guys! I just bought the screen shot on the first page from best buy, the Asus with the i7, and the GTX660M graphics card, and 17.5" screen. I'll wait for pricing to come down on the solid state before investing, plus as time passes it'll probably have more "techy shit" too. I was going to put on the things that make you smile thread, but since i started this thread... Best buy has a no interest option for financing, and the guy helping me was super cool and extended the no interest terms from 18 months to 24 months, so instead of dropping $1400, I'm just paying $50 a month for a couple years. That's some good shit. Now i can get back to participating in ffxiv without my screen dumping on me...

Sounds really nice. My PC is getting on a bit and the CPU's age is starting to bottleneck me, especially on SWTOR which is very demanding on your processor, but it should be okay for another 6 months (which should also give me some time to save for a computer upgrade).