News - pureSilicon reveals the 1TB Nitro solid-state drive

[HEXUS]

We're expecting SSDs to fall in price throughout 2009, and that'll be joined nicely by a steady rise in capacity.

Read more.

[vinnyT]

With all these Co.s developing the ssd market should become affordable to those of us on the lower rung of the ladder, sometime in 2010 (hopefully).

[Keva161]

I didnt know the advantages of SSD's untill I looked on youtube.. I want one just for installing Windows! Who needs Raptors
Do they wear out over time if you use them to much?

[Philipp]

Want one for my notebook

Wonder how much they will cost, taking into account Intel’s 80Gb SSD (similar speeds) is around £450

[shaithis]

2010 looks like the year I may upgrade my RAID

It's really starting to look like conventional hard drives have a very short life-span left, apart from the optical drive, the last mechanical PC parts days are numbered!

[Fraz]

I didnt know the advantages of SSD's untill I looked on youtube.. I want one just for installing Windows! Who needs Raptors
Do they wear out over time if you use them to much?

Yes and no. You can only write to each flash cell so many times before it becomes non-functional, and this behaviour is quite predictable. Thus, you have some cunning wear-levelling algorithm to evenly distribute the writes over the disk, and the disk keeps track of these and can practically give you a time left until failure.

Also, there are two types of flash: MLC and SLC. MLC flash stores two bits of data per cell, and can be writted to fewer times than SLC flash which stores a single bit per cell. Thus, the enterprise guys are gonna be mainly interested in SLC flash devices, and the consumer more interested in the MLC flash drives. SLC flash also has a higher write speed (even more ridiculously high than MLC devices), which is also good for enterprise, but less useful for your standard desktop user.

[GaryRW]

I thought I read something about the Intel SLC SSD having really bad write rates the second time you write stuff to it. apparently there's an issue that if data has already been written to that cell once, it needs to be cleared before anything can be written again, which drastically slows write rates down.

Sounds plausible, but haven't seen any links to solid benchmarks or tests that show the problem. Anyone else seen anything to back this up?? Seen plenty of bench's that show amazing speeds, and would be weird if this problem hadn't been identified. Maybe it's the wear levelling algorithms - did any of the testing use all 80gb of the drive??

[Sim0n]

Gary,
I would personally say that its abit unrealistic to throw HDTach at SSD's and expect them to give real world performance.
After all, no real world application writes to a hard drive in a similar fashion.

I always take hard drive benchmarks with a pinch of salt. Its abit like taking the lastest Ferrari up to 200MPH on a banked oval testing track. Your never gunna do it yourself.

In all honesty, hard drive benchmarks will probably evolve once SSD become more mainstream, rather than "review filling".

[[GSV]Trig]

This years gonna be the year of the SSD thats for sure, if there hitting 1Tb already then hopefully smaller 32/64gb ones will start turning up for decent money and will start working there way into desktops/laptops of normal users.

[baius]

The fact this is going to take ages to hit the market implies that this was just a "boast". "Look, we can do this."

AKA marketing hype?

When this 1TB SSD turns up, I doubt (with some reservations: 640k of ram, etc) they'll be reasonably priced.

Maybe 2010Q3 we can all buy one?

[TheAnimus]

I thought I read something about the Intel SLC SSD having really bad write rates the second time you write stuff to it. apparently there's an issue that if data has already been written to that cell once, it needs to be cleared before anything can be written again, which drastically slows write rates down.

Sounds plausible, but haven't seen any links to solid benchmarks or tests that show the problem. Anyone else seen anything to back this up?? Seen plenty of bench's that show amazing speeds, and would be weird if this problem hadn't been identified. Maybe it's the wear levelling algorithms - did any of the testing use all 80gb of the drive??

I've not had this problem with the E varient, thats the SLC ones.

[master2k27]

omg nice !

it will cost £££

[Funkstar]

it will cost £££

yes. Just like most other things, this will cost money.

[Fraz]

yes. Just like most other things, this will cost money.

Actually, I've heard that PureSilicon are accepting sexual favours in return for these.

[vinnyT]

Actually, I've heard that PureSilicon are accepting sexual favours in return for these.

Bet I still have to pay.

[Fraz]

Bet I still have to pay.

Yes. You do.