Samsung Mass Produces 128GB SSD

Lucas123 writes "Samsung Electronics said today it is now mass-producing solid-state drives with a 128GB capacity, and it will begin production of a 256GB product later this year, ahead of its scheduled 2009 release. Samsung's 128GB and 64GB SSDs are available in 1.8-in. and 2.5-in. Currently, solid state disk costs about $3.45 per gigabyte and spinning disk costs about $0.38 per gig."

Comments

Still no deal

[Daimanta]

And still it is about 10 times more expensive than a hdd. If this doesn't get any cheaper, it won't get any popularity. If a new tech wants to replace an old tech it needs a significant and intrinsic advantage otherwise it will be adopted at a snails pace.

Re:Still no deal

[von_rick]

It wasn't until 2007 that we saw the laptop hard drives hit the 250GB capacity, and they didn't hit the $0.38/GB range until a few months ago. In comparison SSD would be reaching the 250 limit in a much shorter period and as higher capacity drives flood the market, the lower capacity SSD drives would become affordable before you know it.

Re:Still no deal

[tepples]

1) Faster reads

Not necessarily. Sustained read speeds are still faster on (most) spinning disks (vs. most SSDs). They do have orders of magnitude better access time resulting in better random read performance, but that wasn't what you said.

To what extent does a typical desktop work load use random vs. sequential cluster reads, especially when it would matter? Consider for a moment that an SSD controller can stripe data across many flash chips, while a conventional drive can address only one platter at once due to head-to-head alignment limitations.

2) Lower power

Not necessarily.

I read that same Slashdot article from a week ago. I gathered from the comments that the faster random read of SSD caused more transactions to be performed per second, and that shortened the battery life as much as anything else.

I expect that 2-disk setups will become the norm: SSD for the OS, and HDD for data - which is what I've been doing in my own systems for the last 2 years (using CF->IDE converters)

Isn't the OS something that can be read sequentially, if you put the kernel, kernel modules, C library, and services in one big squashfs on the hard disk, like a less-extreme version of Puppy Linux's boot process? Then you get the sequential read speed advantage of platters for stuff that'll become resident in RAM anyway.

Re:Still no deal

[dgatwood]

I gathered from the comments that the faster random read of SSD caused more transactions to be performed per second, and that shortened the battery life as much as anything else.

In general, what eats battery power is writing and erasing flash. If you don't have enough RAM and end up paging to flash, that's going to cost lots of battery life (and SSD lifespan as well). There's also a wide range of power management among flash controllers from those that do little or no power management at all to those that only power up an individual flash part when needed. There's also the problem of computer filesystems being horrible in terms of minimizing writes to the flash. When you have to rewrite an entire 128kB flash block for a 4kB cluster write, you can see why this is inefficient.

More significantly, this means that even small improvements to write caching in the OS can make a huge difference in battery life. I would not be at all surprised if somebody turned around and did the same benchmark on a different OS and finds that the same SSD performs better than the hard drive instead of worse. Indeed, AnandTech did just that on Mac OS X and got very, very different results that showed the SSD providing a significant improvement in battery life.

This is, of course, comparing to drives at the 1.8" size, however. Those same tests with 2.5" drives showed the SSD being slightly worse than the latest hard drives, though still favorable compared to drives from a couple of years ago, and only on the order of 5% worse than the latest drives on average---nowhere near the difference seen in the Tom's Hardware test, and pretty clearly proving wrong what Tom's hardware said about 1.8" SSDs versus 1.8" drives ("As a result, the flash based SSD will lose the power consumption battle against 1.8" mechanical hard drives.").

Thus, the question of SSD power consumption becomes mostly one of how much of the wildly different results is due to better write caching, better hot files clustering, etc. in Mac OS X, and how much of it is due to differences in the workload between the two benchmarks used. Discuss.

Wrong direction

[LordVader717]

We don't need higher capacity. What the market wants is for their 32GB drives to come down in price under the 100$ mark. I'd love to replace the hard drive in my notebook to a flash drive, but if it means splashing out hundreds of dollars for one, when there isn't really that much of a glaring advantage compared to a 30$ hard drive, I have to get back down to earth.

Put this in an iPhone

[Takehiko]

Hopefully Apple will put these in the next round of iPhones. Then I can finally replace my cell and iPod with one device!

Compare to LCDs

[Erioll]

It may be more worth it to compare the adoption of SSDs to how the adoption of LCDs occurred. For quite a long while LCDs were much more expensive than CRTs, with arguably worse performance in some significant areas (response time and color accuracy), but they were THIN, and they were absolutely flat, and they were (generally) lighter.

And now they've taken over, and dirt-cheap LCDs are easily available. So being a much more expensive technology initially is not necessarily a barrier to many consumers who want "the next big thing" because they want the specific advantages.

For myself however, I'm interested to know how they've addressed some of the traditional weaknesses of SSDs, such as number of times you can write to any specific memory element, write speed in general, and lifetime of the memory when no power is applied (this limitation exists for HDDs too in that over time the files will become corrupt (random bit flipping due to the magnetics), but I want to know the numbers for SSDs too).

Perhaps a mix

[PIPBoy3000]

I wonder if we'll see a mix of drives in PCs for different applications, or HDs will end up having a massive SSD cache and information moves from drive to drive as appropriate.

Key read-only OS files would remain on SSD. Bigger files that are rarely used would be on the hard drive. The tricky part would be to minimize the number of times you spin up your hard drive. You could potentially leave it up to the user and have a deliberate mounting process when it's time to do backups or archiving.

A 'disk' is like a 'carriage'

[Ed Avis]

We'll know that the new technology has taken over when people no longer need to refer to it as a solid state 'disk'.