Slashdot Mirror
Samsung Mass Produces Fast 256GB SSDs
Lucas123 writes "Samsung said it's now mass producing a 256GB solid state disk that it says has sequential read/write rates of 220MB/sec and 200/MBsec, respectively. Samsung said it focused on narrowing the disparity of read/write rates on its SSD drive with this model by interleaving NAND flash chips using eight channels, the same way Intel boosts its X25 SSD. The drive doubles the performance of Samsung's previous 64GB and 128GB SSDs. 'The 256GB SSD launches applications 10 times faster than the fastest 7200rpm notebook HDD,' Samsung said in a statement."
My left nut is an unreasonable price.
I don't need no estinkin'
Jeepmeister
Does this mean it's no longer useful to buy 10,000RPM drives? 10x faster? Sweet.
MABASPLOOM!
Damn -- How can I bitch about how expensive it is when they won't even tell me!
Caveat Utilitor
It makes a nice press release. But I like to see a story with a little more meat on the bones.
So it launches applications 10 times faster [sic] (should say in 1/10 the amount of time), but the article only claims speed improvements of about 3.5 to 1. People need to seriously examine how they quote or accept statistics.
Jim Elliott, vice president of memory marketing at Samsung, said the new 256GB drive can store 25 high-definition movies taking up 10GB of space each in just 21 minutes, which he said is a significant advancement over a 7200rpm hard disk drive, which takes about 70 minutes.
signature pending slashdot approval
The spinning disc is slowly beginning to wind its way down.
This guy's the limit!
to put some of these into my servers
ditto, but im waiting for permanent data erasure to become a little more mature. i understand the wear leveling incorporated into SSD can cause current programs to stumble.
On the Oregon Cost born and raised, On the beach is where I spent most of my days
I am thrilled, as a home user I think 250 gigs is the sweet spot for my laptop. While I certainly could fill more than that, I think that mark represents a reasonable amount of space for the average mobile user looking to ditch the problems associated with a spinning platter. I also expect the price to fall quickly, making these drive much more affordable in the near future. SSD is finally getting close to the masses.
You could always use RAM for something that needs high seek speed, frequent access, and constant I/O...
Just sit the device on a massive voltage-regenerating battery-driven power-supply.
The ComputerWorld article says "and are available for resellers today". The Samsung press release says, "announced today that it has begun mass producing". I couldn't find them in any of the usual places.
The Samsung website is particularly un-useful and hard to navigate, though I suppose it's appropriate that they require you to use Flash for this one.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
W O W
I never thought a Single Sided Disc (SSD)
would ever be able to hold so much data . . .
We are Dead Stars looking back Up at the Sky
I know the impetus is to produce big and fast SS drives, but I'm more interested in cheap and fast ones. My desktop machine has 11 Gb of system and apps and <1 Gb of user files. I would be perfectly happy with a 16 Gb SSD that had great performance, was cheap, and was reliable. Reliability is a big issue. Although theoretically a device with no moving parts should always be more reliable than one with moving parts, in reality SSD technology isn't as mature as HD technology, so the failure rate may actually be higher, and there may be no way to recover from a failure.
Find free books.
All these flash drives and solid state drives keep advertising capacities in powers of two: 64 GB, 128 GB, 256 GB. So why do they still say a 256 GB SSD is 256 million bytes?
Oh, say does that Star-Spangled Banner entwine / The myrtle of Venus with Bacchus's vine?
SSDs make use of bulk read/writing to give an average i/o speed that's quite high, however reading and writing hundreds of small files that aren't located sequentially throws a spanner in the works. I'd like to see some real world benchmarks
Me neither. We spent weeks (which translates to tens of thousands of dollars) benchmarking and optimizing a database app. The thought of accelerating it by a factor of 5-10x with a simple hardware upgrade is stunning.
Does a seek-time from one RAM location to the end of the 256GB (or larger drive) as fast as the avg seek time of a 15K drive? And please stop calling them disks! Disks are cicular objects.
TOP DSLR Cameras Reviews of the top DSLRs
ditto, but im waiting for permanent data erasure to become a little more mature. i understand the wear leveling incorporated into SSD can cause current programs to stumble.
The mean time to fail in the new SSD's is a bucket load better that most regular HDD's.
http://www.google.com/search?q=ssd+mean+failure+vs+hdd
Between the 10,000-100,000 writes and the logic used to try not to rewrite the same place over and over they do quite well.
Anyone got a part number on this?
Wonder how many hours this drive would last if used for swap or a database container until the flash cells wear out and start returning errors.
10k*256GB / 200MB/s write speed = 151 days at full write 24/7. And you'll probably get some nice warnings without data loss since the typical failure mode is that they can be read but no longer written. Of course if you're using swap even nearly that much, you're doing it wrong. I'd be very surprised if my swap use exceeded 10GB/day, in which case it'll take me some 700 years to hit the write limit. And if you're running a heavy database there are drives for you, just not this one. So who do you work for? Western Digital? I think they're the only ones that haven't realized the boat is leaving and they're lost in the mountains.
Live today, because you never know what tomorrow brings
Disk I/O is the one area I still have an easy time slamming modern computers on. Most others, it isn't too expensive for me to simply get enough power that handles what I want in realtime without slowdown. Multiple VM, no problem quad cores are cheap. Big audio projects? Hell I can get 4GB of RAM for less than a month's Internet access... However when those projects start hitting the disk, I start having problems, even with a RAID array. The sequential stuff isn't it, it's the random access that kills it.
Audio only takes 172Kbytes per second per track (for 32-bit floating point). So you figure that doing something with, say, 64 tracks isn't a big deal right? Only about 11Mbytes/sec, way under what a single disk can take. However you can find that it'll choke. Reason being is that the audio isn't all nice and sequential. It's written to disk as 32 separate stereo audio files. Also you maybe have some of them reading, some of them writing and so on. The disk gets overloaded trying to seek to the information in time.
VMs are the same thing. Two VMs running computations at the same time on a system works at full speed. They each use a core of the CPU, there's no problem. The do contend for memory bandwidth, but that is plenty high enough. Likewise one VM doing disk access happens at near native speeds. There's not a lot of overhead to read and write to the disk. However get two VMs doing disk access, man things grind to a halt. Your drive is dancing all over trying to service the simultaneous requests from different areas so throughput grinds to a halt.
An SSD would just be amazing for apps like this. Not because it has so much more bandwidth, but because it's bandwidth stays much higher under intense random access. Where a harddrive might obtain 50MB/sec in sequential read, the same drive might struggle to pull even 5MB/sec in random reads. For the SSD it might be more along the lines of 200MB/sec for sequential and 180MB/sec for random. Even though it isn't full speed, it's close enough as no odds. With that, the VM and audio work would have no throughput problems.
Perfect for systems that need to be written to once, then read lots, available with minimal delay (no spin-up) and maximum reliability. ie pr0nz server. Immense sales for this market sector alone should bring prices down.
They whose government reduces their essential liberties for temporary security, receive neither liberty nor security.
Giga is an SI prefix. It is defined as 10^9 and abbreviated as a capital G. So to say you have 200G of something implies you have 200,000,000,000 of them.
Computers do it wrong. When computers say Giga they mean 2^30, not 10^9. That's wrong, for that you use IEC prefix of gibi, abbreviated as Gi.
The reason is that back in the day, computers had little memory. Thousands of bytes was all. So when talking about thousands of bytes, programmers started calling them "kilobytes". After all, it is close. 2^10 is close to 10^3, only 2.4% error. Well memory kept growing, and the incorrect prefix usage kept going on and they kept using bigger ones.
However this has two problems:
1) The error grows. At the giga level it is about 7% off. The large you are talking about, the more the difference between the base 10 prefix and it's "closest" base-2 amount.
2) You get confusion between levels. For example suppose your computer shows you something in megabytes. It says you have a file that is 2000 megabytes. Well that's 2 gigabytes right? Wrong, 2 gigabytes is 2048 megabytes. So it is rather unintuitive to humans. We work in base 10, the numbers displayed are base 10, but the prefixes are used wrong.
Really, the harddrive makers are right. Computers should display amounts according to the base 10 prefixes. Computers have no problems with base conversions, they should be doing that for people.
Well, this year I bought an Acer AspireOne netbook, but I decided for the model with a HDD instead of the SSD version...
However, if SSDs continue improving and becoming cheaper, maybe my next mobile computer will be moving-part free!
No sig for the moment.
One problem is they write in blocks (128kb?) and the entire block needs to be erased to modify a single byte.
That chews those writes faster than you'd expect, and write leveling can only do so much especially if the drive is mostly full.
No fans? Do you want some roasted nuts? :P
Here is a Mil-Spec approved permanent SSD erasure device: http://static.howstuffworks.com/gif/hammer-1.jpg
Wouldn't this make RAM effectively useless?
Don't SS Drives have a negligible performance gain when the I/O is sequential? [or do I have that backwards, I always mess up the most the smallest details] This makes me think a SS Drive would be good for something like a windoze pagefile... but bad for something like reading a really large table that isn't indexed.
The Dell Inspiron Mini 9 has no fans. It's never felt too hot to me.
Stop Global Warming!
Just say no to irreversible processes!
We spent weeks benchmarking and optimizing a database
And just think of all the people who have crappy unoptimized databases who can just buy these and leave their design as-is!
I could be wrong, but you sound like you're being sarcastic, which is a pretty stupid attitude to have here.
Let's say you have a crappy unoptimized database. You can spend tens of thousands of dollars' worth of programmer time to fix it up and optimize it so that it runs fast on your current hardware. Or you can spend perhaps one tenth of the money to upgrade to a super-fast disk, achieving the same end result. Which one is the smarter move?
If you mod me Overrated, you are admitting that you have no penis.
And of course every time someone levels this criticism it completely ignores the complimentary question: how many hours will a mechanical spinning-platter HD last doing all out full-speed sustained reads and writes 24/7?
At least flash drives have a predictable failure timeline, whereas HDs simply have a vague MTBF and could easily fail much sooner (or much later!) than that.
If you mod me Overrated, you are admitting that you have no penis.
Can anybody tell me what data recovery is like on SSDs? Is it advanced as standard harddrives where they can pretty much get at the data even after it was overwritten?
Fixing the database is the right move. If it costs 10's of thousands of programmer dollars to fix it I'd say you have far bigger problems than an unoptimized database. I mean, programmers should only have access to selects updates and deletes. If your programmers are designing your databases, no wonder they're borked.
On topic, I can wait for the SSDs. The price of the 250s, mass produced or no, still presents as a problem for a business case to run out and load up on these. Hardly stocking stuffers.
-- Karma whore? You betcha. --
I recently did some research before buying an SSD and the cheaper SSDs show some odd behaviour, namely, their reads are very sawtoothed (ie not sustained high thruput). Admittedly, the greatest speedup is in random access (no rotation/seek time), but the xfer is important too. I ended up going with a smaller, but very fast 16G mtron 7500 pro, as it's more than enough for system, swap and frequently used (db's etc). Bigger would have been nice (for warcraft and vm's), but it's not great loss. A good ssd for a system drive is a thing of beauty, and can be retained as other hardware is upgraded, of course.
Here is an even more permanent erasure method.
--
"Open source is good." - Steve Jobs
"Open source is evil." - Microsoft
could be wrong, but you sound like you're being sarcastic
I was being 50% sarcastic - some people get lucky and a technological advance allows them to fix poorly performing design for the time being (they then turn into) Other people assume that some advance will come along and then either it doesn't or they can't afford it. Finally, there are people who see that a poorly performing design is going to catch them up and they proactively fix it before anyone loses their job over the order taking system crashing for a couple of days due to heavy load.
Databases are frequently the workhorse backend on which the business runs; deciding whether to spend $$$ now to fix poor initial design is like deciding whether to put safety bladders in the gas tanks of Pintos.
Reads are byte word size
The caveat is that it is necessary to erase the whole block to write a byte when a bit needs to be changed from its erase state. e.g. :
If the erase state is 0 and a bit needs to be cleared (it holds a 1 in or scenario, and we want it to be zero) then it is necessary to erase the whole block. This (obviously) means copying the block contents to RAM, zeroing the FLASH page in which the byte resides, and then writing the page back to FLASH. It sounds worse than it is, and ultimately the overhead doesn't put a dent in the difference between using spinning media and FLASH. For example, what is the overhead to change a single byte on a hard disk?
Update the metadata (itemization of steps not included; you get the idea.)
Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
Who told you this?
Wear leveling is handled at the firmware level, and should be transparent to the application. There's no reason it should cause issues any more than RAID, error correction or sector re-mapping causes issues on current hard drives.
How often do you edit a single byte without updating the rest of the block? How often do you write period? And, what difference does it make weather you burned out a single byte within a block, or the entire block?
Furthermore, wear leveling should keep this from burning out a specific block on the device: every time the block is re-written, it goes on a different part of the disk.
Well, the guy who brought up database optimization said that "we", which implies at least two people, spent "weeks" optimizing a database which was too slow on spinning media. Multiple people with the appropriate expertise spending that much time is going to cost many tens of thousands of dollars.
Perhaps they had "far bigger problems". Didn't sound like it though.
If you mod me Overrated, you are admitting that you have no penis.
SSDs do not allow you to directly read/write/erase flash memory. The firmware includes a flash translation layer that lets the host read/write 512 byte sectors just like any other drive. Sectors do *not* have a fixed location on the disk. Writing a sector simply appends it to the current erase block, and updates the translation table (also an append). When it runs out of blank blocks, it picks one to erase based on its wear leveling algorithm and garbage collection, and copies any live sectors to a fresh erase block. Just like a HDD, there are plenty of spare erase blocks, which are needed for the copying garbage collection and for when erase blocks go bad.
While the basic function of FTL is open, the wear leveling and garbage collection algorithms are fiercely proprietary. (The best ones actually count how many times a block has been erased and keep the counts even - and do this at high data rates.) This is OK for now because there is also fierce competition, and the code runs only in firmware on the device - not on the host. (Same as the controller code on a HDD.) Should the SSD market ever shake out into a monopoly, the basic FTL ideas are available.
Good god, man, only suckers with corporate accounts have to shop at CDW.
Yeah, but when you buy enough from their lame selection of their overpriced stuff, they send you 'free' tins of cookies! We just got our yearly allotment of cookies at work. Yum. Totally worth spending tens of thousands of dollars more than we should so we get 'free' tins of cookies worth probably $15 each. (that's per tin, not per cookie :)
It seems (to me anyway) that people dont think a single byte being defective is a big deal, when in fact its dead blocks which are somewhat larger.
Logs, most databases and many other things have frequent small writes.
It means that the SSD can be fried somewhat faster than people think.
Wear leveling just means it all fails at the same time instead of some bits failing sooner than others.
Not everyone works on simple 5-table MySQL DBs for their online pr0n collection. I know it must be hard to imagine, but some people actually work on multi-million dollar systems which are distributed across multi-node clusters, with TBs of data, etc. Some of those people actually lurk around here too...
Some of us can't wait for SSDs either - like the poster said: to be able to instantly gain 5-10x performance is staggering and worth every cent. When last did computing tech provide that kind of performance jump?
Outside consulting contractor: $120/hr
One week's onsite work: 38 hours.
Mathemagic says that hiring two consultants for more than a week each would be above $10k.
Rampant carbon sequestration destroyed the Dinosaurs' tropical paradise. I'm here to help repair the damage.
Oh, yes, you have the option of spending 100k beforehand and fix that O(e^n) algorithm, or, alternatively, spend some 30k upgrading your disks and waiting a few more months before more data is added and the app crashes again, and, then you spend those 100k.
That is quite a big interest rate here.
Rethinking email
That chews those writes faster than you'd expect, and write leveling can only do so much especially if the drive is mostly full.
Not really. With dynamic wear leveling, you will use the entire disk, not only the empty parts. Some SSD manufacturers (and their customers) have found out the hard way why only using the empty parts is a bad idea...
mov ax, 4c00h
int 21h
I bet if you did a file usage analysis, you would fine the OS files are used 100x more often than your .docs or other work files (avis).
Nothing wrong with 250gig in a laptop for work files, but it cannot hurt to have the primary os on a 16/32gig SSD card.
Any extra LARGE apps, can be install to the HD.
At least the OS can be pretty fast, and your apps wont slow down with OS usage.
It all comes down to cost/usage smarts. Why waste 80% of a SSD being used rarely. Pay peanuts now, upgrade to 4x or 8x in 6months for the same price, dont pay top dollar today, just to be COOL.
Liberty freedom are no1, not dicks in suits.
I hope you're using striping (RAID5 or RAID10) and not doing pure mirroring or concatenation for your RAID arrays. Anyhow, I am really looking forward to SSDs that permit a lot of operations on the same bit. I deal with databases, and they too do a lot of random IO.
Stop the brainwash
If the speed increase due to optimization is the same as the speed increase due to a hardware upgrade, then the benefits are the same....
I don't know why everyone thinks that a hardware upgrade's benefits will be wiped out nearly immediately but carefully optimizing the software will last forever. It just makes no sense!
I can only conclude that Slashdot is home to a cult of optimization worshippers.
If you mod me Overrated, you are admitting that you have no penis.
The fan in my AspireOne usually turns on only when I have been using it for more than an hour in bed
When Im just idly browsing the net in a coffeeshop or some other place it never turns on
No sig for the moment.
Do they seriously have a howstuffworks page for a hammer?
Planning to be moderated ± 1: Bad Pun.
Because too often a slow program is slow because it uses a O(n^8) algorithm where it should be using a O(n) one. So, every doubling at the hardware speed after optimization will lead to a doubling on the amount of data your program can deal with, but every 256 times hardwre spee-up will lead to a doubling on the amount of data the program can handle before optimization.
After some point, even Moore's law get slow.
Rethinking email
Yeah... I was taking that for granted. I've got 4 500GB WD-AAKS', identically split to 9 partitions. The 'outside' (beginning of disk) two are RAID 10, the next two are RAID 1, and the remaining five are RAID 5. The very 'inside' (end of disk) is filled with dd'ed .img that I mount on a loopback to benefit from RAM page caching in the kernel. The raw images are great for virtualBox, since it's over iSCSI I can use the disks with a either Windows or Linux. The RAID 1's hold ISOs for distros, and the RAID 5's are general purpose. The iSCSI server itself is just a high end gaming board with 3 gigs of RAM and a bunch of NICs running BlueWhite64... it's somewhat hacked together, but it works (I'm a student so I had to eat a lot of off-brand Ramen to afford the rig).
At any rate, it can easily handle 5 or 6 VMs beating it down along with serving up regular files to my desktops. But I'm in the opposite side of the spectrum from you; I'm tuned for raw throughput and low IOPS... this setup would fall flat on its face with a database load. Tuning for throughput is much easier than tuning for IOPS. I'm sure you won't be missing spinning disks when you can delegate them to mass storage for backups and run on SSD at a 100th the latency.
If I mod you up, it doesn't necessarily mean I agree with what you've said, sorry.
Drives using interleaved storage methodology have already been available as PCI-e x4 cards for some time now; the FusionIO ioDrive can saturate most PCI busses, and has been available for more than a year. It also blows the doors off of these claimed specs. If I remember correctly, the ioDrive is 64 devices in parallel.
So yeah, way to go playing halfway catchup, Samsung.
StoneCypher is Full of BS
I thought that was only true for 'higher' level partitioning (LVM, etc.). I partition using (c)fdisk mapping cylinder ranges to disk slices and then put the FS right on top of the partition. Do you mean that the hardware doesn't actually correspond cylinders to 'addresses' relative to other cylinder addresses (the 1024 BIOS cylinder limit non withstanding)? I assumed that this was the case from my understanding of the Linux I/O schedulers. This would surprise me, but probably mostly because I've never really considered it before.
Then again, I'm not sure that any mobo chipset/harddrive firmware that I've ever run was open source, so I probably couldn't verify that it's doing what I would assume it's doing.
If I mod you up, it doesn't necessarily mean I agree with what you've said, sorry.
My wife and I have just discovered the joys of digital audio on my/her itouch/iphone. It has not taken us very long to fill a 250 GB drive. People who keep movies must have a real storage problem.
The industry moved from 5" media to 3.5" media, and now 2" hard disks are coming in for more than just laptops. This has mostly come about for the sake of faster seeks, yes?
Given the disparate needs, perhaps the ordinary user needs different kinds of storage. A desktop that is running streaming video doesn't seek much. Do you care if the heads have a 20 ms average access time?
So you guys who have more knowledge about hardware than I do:
How cost effective would it be to create a 12 platter 5" form factor drive that spun at 3600 rpm, but had the same areal bit density as current 2" drives?
***
Part 2. At one point I remember 'tiered' storage. Back when disks were very expensive compared to tape. You had a set of spinning disks, and a silo full of tapes. When a file hadn't been accessed for N days, it was migrated to tape. When it was called for again, you had a long pause while the silo found the tape and loaded the file.
Are we going to have a window where this makes good sense, substituting SSD for disk, and HD for tape?
Third Career: Tree Farmer Second Career: Computer Geek First Career: Teacher, Outdoor Instructor, Photographer.