OCZ's Colossus SSD comes out of its shell
By Donald Melanson 
posted Jul 25th 2009 7:26AM
posted Jul 25th 2009 7:26AM
We got a brief glimpse of OCZ's massive new Colossus SSD drive at Computex last month, and gotten the official word on most of the basic specs straight from OCZ itself, but we've only just now gotten a look inside that slightly mysterious-looking black box and, well, there's a whole bunch of SSD chips inside. Available in capacities ranging from 128GB to 1TB, all of the drives boast a tightly packed design that includes a pair of Indilinx controllers and some associated DRAM cache for each, along with a dual-drive, single PCB implementation that makes it all appear as a single drive to the system. Still no word on a release date just yet, unfortunately, but it looks like these'll start at $299 for the 128GB drive.
Jeebus! How much for the 1TB drive I wonder? And how long until these are standard and cost $100 like today's SATA? Yay technology!
SSD's are also sata...
I believe you mean standard platter based drives.
During a fairly recent discussion with an upper level Apple tech rep, I’m convinced the very next MacPro will provision a medium/large capacity SSD as a BTO.
Regards,
I'm not sure saying Apple will have a BTO Mac Pro with an SSD in the future quite qualifies you for Nostradamus status...
Wow, I never saw that coming. I mean, they've offered SSD drives in the laptops for a couple years now, but I NEVER thought they would make their way to the desktops!
/s
it will be a good day when these large capacity SSD's have similar pricing to platter drives.
No doubt that in
No doubt that in 10 years these will be as much of a joke as a 30GB 4.2K drive is today.
Yay Technology!
(what's up with this stoopid commenting system?)
Topic: OCZ Colossus -> SSD
... as soon as you use certain punctuation it gets cut off. don't _EVER_ use "$message = strip_tags($message)" .
use "$message = htmlspecialchars($message)" instead.
the link to hothardware's site is down, not sure why u posted this now, would have been better to wait until their page is back online.
Any hardware reviews?
FLASH chips come with about 3% extra memory just for wear leveling. Now, not all memory will be perfect, so there may be some bad blocks in there - that'll depend on the manufacturer. That extra 3% can't really replace bad blocks due to how it's arranged.
Example data sheet - see page 5: http://www.hynix.com/datasheet/pdf/flash/HY27(U_S)G(08_16)2G2M%20Series(Rev0.7).pdf
Have OCZ given up on honesty or is this really 128GB? Unlike other SSD and hard disk manufacturers OCZ were being honest about the true capacity of their drives, advertising a 60GB drive as a 60GB drive instead of the 64GB other manufacturers would claim it is. I bought an OCZ on principle due to their honesty, however many people would probably believe OCZ's drives were actually smaller than their lying scum competitors. For this reason it looks like OCZ may have had to resort to lying about the capacity of their drives as well.
Something really does need to be done about this since the current situation is absurd and on a 1.5TB drive you're now getting 140GB less than the claimed size of the drive. There's nothing stopping drive manufacturers claiming a drive is 100TB* and then hidden at the bottom of the page *1TB = 10bits. I was hoping OCZ would take a stand for honesty but it looks like doing so never made business sense.
In SSDs there isn't that problem. 128 GB is 128GB (and sometimes more, with some extra hidden around to help with wear leveling and such).
I think some do count the "hidden" space for wear leveling in the advertised specs, so I usually think of a 64gb drive as 60gb.
Those platter drives having less space then advertised; it's the cause of microsoft(or was it ASCII? can't remember) thinking 1024 bits as 1 kb, instead of 1000 bits as 1 kb as most hardware manufactures think of it.
The SSD's? no idea. (wouldn't be surprised if they DID think of it the same way, though)
I dont think the manufactures are lying Its just they are using a different standard. There are two standards for units of digital information storage.
There is the metric standard (also known as the SI prefix) which includes kilobit megabit gigabit and terabit and their respective numerical representation 10^3, 10^6, 10^9, and 10^12 bits.
The other standard, the manufacturers standard, is based on a binary prefix which includes kibibit mebibit gigibit and tebibit and their respective numerical representation 2^10, 2^20, 2^30, and 2^40 bits.
The decimal, metric, SI prefix is the easiest one for consumers who generally are not informed as to what binary is and number systems of a different bases. Representing digital information how its stored in base 2, as a binary prefix makes sense as that is how it is manufactured physically. However this would more than likely confuse most people.
However with that said there is a bit of a disconnect as the drives are manufactured and rated in storage size under the binary prefix but are marketed and sold to consumers using the SI prefix.
Hope that clears it up :)
Forest, that's not Microsoft's fault. Hard drive manufacturers know that computers operate in binary (and thus 1024), yet they still neglect to measure by an actual real-word number.
Kelth, hard disks and SSDs are exclusively used for binary data storage so it is entirely dishonest to use a decimal measure of their capacity. When you say 1GB in relation to a computer product people think of it in binary/computer terms (i.e. 1,073,741,824 bytes) not as 1,000,000,000 bytes. You can see from some of the above comments that many people don't even know why that they are losing 73MB for every gigabyte and when they do know believe that it's Microsoft's fault or it's to do with hidden space rather than deception by hard disk manufacturers.
Using a decimal measure may be technically correct and may follow a standard but it's the same as selling petrol in the UK in US gallons (which are smaller) - it's technically correct but is not the appropriate means of measurement and not the means of measurement the buyer expects, thus it’s deceptive.
@Kelth
Yep, that's it exactly. Good thing you said it, cuz I'm too lazy to type that out. XD
Any news on the release date in the UK?
tomorrow, deuche!
Richard,
No one is lying to you. It has to do with the differences in how the drive manufacturers and the OS makers address the bytes on a drive. The drive makers use the decimal system where 1000 bytes = 1 kilobyte. The OS makers use the Binary system where 1024 bytes = 1 Kilobyte. Poof, space is "lost" in the conversion. You're still getting the same number of bytes in both cases, but when you take those extra 24 bytes in each Kilobyte your Gigabytes appear to be less. This is all stated in fine print on the drive box. Do a little research before you accuse people of lying. This has been going on for years, and don't expect the drive makers to budge, because they're right, and so are the OS makers.
Read for yourself (if you're not too lazy): http://www.lmgtfy.com/?q=byte&l=1
The "metric system" to which Meikat refers was invented by marketeers and not technologists; it exists purely to make storage capacity appear to be larger than it really is, it has nothing whatsoever to do with how data is accessed or stored on the drive.
Funny how this marketing practice is acceptable when selling technology, can you imagine if car manufactures' marketing departments decided to redefine miles (making them shorter, thus car faster) or gallons (making them larger, thus more efficient)?
"The "metric system" is a tool of the devil!"
Funny how this marketing practice is acceptable when selling technology, can you imagine if car manufactures' marketing departments decided to redefine miles (making them shorter, thus car faster) or gallons (making them larger, thus more efficient)?
You mean like how some auto makers report gross, rather than net horsepower? Or how some German manufacturers report DIN, rather than SAE horsepower?
Seriously, the world has real problems, put away the OCD and let this go.
This has nothing to do with the metric system, other than the SI prefixes of orders of magnitude are used in the metric system.
And to the OP, no, they are not both right. By definition, a KILO-byte is a thousand bytes. The official name referring to 1,024 bytes is a KIBI-byte or Kilo Binary byte.
So in fact, the harddrive manufacturers are correct; it is the OS and application designers that continue to use the wrong naming. Of course, this is because the wrong use of kilobyte/megabyte/etc entered into common usage.
Pricing for these drives can be seen here (I love anandtech.com):
http://www.anandtech.com/storage/showdoc.aspx?i=3608
Once a super fast 500GB SSD is available for $2-300 I'll jump on the ssd bandwagon for my PC, for the laptop however, I want one now, even if its slightly slower than my rotational 7200rpm drive.
My 80GB Intel x25-m blows the pants off my Seagate 500GB 7200RPM ... so I'm not sure where you're getting the 'slower' idea. An intel x25 or ocz vertex should easily kill a 10kRPM raptor as long as we're talking application performance (random read/write) as opposed to sustained transfers (moving multiple GB's of data).
Are these those 3.5 inch drives?
Yeah, they're 3.5" drives.
How about a recession antidote? :)
Im personally waiting to upgrade to all ssd in the next 2 years. For me they ar just so much more stable, and are much quicker. I do a lot of video work on the go and i cant afford to lose footage because of the platter needle technology.
You're not gonna lose dat if you RAID or backup properly - you WILL lose it though if you are relying on a single drive, doesn't matter if its platter, SSD, optical, floppy or tape.
Happiness is multiple backups.
Let's say that you have a laptop in RAID 1 (mirror mode). When you are writing to disk, the needles are working simultaneously to produce the mirror images. If you drop your laptop when this is happening, it is likely that both disks will be screwed up costing you $1000 to recover the data. Not to mention running 2 HDD drives in a laptop (to avoid possible data loss) has its other downsides (heat, power consumption, size, etc.).
The 2.5 intel and ocz ssd's operate at similar read speeds, but these Colossus 3.5 ssd have dual controllers and should operate at raid 0 (almost double) speeds, in one drive. Although I'm not sure how they can pull that off with a single SATA 2. SATA 3 aren't ready yet and the SATA 2 witch is supposed to go to 300 MB/s I think may top out at around 260 MB/s or so...
Obvious answer is it WON'T be double the performance with sequential reads. Sequential writes maybe. And random reads/writes should definitely double. So yes the SATA 2 interface will be a bottleneck for certain performance tests, but not for others.
Didn't intel just release new SSD's on the 32nm process that makes 128gb drives $224? Doesn't that make these drives useless?
Intel's drive is 80GB.
To Luffy - Yes intel is releasing their new drives at that price point. Does this make these drives useless? absolutely not. These 3.5" ssds are managed by dual indillux controllers, and two sets of dram cache. The single indillux controller configurations are already as fast as the intel drives so expect speeds SIGNIFICANTLY faster for these 3.5" beasts.
Ok is this really 2 64GB drives in one RAID 0? Or is it 2 128GB drives in a RAID 0? If it's the later it is one hell of a deal at $299 the first one isn't too bad either.
2x 64gb in raid 0 is 128gb. The 256gb drive is $650
Hell, at least its cheaper than 700 freaking dollars..
The OCZ Colossus may not be a good idea, particularly since it is going to be used in a desktop with relatively abundant space and connections.
I guarantee that two OCZ Vertex drives running in RAID0 will be far faster, at least in sequential speeds. The problem is that the Colossus only uses one SATA connection which is limited to just under 300MB/sec in SATA II. Considering the fact that one OCZ Vertex (same controller) can hit over 200MB/sec in seq and 250MB/sec in burst, there is no doubt in my mind that a Colossus will be bottlenecked by the SATA interface. They should really just focus on the PCI express based product for desktops, or just make the 3.5 drives use one controller.