OCZ Core Series V2 SSD (250Gb)

[aztec]

http://www.ocztechnology.com/products/f ... ii_2_5-ssd

Featuring new architecture, the Core Series SSD V2 will be available in up to a massive 250GB capacity and delivers enhanced speeds of up to 170 MB/s read and 98 MB/s write speeds with an improved seek time of less than 0.2-0.3ms, making the Core V2 significantly faster when it comes to both Read/Write and seek-time performance. The addition of a mini-USB port also makes it possible for customers in the field to update the Core V2’s firmware should new versions become available, to further enhance compatibility or performance with future platforms.

The OCZ Core Series V2 delivers additional performance and reliability of the latest SSDs at a 50% less price per gigabyte than other high speed offerings currently on the market. The Core Series is ideal for an energy-efficient mobile or a RAID configured desktop to provide blazing speeds and advanced access and seek times. Because SSD drives feature no moving parts, the Core Series V2 will keep your laptop or desktop environment cool and quiet and provide a durable alternative and superior shock resistance. With even greater capacities up to 250GB, Core Series V2 now offers sufficient storage for large media files.

Designed for ultimate reliability, Core V2 SSDs have an excellent 1.5 million hour mean time before failure (MTBF) ensuring peace of mind over the long term. All Core Series SSD drives come backed a two year warranty and OCZ’s legendary service and support.

That was a quick launch. Didn't they just release the Core Series a couple of months back?

[Bar81]

Well, after years of buying OCZ products, I'm done. Their lame "technology is moving so fast" excuse doesn't cut it; they used their customers to beta test product (with most places offering no return except for "defect") and then when they realized those same customers would let others know that the product sucked, they go ahead and offer a new product with the specs that should have been on the original product.

As soon as 64GB SLC SSDs from Mtron and Memoright drop in price, I'm on that train. Good job OCZ, you've lost a customer.

[krille]

The Core V1 line will be continued as a slightly cheaper alternative, with the Core V2 line being more of OCZ's premium line.

The plan is to make Core more affordable and Core2 just a little above it. SSD tech is moving at a furious pace, much faster than i have seen anything move for quite a while. The push to make the drives faster is a weekly thing, it really is quite scary at the moment.

Drives arrive with OCZ early next week I hear, in stores immediately after.


USB is for firmware upgrade only ( i will recheck) and the drives are sata

http://www.xtremesystems.org/forums/sho ... ost3219447

[dhanson865]

Wow, so instead of calling it a 256GB product they are reserving 6GB of cells for spare sectors. (edit speculation on other sites is they just are addressing the SI naming convention issues, not a technical addressing spare sectors effort).

The part numbers show the progression of new products. 1C, 2C, 3C...

Old
Core 32GB - OCZSSD2-1C32G
Core 64GB - OCZSSD2-1C64G
Core 128GB - OCZSSD2-1C128G

New
30GB - OCZSSD2-2C30G
60GB - OCZSSD2-2C60G
120GB - OCZSSD2-2C120G
250GB - OCZSSD2-2C250G

Presumably the formatted size of the Core 128 and Core V2 120 are the same. At least it's a more honest name from a marketing standpoint.

Honestly if the price is right and the performance and reliability are there I don't need the 250GB. I'd be happy with 120GB or even 60GB.

[krille]

While I could afford an SSD disk, I don't think it's worth it at this point (especially since I'd want at the very least 120GB for my system).

I remember paying something like $400 USD for a memory card (2GB) for my cellphone a few years ago. Now the same card costs $30 USD. I fear (and hope!) something similar will happen to SSD drives. Much more space (and probably slightly better performance) in the future for much less. I will stick to HDDs till then.

Instead I'm investing in a dedicated file server for my main storage (which I will put in my wall closet).

[aztec]

While I could afford an SSD disk, I don't think it's worth it at this point (especially since I'd want at the very least 120GB for my system).

I remember paying something like $400 USD for a memory card (2GB) for my cellphone a few years ago. Now the same card costs $30 USD. I fear (and hope!) something similar will happen to SSD drives. Much more space (and probably slightly better performance) in the future for much less. I will stick to HDDs till then.

Instead I'm investing in a dedicated file server for my main storage (which I will put in my wall closet).

After debating on this, I think I'm going the same route. 120Gb SSD and a DYI NAS. Still cheaper than a 250Gb SSD at this point in time.

[highlandsun]

I'm thinking of buying one of these 120GB SSDs for my new notebook. But they're not actually available for sale yet anywhere I can find. A couple places are listing them as 1-2 weeks till availability.

[jontseng]

Seems quite a quick replacement cycle. The Core 1 series only hit retail in the UK about two weeks ago. Currently running a Core 1 64 GB to set up my new HTPC. Working absolutely fine so far and (obviously) completely silent.

I've been waiting for an affordable SSD for a while; how I rationalised it was a) I want an SSD big enough to run vista + a couple of apps i.e. 64Gb minimum and b) I want to pay less than UKP200. When the cores pitched up at UKP160-180 I pushed the button.

IMHO there's no point hanging around for cheaper / faster models six months down the line - there will ALWAYS be cheaper/faster/bigger drives further down the line. Just find a capacity which suits your CURRENT needs and if the price is acceptable in absolute terms go for it. There is NO POINT holding out for higher capacity unless you need it TODAY. The reason is in twelve months time you will be able bolt on incremental SSD capacity for a nominal price given how pricing is going.

J

[andyb]

Good job OCZ, you've lost a customer.

I assume then that you gave up being a customer of Western Digital some years ago.!!! They have done the same thing for a very long time, but at least OCZ actually use a different part number unlike WD who change a drive's specs but not the part number.

Anyway, back on topic.

This can only be a good thing, and this is a direct response to the rapidly improving SSD technology, keep up the good work OCZ. I have no doubt I will buy an SSD some time in the next year or 2.


Andy

[Bar81]

That wasn't even a good attempt at trolling. Why would I give up on WD when every product I've bought from them performs as advertised (no hitching, freezing, etc.)

[Turas]

I am looking forward to the new OCZ drive. I have the core1 128GB model now and am pretty happy with it. Although I have now experienced some of the hitching it is tolerable as this is still developing technology. I have been running SSD drives now for over a year. I have the Sandisk SATA 5000, First gen Mtron, MLC SAmsung and now the OCZ core1. I have seen the technology improve in each of the MLC models. The MTron in there was nice but was just too small/expensive for me. I figured Q4 this year to be the prime time for my next upgrade. The Micron/Intel ones will be around along with the new Core ones and I think we will see the same improvements in these new models.

I am happy to see the quick dev cycles as each one gets a lot better.

[highlandsun]

I just saw that the Series 2 drives became available today on a number of online retailers. Just ordered one, 120GB, hopefully it will arrive by tomorrow...

[frostedflakes]

These look really good. Still a bit too expensive for me, but nice to see performance increasing and prices going down.

[andyb]

That wasn't even a good attempt at trolling. Why would I give up on WD when every product I've bought from them performs as advertised (no hitching, freezing, etc.)

You missed my point somewhat. You are pissed off at OCZ for bringing out an even newer SSD a 1-2 months after their last one, yet WD have been changing things behind the scenes without even chaning their part numbers for years.

So why get bitchy about OCZ who have a different model number, a different listed capacity, different performance, and a different price tag, when WD have several times sold different actual products under exactly the same name.

Examples: Different platter count vs density adding up to the same overall capacity. Short stroke platters vs less platters used. Selling for months identical products with and without Fluid Bearings. Selling 5400 rpm drives as 7200rpm drives (although they have changed this now). And most recently selling HDD's with different physical ecasements with the same part numbers.

My point is that you should not be pissed of with OCZ at all, and this only makes you look two-faced stating that OCZ are bad because of this but its OK for WD to have done worse things for years.


Andy

[Bar81]

My point is that you should not be pissed of with OCZ at all, and this only makes you look two-faced stating that OCZ are bad because of this but its OK for WD to have done worse things for years.

Andy

I understood your point the first time and my response stands.

[Turas]

Highland, please respond back with some testing results when your arrives.

[highlandsun]

I spent a few hours copying my laptop's hard drive onto the SSD. I seem to have botched my Windows partition, but my Linux partition is working. For sequential reads, I'm seeing 140MB/sec. For sequential writes, about 60MB/sec.

It's definitely terrible for random writes.

[Turas]

I was hoping the random ones would start to improve a little more. I guess I may wait a tad longer for the Intel or Micron models and see how they are.

[AZBrandon]

It's definitely terrible for random writes.

I don't understand that at all. What's the technical reason for this? I would think that with no moving parts, no need to seek, wait for rotational latency, no mechanics of any kind that SSD would be made in the shade for random read/write performance. Is there any sort of tech doc online that describes the limitations of SSD in this respect?

[highlandsun]

NAND flash can generally be written in sectors of 512 bytes at a time, but if you need to rewrite existing sectors they have to be erased first, and erasing is generally done 2MB at a time. So, if you're only rewriting a region of data that's only part of a 2MB block, the whole thing has to be read, the changes made, and then the whole thing written again.

If you can guarantee that all of your writes are even multiples of 2MB, I think you can get pretty good speed all around.

[frostedflakes]

Exactly, one of the weaknesses of NAND flash is that it has to be erased in large blocks. So let's say you want to write a 20KB file to the disk, the way I understand it, this actually requires writing 2MB of data (erasing the entire block, then writing the 20KB of new data, and rewriting the 1980KB of unchanged data). This reduces write efficiency, resulting in a lower effective write rate.

There is actually a fairly simple solution to this problem, and I have no idea why it hasn't been implemented on the controller level yet. What you do is buffer random write data (on a DRAM chip, for example) until you have 2MB of data, then sequentially write all that data to the flash disk. This can significantly increase random write performance, because you're never doing random writes to the flash, only sequential writes (theoretically, random write speed should be equal to sequential write speed). It also has the added benefit of increasing the life of the flash chips.

[AZBrandon]

There is actually a fairly simple solution to this problem, and I have no idea why it hasn't been implemented on the controller level yet. What you do is buffer random write data (on a DRAM chip, for example) until you have 2MB of data, then sequentially write all that data to the flash disk. This can significantly increase random write performance, because you're never doing random writes to the flash, only sequential writes (theoretically, random write speed should be equal to sequential write speed). It also has the added benefit of increasing the life of the flash chips.

With so many hard drives having 8-16mb buffers to begin with, you would think this would be a very simple and logical change. I get the feeling that with Intel jumping into the SSD fight, they'll be pulling out all the stops to kill all comers in performance and we'll see technology like what you mention very soon.

It seems like SSD when from an interesting sidebar technology to the most exciting nearly-mainstream tech for silent computing in years. Maybe the most exciting advancement in hard drives since.. I dunno, since the introduction of the 3.5" form factor? I get the feeling that 10 years from now, the storage market will have crested 50% of total sales going to SSD rather than rotating discs of any type and 20 years from now, the magnetic hard drive will have gone the way of the floppy disk.

[highlandsun]

Agreed, it's pretty stupid of SSD manufacturers not to include DRAM caches in their products. A cache would also allow the SSDs to have burst transfer rates equal to the interface speed, e.g. 300MB/sec on 3Gbps SATA2.

Given how common 8-16MB caches are in current HDDs, there's really no excuse for it.

[andyb]

With so many hard drives having 8-16mb buffers to begin with, you would think this would be a very simple and logical change. I get the feeling that with Intel jumping into the SSD fight, they'll be pulling out all the stops to kill all comers in performance and we'll see technology like what you mention very soon.

The problem with that is data security, if you have power failure you could loose quite a few 20Kb documents. The idea is sound bar that point. This is really where I hoped that Hybrid drives would win out, you get the high read performance of the Flash (say 64GB, and the high storage capacity of a HDD (say 1TB, you could write all of those tiny little files to the HDD as you go, and then the drive moves stuff about internally (essentially a built in defrag and file organisation) when the drive is idle.

The other option would be to create a SSD with cache, and have a battery attached to it, so in the event of power failure it writes its 8MB of data stored in the cache to the SSD, this would be a slightly costlier item, but so much better in so many ways (mainly random writes).


Andy

[Turas]

Yeah supposedly Intel says their controller does not suffer from the horrible write performance of the previous ssd's. I guess we will have to wait and see on this one. I guess the only way would be to us on of the caching controllers and setup a raid array. If there was no cache then I suppose you could set the stripe size to the saize of the memory cell on the hard drives but I think that would be rather wasteful of space. The other option would be for a manufacturer to come up with a battery backed up controller which they already have and adap the kllogic on the controller to only issue writes when it has the data to do a whole block at at time. One of the problems i see is that I dont know if there is a standard size on all drives. I heard earlier 2MB but I assume that number would grow as drives get bigger or would there just be more 2MB cells in there? I am not aware of any ssd aware controllers at this point.

[AZBrandon]

Considering most SSD's will be going in laptops for the next couple years, they already have a battery backup attached to the drive. I suspect drive makers will be willing to overlook the potential for data loss in some respect due to this.

[highlandsun]

These days, 16MB of DRAM probably runs on less than a milliwatt. A capacitor would be enough to keep the buffer intact for quite a long time. Writing to flash takes a bit more juice, but I bet you could use a cap to handle this as well, so the cache would automatically flush when power was removed from the drive.

[highlandsun]

Yeah supposedly Intel says their controller does not suffer from the horrible write performance of the previous ssd's. I guess we will have to wait and see on this one. I guess the only way would be to us on of the caching controllers and setup a raid array. If there was no cache then I suppose you could set the stripe size to the saize of the memory cell on the hard drives but I think that would be rather wasteful of space. The other option would be for a manufacturer to come up with a battery backed up controller which they already have and adap the kllogic on the controller to only issue writes when it has the data to do a whole block at at time. One of the problems i see is that I dont know if there is a standard size on all drives. I heard earlier 2MB but I assume that number would grow as drives get bigger or would there just be more 2MB cells in there? I am not aware of any ssd aware controllers at this point.

The 2MB block size is actually a property of the NAND flash chips, not the SSD as a whole. If an SSD is using multiple channels to write multiple chips in parallel to achieve high sequential speeds, then its erase constraints will get even worse.

The logic for caching and dealing with erase blocks really belongs on the drives themselves, not in the controller. If you only do caching in the controller, then it's tied up while flushing its cache buffers to the drives at their native write speed. If you do the caching in the drive, the controller can dump data at full interface speed and then go away and do other work.

[Turas]

good point, I would think it is best handled in the drive itself too, but I was just wondering who would come up with something faster.

The post from earlier about the multiple channels actually making write performance worse. I am curious as to why? Wouldn't you be able to be writing to one channel and reading from the others. I would think that could improve some things as long as the data you are reading is not influenced by what you are saving at the same time.