Looking good. Pay attention to non-copressible numbers. Source.
Printable View
Looking good. Pay attention to non-copressible numbers. Source.
Making some very rough/basic estimations from the 4kb reads at QD=3, the access times (QD1) should be less than 0.065ms. Very good, if it is true. I am a little skeptical but we will see.
AFAIK this Toshiba NAND is a lot faster than the Intel NAND...Quote:
The drives are using 32nm Toshiba toggle-mode NAND (effectively DDR NAND), however OCZ will go to market with 25nm Intel NAND when the drive is ready.
Looking forward to this as well, very strong figures from the SF-2xxx controller.
With all these paper-launches I guess Intel has to speed up their 6Gb/s implementation.
As a sidenote, I've started testing using the iometer RC on my SF drives and as expected, random data makes quite an impact on both read and write.
(the RC includes random data generated in 3 different types of data, repeatable, pseudo and full random)
It looks like the Toshiba NAND and controller is pretty fast, Link
close to 100.000 iops (read) if you look at the video in the link.
can you teach me how to switch between these...i never knew that!Quote:
(the RC includes random data generated in 3 different types of data, repeatable, pseudo and full random)
if these SF numbers are real with the incompressible data i can see myself warming up to them quite a bit....
Do you know if any production drives will utilize Toshiba's toggle-mode NAND and controller? Would be nice...Quote:
Originally Posted by Anvil
..must...temper....excitement...wait for G3!
Isnt the toshiba SSD a SAS drive?
for us uber nerds sas is like music to our ears....but too damn expensive:p:
So nothing on the G3 at CES, although the Hitachi/ Intel Enterprise Ultrastar SSD400S was showcased - fibre channel and SAS interfaces, with 535MB/s read, but only 46,000 random 4K read IOPS and 13,000 random 4K.
35PB random writes or 19TB random writes per day.
Both Toshiba and Hitachi/ Intel talk about features to enable data reconstruction in the event of failure.
Looks like this gen of SSD's is going to make big inroads into the Enterprise market.
I'll create a thread for the RC.Quote:
Originally Posted by Computurd
@F@32
The Toshiba and HGST SSDs are not consumer drives, I expect they'll be pricy.
The technology used should rub off on consumer drives at some point in time.
I'm bummed out a bit because it is not SATA3 and can't touch SF-2xxx and new Micron drives in sequential non-compressible data throughput...Quote:
Originally Posted by Computurd
Well 4k random should be better on the intel drives, thats what they do.. but not having sata 6gbs is kinda old as all the new upcomers have Sata 6.. Who knows though they could change alot tho seeming as they werent at CES this year.. :shakes:Quote:
Originally Posted by F@32
The enterprise market is showing the higher speeds because they can display the SF-2500 which several did. Other than OCZ, manufacturers have yet to have completed product to display the SF-2000 and even OCZ will admit that they are still fine tuning their SF products. Corsair will have the new Marvell controlled drive out by Feb but also stated they will not comment with respect to the SF situation which means that they will be hitting with two SATA3 products once SF is completed final testing of the SF-2000.
It was expected to be out already but for a delay with respect to stability in some areas. Credit to SF who won't comment but will not allow its release with this minor deficiency which will be fixed soon enough. March is going to be a HUGE month for solid state drives and we will see alot more than anyone expects!
First time at CES and I have had meetings with every SSD manufacturer. I also got front and center 5ft from Stevie Wonder while he did a duo with EW&F and took the most amazing shots (pics below) . If one hasn't experienced Vegas or CES, it simply can't be described.
What you're hearing from these vendors sounds like code for "SF-2000 isn't performing up to expectations in the lab and everyone's in scratch their heads and wait mode."
Last I heard (about 3 months ago) there were stability problems at max throughput speeds.
From Anandtech:
“The drives won’t see the light of day for months (sometime in Q2) and what OCZ is showing today is very, *very* early silicon and hardware. The drives are using 32nm Toshiba toggle-mode NAND (effectively DDR NAND), however OCZ will go to market with 25nm Intel NAND when the drive is ready.”
New gen NAND, new gen controllers, new problems to resolve.
gonna be a huge jump to get from where they are with incompressible data to where they want to be with the new controllers....hopefully they can get it though...man that would be a helluva feat, but the big question in my mind is how does that affect the latency...not real impressed with current gen SF and the latency.