backtop


Print 20 comment(s) - last by Bladen.. on Oct 23 at 5:40 AM


Prototype SSD using JMF612
Performance questions linger

JMicron is a company that makes a lot of different controller chips, and their products have been used successfully for many years by companies like ASUS to add functionality to their products. However, the company has run into a rough patch due to random read and write latency problems with its SSD controllers.

The company came up with the JMF602B NAND flash controller to overcome some of those problems, but it was just a stopgap solution. They've been working on a new controller chip that would offer similar performance to the highly touted Barefoot controller made by Indilinx, but at a much lower cost. The new controller is also able to use 3xnm NAND flash, thus lowering the cost of SSDs using those smaller chips.

The first SSDs to use the new JMicron JMF612 flash controller were supposed to have been shown off at Computex in June. However, some issues came up that the company has been quietly resolving since then, and it appears that the chip is now in mass production.

The first SSDs to use these chips are from a little known company named Active Media Products. They are paper launching their Predator-X7 series of SSDs featuring the JMF612 along with 128MB of DRAM cache. There aren't a lot of specs available, but the company is promising sequential read speeds of up to 230 MB/s and write speeds of up to 180 MB/s. This falls short of the 250/200 MB/s sequential read/write rates that JMicron provided to DailyTech in May.

The big news though is that SSDs using the new controller are coming, and there will probably be faster implementations at prices that will challenge Intel and OCZ. Lower SSD prices are good for the consumer, but questions still remain about availability and performance.

Active Media is promising availability in November, and that's when we'll get our first reviews.


Part Number

Description

Expected
Street Price

S32G-25PX7

32GB 2.5-inch SATA-II JM612 SSD

$149

S64G-25PX7

64GB 2.5-inch SATA-II JM612 SSD

$219

S128G-25PX7

128GB 2.5-inch SATA-II JM612 SSD

$349

S256G-25PX7

256GB 2.5-inch SATA-II JM612 SSD

$699

 



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Lower cost SSD'd really....?
By Bigjee on 10/20/2009 12:46:11 PM , Rating: 3
If these are going to cost lower then y is it that the expected street prices are higher than what they are for current Barefoot based SSD's.

With all the smaller 32nm Toshiba NAND flash and the supposedly cheaper JM 192 controller I was hoping for a lower price tag.

If the prices are what they are expected to be then I see no reason not to keep buying the Indilinx based SSD's




RE: Lower cost SSD'd really....?
By Bigjee on 10/20/2009 12:47:57 PM , Rating: 2
JM 612 controller*


RE: Lower cost SSD'd really....?
By Pakman333 on 10/20/2009 1:18:20 PM , Rating: 3
Where it says it uses 32nm? This is probably using 43nm flash...

Anyways, who knows what price will be when it launched? Prices already dropped a a lot from May.


RE: Lower cost SSD'd really....?
By fic2 on 10/20/2009 2:18:19 PM , Rating: 3
I think the OP took
quote:
The new controller is also able to use 3xnm NAND flash
to mean this was using the 3xnm NAND flash.

I agree with the OP - I would like to see closer to $1-2/GB.


RE: Lower cost SSD'd really....?
By AskTheChief on 10/20/2009 1:11:24 PM , Rating: 2
I agree, this is not a lower price!


RE: Lower cost SSD'd really....?
By BZDTemp on 10/20/2009 1:11:07 PM , Rating: 2
This is only the beginning. Right now the SSD business is sellers market as all but the "old" and "slow" are sold out times over.

Chances are in a year we shall see bigger, faster, and cheaper units. Just imagine having a 100% silent super fast NAS :-)


RE: Lower cost SSD'd really....?
By jay401 on 10/20/2009 1:57:43 PM , Rating: 2
yeah those prices are a joke if the author is trying to imply they're cheaper. Try $50 cheaper on each of those.


By AnnihilatorX on 10/20/2009 6:15:36 PM , Rating: 2
RRP
aka Ridiculous rip-off price


Jmicron & Indilinx
By Obsoleet on 10/20/2009 2:14:14 PM , Rating: 2
I have a jmicron 602B ssd and a barefoot ssd. The barefoot is so much better it's not comparable. I'd rather have a standard hdd than a jmicron (even this new one) based drive.

It's Indilinx or Intel, depending on what you're doing with your computer. There's no doubt if I was into photoshop or anything using sequential writes I'd accept nothing else but Indilinx. For general use the Intel is great. Both are good at everything though, hard to go wrong.

I still use my jmicron on a daily basis but if I could swap it out for an equivalently priced HDD with no hassle I would in a heartbeat.




RE: Jmicron & Indilinx
By therealnickdanger on 10/20/2009 3:01:31 PM , Rating: 2
quote:
I'd rather have a standard hdd than a jmicron (even this new one) based drive.

Even though you haven't even tested the new one? hmm...

I have a 602B SSD as well (see link below) and it has been working just fine since I bought it in February. It's about 70% full. I did most of the SSD tweaks for Vista, including moving my Internet temp files to my other HDD, so there are few occassions where I notice any small-file-write issues. In fact, they are so far and few between that I wouldn't EVER go back to a standard HDD again. I have been very, VERY happy with the performance of this drive. Still the best $120 I have ever spent on a computer, in terms of performance enhancement.

Of course, in today's market, I could NEVER recommend it compared to the Intel and Indilix SSDs available. They are so remarkably faster/better that it boggles my face. haha

http://www.guru3d.com/article/gskill-ssd-solid-sta...


RE: Jmicron & Indilinx
By Obsoleet on 10/20/2009 4:14:24 PM , Rating: 2
I feel burned enough, that I wouldn't recommend anyone even risk going with this new Jmicron. There's just no reason, it's not cheaper, not faster, and that leaves nothing for an advantage over the Indilinx stuff.

I run an app that creates a 32mb cache out of my RAM to speed up my OCZ Solid Series drive, did all the OCZ tweaks, and it still sucks.
My Vertex is amazing with 1/8th the work. Basically plug and play at full performance, turn off defrag, turn off indexing.. put wiper.exe in your system startup and it's done. That's far less than required to get a Jmicron to even be a tenth as smooth or fast.
I've even updated the firmware on my Jmicron to their latest version, it helped but it still runs like crap.


RE: Jmicron & Indilinx
By sigmatau on 10/21/2009 5:07:44 AM , Rating: 2
Agreed.

Why would anyone buy a Jmicron based SSD? Could anyone give me a reason? Lower speeds, same prices, bad reputation? LOL!


RE: Jmicron & Indilinx
By therealnickdanger on 10/21/2009 8:57:54 AM , Rating: 2
quote:
Why would anyone buy a Jmicron based SSD? Could anyone give me a reason? Lower speeds, same prices, bad reputation? LOL!

The product isn't out yet and in a market as competetive as this, I'm sure they will be priced appropriately given their "lower" speed (which, 230/180 is the roughly the same rating as the non-overclocked Indilix drives and faster writes than Intel). Like the Indilix drives, you may or may not reach that speed or you might be faster, all depends on the rest of your hardware.

All things being equal performance-wise, JMicron will probably STILL have to lower its prices below comparable SSDs almost entirely because of their bad reputation. To be completely fair to JMicron, its controllers were not intended to be used the way SSD MANUFACTURERS used them. These were low-performance controllers being peddled in "high-performance" SSDs. Every SSD manufacturer that used a JMicron controller WITHOUT properly testing it or warning its customers about its poor write performance is infinitely more guilty than JMicron in these scenarios.

With faster flash, faster controllers, and faster interfaces, the future of SSDs is going to kick some serious arse. But you guys are right, JMicron has its work cut out for it in order to be a player. To JMicron's credit, it initially pushed out the dual-controller 602B to help, but then it has been taking its time developing a cache-based controller. Hopefully they get it right, because we really do need more competition to drive prices down.


RE: Jmicron & Indilinx
By leexgx on 10/20/2009 9:17:08 PM , Rating: 2
your lucky then but you did state small write issues still

you have to be mad to buy any thing from JMicron for SSDs until anandtech and 2-3 other sites have had there hands on them, as this SSD will have cache on it problem hopefully be solved


JMicron
By Pessimism on 10/20/2009 1:21:21 PM , Rating: 5
They released glitchy, poor performing SATA controllers a few years ago when onboard SATA was becoming standard, and now they are releasing glitchy, poor performing SSD controllers. Sometimes price isn't everything.




cost or price?
By motigez1 on 10/20/2009 2:14:26 PM , Rating: 2
Do not get confused, regardless of the memory technology inside, 43/32nm, the prices are set according to principal market force, demand/supply, as long as the market of NAND is in over demand, you can forget about lower prices...




RE: cost or price?
By SAnderson on 10/20/2009 3:40:52 PM , Rating: 2
True, NAND prices are going up, even for the same part. Sept/Oct are the busy months for NAND suppliers before the Christmas season. December onward prices will probably go down somewhat or at least level off. Overall $/Gb will still go down even more, but never to the point it will compete with HDDs.


Shame on DT
By Concillian on 10/21/2009 11:47:44 AM , Rating: 2
quote:
There aren't a lot of specs available, but the company is promising sequential read speeds of up to 230 MB/s and write speeds of up to 180 MB/s. This falls short of the 250/200 MB/s sequential read/write rates that JMicron provided to DailyTech in May.


Why does the article mention this? The older Jmicron controllers had fine sequential read / write performance. Where they stank up the room was on random transfers of very small files, where performance was so bad, that the drives would sometimes cause ~half second delays.

The writers and editors letting this through written like this are doing a great disservice to the public by focusing them on specs that are not important to the issue.

It's media coverage like this that caused the initial problem with the Jmicron controllers. Media focused so much on sequential throughputs that Jmicron believed they needed to maximize sequential throughputs at the expense of all else. They did so and in the process trashed the performance of small file random access transfers (the exact kind of things that SSDs should excel at compared to rotating storage). Guess what happened in the real world? Well, we all know the story there... the drives were so bad, they were considered unusable by the majority of users.

Why is DT being part of the problem and not part of the solution with this article? We need the media working to turn around the perception that sequential rates are all that matter, not perpetuating it.




RE: Shame on DT
By Bladen on 10/23/2009 5:40:26 AM , Rating: 2
I think the problem is that there is now easy way to measure small read/write performance, whereas sequential figures smack you in the face.


Savings?
By KoolAidMan1 on 10/20/2009 3:47:26 PM , Rating: 2
The Idilinx based OCZ 256GB SSD I have coming in is also $700, I don't see where the savings come in.




"Intel is investing heavily (think gazillions of dollars and bazillions of engineering man hours) in resources to create an Intel host controllers spec in order to speed time to market of the USB 3.0 technology." -- Intel blogger Nick Knupffer














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki