backtop


Print 32 comment(s) - last by fteoath64.. on Feb 9 at 8:15 AM

DailyTech chats with JMicron about its controllers used in the majority of today's mainstream SSDs

AnandTech’s Anand Lal Shimpi investigated the performance of Intel's X25-M MLC (Multi Level Cell) SSD, and compared it to OCZ Technology's first generation Core SSD, which used Samsung NAND flash along with a JMicron controller. He found that random write performance was abysmal, due to extremely high write latencies.  This was a problem attributed to the JMicron controller, which was problematic since many other SSD manufacturers used JMicron's controller as well.

JMicron has not been too happy about the negative buzz surrounding its controller around the internet. They have been working on the problem but it is hard to change perceptions once first impressions have been made. In the following interview we asked a few questions and gave JMicron the opportunity to tell their side of the story.

Tell us a little bit about JMicron.

JMicron was founded in 2001 and our headquarters are located in Hsinchu Science Park, Taiwan. As a fabless IC design house, JMicron focuses on high speed serial link technology such as Serial ATA, PCI-Express, USB, RAID and Storage applications. Our products are widely adopted by major motherboard and notebook vendors such as ASUS, Gigabyte, ACER, HP, Dell, Toshiba, Lenovo, etc. JMicron is also the first fabless company in Taiwan that can mass produce SATA II application products.


When did you first find out about the write latency issue?

We have been developing SSD technology since 2006 and launched our first generation SSD controller, JMF601A/602A at the end of 2007. It soon attracted the attention of SSD makers because of the feature set and high performance. We found the write latency issue around March, 2008. The issue only happens under a special condition, when the system data is close to full and the host keeps writing data on it. It takes time to do internal garbage collection, data merge and housekeeping.


What did you do to solve it?

We revised the hardware architecture and launched JMF601B/602B in June 2008.  JMF601A/602A was the old version after B version was available. Currently, all JMicron customers are using latest version, including ASUS NB/EeePC, OCZ, Super Talent, Transcend, etc. The B version improves the write latency a lot. Besides, JMicron also can reserve more spare blocks to alleviate the issue. Because more spare blocks reservation would decrease the drive capacity, most SSD makers tend to not enlarge the spare size.

Note by author: This is part of the reason why OCZ Technology's drives are labeled as 30, 60, 120, and 250 GB instead of the regular 32, 64, 128, 256 GB. Almost all SSDs make use of spare blocks; it is not a feature specific to JMicron.

It should also be noted that AnandTech's testing used OCZ's Core V1 , the Core V2 was meant to address deficiencies and integrate some improvements.

OCZ created a new design that uses up to 64MB of cache to eliminate the write latency issue in their Vertex series of SSDs.

What is the current status of JMicron's controllers?

The JMF601/602 is designed for netbooks and portable applications. They are not so good for servers and  heavy access loading (for example, multi-task access at the same time). We think that's why most users have good performance but some don't. We strove to solve the write latency issue after the AnandTech article was published. And we made some progress in the new firmware versions.

Note by author: Each SSD vendor has the ability to use JMicron's own firmware, or to use their own version. The firmware used can make a big difference. More on this in a future article.

What do you have planned for the future?

Some customers have introduced high speed SSDs with JMicron's RAID controller JMB390, plus two JMF602B controllers. The target performance is 233MB/sec on sequential read and 166MB/sec on sequential write. Moving forward, JMicron is developing SSD controllers with DRAM cache and it is expected to be available in Q3 2009. That will totally solve the random read/write performance issue.

DailyTech will present highlights from JMicron's roadmap in a future article.



Comments     Threshold


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

So the real question is...
By djc208 on 1/22/2009 7:38:23 AM , Rating: 5
When will Anand pick up one of the Rev B SSDs and verify these claims?




RE: So the real question is...
By therealnickdanger on 1/22/2009 8:00:23 AM , Rating: 5
+1

Yeah, I'm inclined to give JM the benefit of the doubt - no company expecting to survive willingly makes poor-performing products. Unfortunately, the market is flooded with SSDs right now with so many internal configuations that it makes it hard to trust anything but the Intel X25-M.

Without a full, mega-review of just about every model out there by AT, I'm hesitant to spend my money on anything else.


RE: So the real question is...
By Gideonic on 1/22/2009 8:20:35 AM , Rating: 5
an interesting review can be found here:

http://www.bit-tech.net/hardware/storage/2009/01/2...

Concerning the following part of the interview more precisely:

Some customers have introduced high speed SSDs with JMicron's RAID controller JMB390, plus two JMF602B controllers. The target performance is 233MB/sec on sequential read and 166MB/sec on sequential write.

The RAID controller really seems to works , reducing the latency and adding write-speed. The drive seems pretty close to X25-M, costs about the same and has 3x more storage space.

I'd really like to see Anandtech review of that drive.


RE: So the real question is...
By therealnickdanger on 1/22/2009 9:21:17 AM , Rating: 2
Thanks Gideonic. Geez, at this pace, the selection in 6 months will be ridiculous!

If the FM-25S2S-128GBT1 uses the same method as the FM-25S2S-256GBT1 reviewed in your link, I may just grab two for a RAID-0... :D

Here's a review of the 128GB Titan, which appears to be the same controller/method as the 256GB version:

http://www.legionhardware.com/document.php?id=804&...

Write speeds are certainly better than the Intel drive, but the lack of on-board cache definitely hurts the G.Skill drives with small read/writes. No stuttering though, which is great!


RE: So the real question is...
By Gideonic on 1/22/2009 10:06:41 AM , Rating: 3
Good find, really seems to be a good deal. Wouldn't hurry buying it though, as according to some rumors (http://www.engadget.com/2009/01/21/intel-readying-... ) bigger Intel drives are just around the corner with price-drops (hopefully).

Seeing how fast SDDs evolve and that they are still a bit too pricey, it seems to be a smart move to wait for the time being.


By therealnickdanger on 1/22/2009 2:31:42 PM , Rating: 2
Haha, if there's one thing I've never been called, it's "smart". The example I always give are the two 36GB Raptors I bought for ~$500 (combined) several years ago. What I can get for my dollar today absolutely crushes what I got 4 years ago. IMO, it's every bit as smart to buy now. ;-)


RE: So the real question is...
By Jansen (blog) on 1/22/2009 8:56:27 AM , Rating: 2
RE: So the real question is...
By mindless1 on 1/22/2009 11:18:26 AM , Rating: 4
On the contrary, they knew the performance parameters of their chips, decided to release them anyway instead of accepting they had overlooked a crucial aspect of performance, and it stung a lot of early adopters.

It would be ridiculous to think they had not found the issues during in-house testing, when even casual use uncovers them. Their sole claim to fame was being earlier to market (by not spending enough time to develop the tech properly) at a good price point.

In other words, screw the customer, so long as we can make a buck. Nevermind the senseless waste on everyone's part we are just too vain to accept we didn't do enough work to have a good product and will let people pay a second time for what they should've had the first time.

That goes for their PATA chips too, this is a recurring theme with JMicron.


RE: So the real question is...
By Kary on 1/22/2009 12:33:25 PM , Rating: 2
of course, and all of the makers of SSD's looked at the JMicron and said "Man, this is the crappiest choice out there, lets go with it."

(end sarcasm)

Even if everyone knew it had a problem, from the number of companies using the design I still have to think it was the best option available at the time.


RE: So the real question is...
By JKflipflop98 on 1/22/2009 12:57:49 PM , Rating: 5
You just said what the poster you replied to said. Except you weren't nearly as eloquent.

Jmicron knew they had the only game in town. They knew it had issues, but since they were the only ones around with a "functional" part, they went ahead and shipped, then collected the cash.

This is a pretty stupid way to do business, because everyone you just screwed over is going to remember that going forward. Jmicron, in my mind, is now a 2nd rate supplier of parts. Ya, they work - but they kind of suck. Just like AMD. Ya, a Phenom works, and it's alright for the money. But what everyone really wants is an i7.

Corporate greed kills another one. Say hello to Creative for us when you get to Hell.


RE: So the real question is...
By leexgx on 1/22/2009 10:19:01 PM , Rating: 2
never really liked jmicro before SSD came any way speed sucked compared to native Sata (more of that was due to been linked to the PCI bus not PCI-e 1x lane)

most of these performance problems are Very apparent when using first gen SSD jmicron, problem should of not taken an year to notices an problem as steep as this one the Write latencys was (V2 fixed) so high any sort of multi writes would kill any benefit of the SSD (some was as bad as 1 second ! delay but most doing 250ms delay) its that slow Vista would not install on some of the SSD drives come up with Delayed write fails on install lol (if you own an intel chip set and have 2 SSD in raid make sure delayed write is turned on useing the intel matrix program should improve performace)

but must admit the hack thay used 2 jmicron chips basically raid to sides of the SSD to remove the problem works very fast as well as lots of storage (250gb) and priced at the same as an intel x35-m (guess on model there) was very price vs storage + speed


RE: So the real question is...
By mindless1 on 1/23/2009 11:27:52 PM , Rating: 2
If you consider AMD 2nd tier, JMicron would have to be 3rd or 4th tier. With AMD you at least get adequate value, the performance you expect that meets the basic need of the part. JMicron could more similarly be equated to Via, netbooks aside since finally something that needs to be cheap, and therefore we don't mind if it's slow and limited, found a niche.


RE: So the real question is...
By ipay on 1/22/2009 8:50:42 PM , Rating: 2
fool me once shame on you, fool me twice shame on me.. u see a pattern??


RE: So the real question is...
By RU482 on 1/22/2009 4:04:25 PM , Rating: 2
yes, please review the OCZ rev2, and Vertex. Hell, review drives with a Samsung controller as well. End user reviews can only be trusted so much


I'm not buying it
By dj LiTh on 1/22/2009 7:35:01 AM , Rating: 2
Just nitpicking but: Change 256MB to 256GB

quote:
Note by author: This is part of the reason why OCZ Technology's drives are labeled as 30, 60, 120, and 250 MB instead of the regular 32, 64, 128, 256MB. Almost all SSDs make use of spare blocks; it is not a feature specific to JMicron.


Fool me once, shame on you, fool me twice shame on me. I myself will be steering away from all their future controllers until there's several generations of their controller between me and their random write problem.




RE: I'm not buying it
By StevoLincolnite on 1/22/2009 7:39:18 AM , Rating: 3
The drives are perfectly fine in something like a Netbook, but in the desktop world, I'll choose something else first as well for a few generations.


RE: I'm not buying it
By paydirt on 1/22/2009 9:24:01 AM , Rating: 1
dj, what you are experiencing is the "snake bit" emotional inefficiency. In essense, you are "selling" JMicron or SSDs at a low point because you got snake bit by a bad experience.

You were an early adopter knowing that it may not work out and it didn't work out and now the new drives greatly mitigate the problem, but now you are "snake bit" and don't want to play anymore.

I'm not saying you are "bad and wrong" just pointing it out.


RE: I'm not buying it
By mindless1 on 1/22/2009 11:22:34 AM , Rating: 2
It's not "snake bit", it's "don't want to give any more money to a company that screws people in order to have more profits".

Obviously we will all buy SSDs in the future, at least anyone who lives a few years longer, but as always it is reasonable to voice your opinion of how a company treats it's customers by spending your money elsewhere.


RE: I'm not buying it
By dj LiTh on 1/22/2009 11:26:20 AM , Rating: 2
That'd be true if i acually had boughten a SSD with the JMicron controller. I'm among the attitude that its too new a technology and people are paying for the R&D of it rather than the acual worth of the parts/labor. Just because i havent been burned by them doesnt mean i cant take note of others who have been and learn from their experience.


RE: I'm not buying it
By surt on 1/22/2009 3:41:17 PM , Rating: 2
It's an emotional strategy, not an emotional inefficiency. Of course, calling it an inefficiency is an emotional inefficiency.


RE: I'm not buying it
By foolsgambit11 on 1/22/2009 6:35:41 PM , Rating: 2
emotional strategy? Have you ever been in PR? Because you're great at creating oxymorons.

What the OP was referring to is an economic phenomenon which creates an inefficiency in the capitalist system. Capitalism functions well when each individual does what's in their best interest (or so the theory goes). When a person doesn't do what's in their best interest, e.g. doesn't buy the product that best fits their needs, then that makes the system less efficient. When the reason for that choice is not based on fact, it's an emotional decision. Ergo, an emotional inefficiency.

Granted, where to draw the line between emotional inefficiencies and simple consumer trust can be difficult. But I'd argue that this case falls closer to illogical rejection. The argument against JMicron is that they produced a bad product, knew about it, but could release it and make big bucks because they were the only game in town. Well, JMicron has made many products, and the vast majority are good. They are no longer the only game in town. Thus, it stands to reason that their next product must be of high quality to compete, and they can't afford to release an inferior product because then it would look like a pattern of failure. So we can expect that their new SSD controllers will perform well, and we can be sure before purchase thanks to benchmarking. So why avoid them?


RE: I'm not buying it
By surt on 1/22/2009 8:22:13 PM , Rating: 2
It's a problem of short term vs long term optimization. You call it an inefficiency, because you're focused on the short term. But in the long term this behavior is a net win for all players in the system.


RE: I'm not buying it
By piroroadkill on 1/22/2009 9:10:26 AM , Rating: 1
I'll nitpick more.

I thought that OCZ's reasons for reporting the capacity as less was because of the disparity between binary and decimal, gigabytes vs gibibytes, and never heard anything about extra blocks being reserved.

I'd like that be clarified..


RE: I'm not buying it
By Brandon Hill (blog) on 1/22/2009 9:19:27 AM , Rating: 4
Here's your clarification:

quote:
*Consumers may see a discrepancy between reported capacity and actual capacity; the storage industry standard is to display capacity in decimal. However, the operating system usually calculates capacity in binary format, causing traditional HDD and SSD to show a lower capacity in Windows. In the case of SSDs, some of the capacity is reserved for formatting and redundancy for wear leveling. These reserved areas on an SSD may occupy up to 5% of the drive’s storage capacity. On the Core V2 Series the new naming convention reflects this and the 30 is equivalent to 32GB, the 60 is equivalent to the 64GB and so on.


http://www.ocztechnology.com/products/flash_drives...


RE: I'm not buying it
By JKflipflop98 on 1/22/2009 9:30:51 PM , Rating: 2
This is a "twofer". One, it is true that blocks are reserved for write leveling and page failure. Secondly, this is a chance for the industry to realign something that has never made a lick of sense into something you can understand. No more "Help! I can only use 109 gigs of my new 120 gig drive!" posts on forums around the world.

Win-win both ways if you ask me.


I haven't had any problems with core
By KITH on 1/22/2009 2:49:03 PM , Rating: 2
using some of the fixes on ocz's forums, such as moving regular writes to another drive, I have had no problems with the core v1 as my main drive.

I understand that others have had problems but my point is that the drives do have a use. Some users might have problems and be unhappy with them.

Personally, I am happy with the drive's performance. It loads apps very quickly and I haven't had any issues.

I've also used them for business use hosting quickbooks files. Makes a huge difference for load times and responsiveness.

I suspect the more vocal people bad-mouthing the drives have never tried them much like those bad-mouthing vista.




By KITH on 1/22/2009 2:58:15 PM , Rating: 2
something I forgot...

These drives at least for the lower capacity i.e. 30-32GB units were very affordable for small amounts of data that needed to be accessed quickly. They do admirably in that capacity.

There is no way I would be using SSDs without these more affordable drives available. They opened up SSDs to the masses. Were they perfect drives, no. But I feel they were well worth the price.

I can understand people being upset over purchasing larger drives where prices might have been more comparable to better performing devices.


RE: I haven't had any problems with core
By Brandon Hill (blog) on 1/22/2009 3:03:47 PM , Rating: 2
I guess the problem is, people expect the drives to just work. I have an SLC-based 64GB SSD in my Lenovo X300. It's my primary drive and I perform all of my work on that drive. I don't have to offload regular writes to another drive or anything like that -- and I shouldn't have to either. I didn't need to do anything to the drive, I just clean installed Vista Business and it just "works". No stuttering, no nonsense.

I wouldn't expect an MLC drive to be any different. It should work as specified with no tweaks/hacks performed by the end-user to get it to work properly. Intel can make MLC perform consistently without hacks -- I expect others to do the same.


By kensiko on 1/22/2009 7:10:00 PM , Rating: 2
It's only a matter of time.

And at the speed it's going, this time will be short. We already see the Samsung MLC SSD being very performant for a MLC SSD.


BTW, thanks for doing this, DT!
By therealnickdanger on 1/22/2009 8:03:26 AM , Rating: 2
It's sadly a bit late in the game for JMicron to get the word out, but I'm glad that there is some clarity now.




By Jansen (blog) on 1/22/2009 8:58:59 AM , Rating: 3
JMicron makes a lot of products for different markets, sometimes to save costs SSD makers will run chips out of spec and overclock them.

It is up to the SSD makers to verify their design, the controller is only one part of the design.

The roadmap article will have some interesting stuff...


By fteoath64 on 2/9/2009 8:15:21 AM , Rating: 2
Jmicro needs to redeem themselves and create fast and reliable products. At this stage, they must be able to quality test some of the SSDs that uses their chips becuase he SSD maker can "damage" the reputation of Jmicro again, even if it were not their controller's (new design) fault.

I want to see Mtron going for gold and cream the rest of the current runners.


"If you can find a PS3 anywhere in North America that's been on shelves for more than five minutes, I'll give you 1,200 bucks for it." -- SCEA President Jack Tretton

Related Articles
OCZ Vertex SSDs Delayed
January 13, 2009, 9:43 AM



Latest Headlines
4/21/2014 Hardware Reviews
April 21, 2014, 12:46 PM
4/16/2014 Hardware Reviews
April 16, 2014, 9:01 AM
4/15/2014 Hardware Reviews
April 15, 2014, 11:30 AM
4/11/2014 Hardware Reviews
April 11, 2014, 11:03 AM










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