FMS 2015: Silicon Motion SM2260 PCIe 3.0 x4 NVMe Controller Spotted

Subject: Storage | August 14, 2015 - 04:44 PM |
Tagged: FMS 2015, silicon motion, SM2260, SM2256, SM2246EN, pcie, NVMe, ssd, controller

We’ve reviewed a few Silicon Motion SSDs in the past (Angelbird | Corsair Force LX | Crucial BX100), and I have always been impressed with their advances in SSD controller technology. Their SM2246EN SATA controller was launched two years ago, and strived to be a very efficient and performant unit. Based on our reviews that turned out to be true, and this allowed Silicon Motion to slide into the void left by SandForce, who repeatedly delayed their newer developments and forced the many companies who were sourcing their parts to look elsewhere.

View Full Size

The many SSDs using Silicon Motion’s SM2246EN controller.

Silicon motion pushed this further with their SM2256, which we first saw at the 2014 Flash Memory Summit and later saw driving SLC/TLC hybrid flash at this past Consumer Electronics Show. While the SM2256 makes its way into more and more products, I was glad to see an important addition to their lineup at this year’s FMS:

View Full Size

Finally we see Silicon Motion doing a PCIe controller! This is the SM2260, seen here in the M.2 form factor…

View Full Size

…and here in SATA Express. While the latter will likely not be as popular due to the more limited PCIe lanes present in SATA Express, I’m sure we will see this controller appearing in many PCIe devices very soon. The stated performance figures may be a bit shy of currently comparing SSDs like the Intel SSD 750 and Samsung SM951, but with the recent introduction of Z170 motherboards and RST PCIe RAID, it is now easier to RAID a smaller capacity pair of these devices, increasing the performance of slower units. Further, the point of the SM2260 is likely to get a low cost NVMe PCIe SSD controller into the hands of SSD makers, which can only mean good things for those looking to make the move away from SATA.

I’ve included Silicon Motion’s FMS press blast after the break.

Silicon Motion Demonstrates its SSD Controller Leadership at 2015 Flash Memory Summit

Milpitas, Calif., August 10, 2015 – Silicon Motion Technology Corporation (NasdaqGS: SIMO) (“Silicon Motion”), a global leader in designing and marketing NAND flash controllers for solid-state storage devices, today announced it will be presenting at the upcoming Flash Memory Summit in Santa Clara, CA from August 11th to 13th and will be showcasing an array of SSD solutions at booth #313, including the award winning SM2246EN SATA 6Gb/s SSD solutions, the new products using SM2256, the world’s first turnkey SSD controller supporting TLC NAND and the new SM2260 PCIe 3.0 SSD controller.

Products showcasing at Silicon Motion’s Flash Memory Summit booth include:

PCIe 3.0 NVMe 1.2 SSD Controller- SM2260

The SM2260 is the first turnkey merchant PCI Express (PCIe) 3.0 Non-Volatile Memory Express (NVMe) 1.2 solid-state drive (SSD) controller solution supporting MLC, TLC and 3D NAND.  The SM2260 delivers PCIe-level performance with low power consumption and features four 8 gigabit-per-second (Gbps) lanes (PCIe Gen 3.0 x4) of simultaneous data flow coupled with eight NAND channels. It is designed with fourth-generation NANDXtend™ technology, enabling the SM2260 to support 1y/1z nm TLC NAND as well as the upcoming 3D NAND technologies from all major NAND suppliers.  The SM2260 solution optimizes hardware and software to take full advantage of the NVMe 1.2 SSD standard, offering high performance and reliability while delivering more cost-effective PCIe SSDs for computing systems, ultra-portables and tablets.

TLC SSD Controller- SM2256

The SM2256 is the world’s first and only turnkey merchant SATA 6Gb/s SSD controller solution supporting 1y/1z nm TLC NAND and the upcoming 3D NAND from all major NAND suppliers.  Leveraging Silicon Motion’s proprietary NANDXtend™ error-correcting code (ECC) technology (which triples the P/E cycles for TLC NAND), the SM2256 delivers unparalleled, no-compromise performance, endurance and reliability for cost-effective, TLC-based SSDs.

MLC SSD Controllers- SM2246EN and SM2246XT

The SM2246EN is an ultra high-performance and low power SATA 6Gb/s SSD controller supporting 1y/1z nm MLC NAND from all major NAND suppliers.  The SM2246EN's performance and power consumption specifications place it squarely among the top SATA SSD controllers on the market.  The SM2246XT is a DRAM-less SATA 6Gb/s SSD controller ideally suited for cost-effective, small form factor and low power client SSDs and industrial storage solutions.  Combined, the SM2246EN and SM2246XT controllers power more than 70 SSDs available in the market today.

Industrial Grade Single Package SSD and eMMC Solutions- FerriSSD® and Ferri-eMMC®

The FerriSSD® and Ferri-eMMC® are optimally designed for a wide range of embedded applications, such as server, thin client, POS, office automation, HMI devices, factory automation, gaming, medical equipment, surveillance and in-vehicle infotainment systems.  FerriSSD and Ferri-eMMC come with advanced firmware features including error correction, bad block management and NAND health monitoring, while supporting customized firmware, enabling highly reliable and differentiated storage solutions.

In addition, Silicon Motion representatives will present at the following forums at Flash Memory Summit:

Forum C-12: SSD Technology (Tuesday, August 11, 3:15pm-5:45pm): Stanley Huang, Director of Product Marketing of Silicon Motion, will be presenting “DRAM-Less SSD Facilitates HDD Replacement” as part of Forum C-12.

Forum E-22: Controllers and Flash Technology Part 2: Decoding the Future (Wednesday, August 12, 3:10pm to 5:45pm): Jeff Yang, Principal Engineer of Silicon Motion’s Algorithm & Technology team, will be speaking about “New ECC/DSP Solution Helps Migration to 3D NAND Era” as part of Forum E-22.

Session 204-B: Embedded Applications, Part 2 (Wednesday, August 12, 4:30pm-5:45pm): Jason Chien, Product Marketing Manager of Silicon Motion, will be presenting “Tailor-Made SSDs to Excel in Specified Storage Fields” as part of Session 204-B.

For more information on Silicon Motion, please go to http://www.siliconmotion.com.

About Silicon Motion:

We are a fabless semiconductor company that designs, develops and markets high performance, low-power semiconductor solutions to OEMs and other customers in the mobile storage and mobile communications markets. For the mobile storage market, our key products are microcontrollers used in solid state storage devices such as SSDs, eMMCs and other embedded flash applications, as well as removable storage products. For the mobile communications market, our key products are handset transceivers and mobile TV IC solutions.  Our products are widely used in smartphones, tablets, and industrial and commercial applications.  For further information on Silicon Motion, visit www.siliconmotion.com.


August 14, 2015 | 09:54 PM - Posted by MRFS (not verified)

Thanks for the updates, Allyn.

MRFS

August 14, 2015 | 10:41 PM - Posted by MRFS (not verified)

I may be dreaming (for now), but I do look forward to:

(a) PCIe 4.0 with 16GHz clock rates and 128b/130b jumbo frames;

(b) NVMe RAID controllers with x16 edge connectors and 32GB/s raw bandwidth; and,

(c) 2 or 4 U.2 ports on such a RAID controller connecting 2 to 4 x 2.5" NVMe SSDs.

If modern chipsets can feed 3 and 4 high-speed video cards,
we should be able to feed one such RAID controller
with the required number of PCIe 4.0 lanes, even if
a PLX-type bridge chip is required on-card, in the interim.

Then, the wiring topology will be identical to what
the industry has already been providing to workstations
using x8 PCIe RAID controllers and SFF-8087 "fan-out" cables.

Ideally, upcoming motherboards will support multiple
integrated U.2 ports, replacing SATA ports 1-for-1,
similar to the disappearance of PATA/IDE ribbon cables.

One of the main reasons for my "dreaming" above is the
extraordinary maturity that comes with integrated chipset
support for a full range of modern RAID levels e.g. RAID-5 and RAID-6.

PCIe SSDs do suffer from a single point of failure,
a weakness that is not being discussed at industry websites
I visit regularly.

p.s. THEN, we can start talking about variable channel
bandwidth, as a way of "pushing" Non-Volatile DRAM
installed in 2.5" form factors, possibly even using
fibre optical channels instead of copper wires:

http://supremelaw.org/patents/BayRAMFive/Metal.Mock-up.Assembly.Sequence...

The latter invention will benefit a lot:

(a) from Non-Volatile SO-DIMMS that are plug-compatible ("PnP") with DDR3-1333 and DDR3-1600 JEDEC standards, and

(b) from replacing the SATA cables with much faster fiber optical channels.

Here's a paper I wrote re: overclocking data storage subsystems -- demonstrating one approach to variable channel bandwidth:

http://supremelaw.org/patents/BayRAMFive/overclocking.storage.subsystems...

August 15, 2015 | 01:16 AM - Posted by Hakuren

Ahh still only x4. C'mon guys give us the 'Shiny Bitz', the full x16 length cards.

Competition is good for end user, but let's not get in the same trap again with all NVMe drives. Nearly identical (x4) with about the same performance like the AHCI versions are (500-550 R/W) no matter what the sticker.

What I would love to see is just plain and simple RAID AIC with 4 ports (or 8 when PCI-Ex 4.0 is out) and some nice cache as a support. Yes these things are mega fast, but you can always benefit from more speed.

August 15, 2015 | 11:49 AM - Posted by MRFS (not verified)

> just plain and simple RAID AIC with 4 ports (or 8 when PCI-Ex 4.0 is out)

DITTO THAT!

Here are some modest calculations from the college of hard knocks: please check my numbers and RSVP if you find any error(s):

PCIe 3.0 = 08GHz clock
PCIe 4.0 = 16GHz clock

128b/130b jumbo frame = 130 bits / 16 bytes = 8.125

08G / 8.125 bits per byte = 984.6 MB/s per lane

16G / 8.125 bits per byte = 1,969.2 MB/s per lane

1,969.2 x 16-lane edge connector = 31,508 MB/s per AIC

16-lane edge connector / 4 lanes per SSD = 4 SSDs per AIC

31,508 MB/s / 4 SSDs = 7,877 MB/s MAX HEADROOM per NVMe SSD

I predict that PCIe 4.0 chipsets will have sufficient bandwidth for 4 x U.2 -and- 8 x U.2 ports per AIC.

RSVP if you can confirm any error(s) above:
I would like to refine these numbers for
general distribution.

THANKS!

MRFS

August 16, 2015 | 12:32 PM - Posted by fredey (not verified)

maybe take into account 10 to 15 percent overhead. I am still hopeful somebody will find ways to reduce it but it still has to be considered

August 15, 2015 | 01:43 PM - Posted by godrilla (not verified)

I wonder if Zen will have anything competitive, a good cpu goes so far and then there is storage.

Fyi the 750 series ssd 400 gig is on sale with code for $339 until 8/20 at newegg.

August 16, 2015 | 12:41 PM - Posted by fredey (not verified)

I still don't understand the lack of interest in sata express I understand it is not ideal as a primary drive with u.2 ahead but it will in the foreseeable future as a secondary drive I imagine single digit TB u.2 primary drive and maybe 2 drives in raid 1 multi digit TB sata e drives.

August 16, 2015 | 06:33 PM - Posted by Anonymous (not verified)

SATA express seems like a kludge to me. It is an awkward looking cable since it is made up of essentially 3 separate cables. Once they came out with m.2 (x4 speed) and actual PCIe cards, I was already under the impression that SATA Express would be little used. The u.2 connector doesn't seem like that good of an idea either. I think a lot of people will still go with PCIe cards if they have the space and m.2 if they don't. The u.2 connector has a somewhat ridiculous number of pins for all of the different signals it attempts to support. It may be cheaper to do it this way since you can route sata, usb, or PCIe through it without bridge chips at either end. Thunderbolt is essentially the opposite since you have one type of signal routed through the cable, but you need a thunderbolt bridge at each end to deliver PCIe or usb. I think I would rather that they make a compact cable which carries PCIe only. This would reduce confusion.

August 16, 2015 | 06:33 PM - Posted by Anonymous (not verified)

SATA express seems like a kludge to me. It is an awkward looking cable since it is made up of essentially 3 separate cables. Once they came out with m.2 (x4 speed) and actual PCIe cards, I was already under the impression that SATA Express would be little used. The u.2 connector doesn't seem like that good of an idea either. I think a lot of people will still go with PCIe cards if they have the space and m.2 if they don't. The u.2 connector has a somewhat ridiculous number of pins for all of the different signals it attempts to support. It may be cheaper to do it this way since you can route sata, usb, or PCIe through it without bridge chips at either end. Thunderbolt is essentially the opposite since you have one type of signal routed through the cable, but you need a thunderbolt bridge at each end to deliver PCIe or usb. I think I would rather that they make a compact cable which carries PCIe only. This would reduce confusion.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

CAPTCHA
This question is for testing whether you are a human visitor and to prevent automated spam submissions.