Samsung SSD Update: 48-layer 256Gbit TLC 3D NAND & Three New TLC SSDs Announced
by Billy Tallis on August 13, 2015 8:30 AM ESTWhen Samsung took the stage at the 2015 Flash Memory Summit, they admittedly didn't deliver any bombshell announcements on the scale of the Intel/Micron 3D XPoint surprise, but they still had a lot to talk about.
We knew that Samsung's third generation of V-NAND/3D NAND was on the way with mass production scheduled for the second half of this year. Samsung has now disclosed that mass production is starting this month, and that it's a 48-layer design with a 256Gb TLC being the first die announced. Samsung's current second-generation 3D NAND is a 32-layer design available as 128Gb TLC or 128Gb MLC.
With mass production imminent, Samsung has ensured that neither SK Hynix nor the Toshiba/SanDisk joint venture will be able to leapfrog them with their respective 48-layer 3D NAND designs, both scheduled for mass production starting in 2016.
Samsung says the new 256Gb TLC will have about 30% lower power consumption than an equivalent capacity of their current 128Gb TLC, and a switch to a dual-plane organization ensures that one 256Gb die will perform at least as well as a pair of the current 128Gb dies. Density is improved by about 40% while production costs only increased slightly, so price per GB will be going down. At FMS, Samsung is pushing the idea that their 3D NAND TLC is ready to replace MLC for most uses, and they're optimistic about scaling up their 3D NAND layer count past 100.
New Samsung 48-Layer TLC SSDs | |||
Drive | PM953 | PM1633 | PM1725 |
Form Factor | NVMe over M.2 22110 and 2.5" | 2.5" SAS 12Gb/s | NVMe PCIe HHHL card |
Capacities | 480GB, 960GB, 1.92TB (2.5" only) | 480GB, 960GB, 1.92TB, 3.84TB | 3.2TB, 6.4TB |
Sequential Read | ? | 1,100 MB/s | 5,500 MB/s |
Sequential Write | ? | 1,000 MB/s | 1,800 MB/s |
4kB Random Read IOPS | ? | 160k | 1,000k |
4kB Random Write IOPS | ? | 18k | 120k |
Endurance Rating | ? | ? | 5 DWPD (6.4 TB model) |
Samsung also shared information about three upcoming drives, all using TLC though not necessarily the new 48-layer parts. The PM1633 Enterprise SAS drive was previewed at CES in January and is intended for read-heavy workloads. A follow-on PM1633a model was mentioned to use the new 48-layer TLC to reach 15.36TB capacity, but we don't have any other information about that update. The PM953 is a enterprise NVMe drive in M.2 or 2.5" form factors, and is the counterpart to the MLC-based SM951. Of particular interest, the M.2 version is using the M.2 22110 form factor (22mm x 110mm, the maximum length for M.2), with Samsung using the extra space to implement power loss protection.
Meanwhile the PM1725 is a fast multi-TB PCIe expansion card that Samsung intends to use to challenge the assumptions about what uses TLC is suited for. Relatively speaking it appears to be intended for workloads that aren't very write-heavy, but it still manages 120k IOPS for writes. That just looks small compared to 1M IOPS for reads and a sequential read speed of 5.5GB/s.
All three drives are intended for OEMs, but the PM953 will probably find its way into the retail channel just like the SM951.
Finally, along with Samsung's new 3D NAND appearing in the afformentioned new drives, it will also be appearing in at least one of their existing drives. The 850 EVO, Samsung's current consumer TLC drive, will apparently be getting an update to use the new 48-layer TLC, though it's not clear if this will be new capacities and/or a wholesale NAND update.
61 Comments
View All Comments
Billy Tallis - Thursday, August 13, 2015 - link
The 40% density increase is definitely affected by the fact that the logic and cache portions of the chips aren't growing vertically. The dual-plane addressing adds some complexity but not enough to show up in rough numbers like these.I'm not sure how how yields for 3D compare to planar. Costs are definitely higher due to the more complex fabrication process, but the flash cells are bigger and everything about flash has to be built with fault tolerance/repairability anyways.
jjj - Thursday, August 13, 2015 - link
So you assume the density is for the entire die and not the cells? If that would be the case, given that their 128Gb 32 layers TLC is 69mm2, this one would be 98.57mm2 and bit density at about 2.6Gb per mm2. But if you take AT's estimates of 72% array efficiency and you assume cell size and the logic remain the same size while going bigger and adding 50% more layers , you get to 85.5mm2. If they are talking 40% theoretical increase if the 48 layers die would be same size in mm2 then ,this die would be some 92mm2. So ,unless i messed up some of the math, i'm lost. In any case it's much higher density than what Micron announced and this one fits in microSD cards too,that's a plus. We'll see how Toshiba/Sandisk compares, they got same capacity and number of layers, the process is unclear though.For yield there are speculations but no idea if any are reliable.
Billy Tallis - Thursday, August 13, 2015 - link
Samsung's slides put the 40% as being an increase of capacity per wafer. If the die size or shape changed, then the number of dies per wafer would change and that would throw off the math a bit. But then, the diagrams they had on the slide depicted a significant die shrink, which is obviously wrong, so we should take everything with a grain of salt and hope that we can get a good wafer photo to clear things up.jjj - Thursday, August 13, 2015 - link
In the end the cost of adding layers is the bigger news and encouraging. Nothing is problem free but if the industry manages to scale up at a decent pace , prices should keep the trajectory of healthy declines we are used to.frenchy_2001 - Thursday, August 13, 2015 - link
I don't think anyone is targetting such a number of layers (yet).Road maps talk about 48 (now), 64 (next year), 96, then 128.
So, a linear growth, not geometric.
They are also talking about improvements in process, layer thickness, edges, etching and precision, which should all have positive effects on density and yields.
Companies went on record against scaling in X/Y, planning only to improve the number of layers.
sonicmerlin - Friday, August 14, 2015 - link
That's incredibly lame. Shrinking the node would bring dramatic density increases and drop the cost/GB like a rock.Scootcha - Thursday, August 13, 2015 - link
I care more about the "SSD Update" to fully support the use of Magician with Windows 10, Samsung.Impulses - Thursday, August 13, 2015 - link
Read elsewhere the 2.5" version of the PM953 uses SATA Express, wouldn't that bottleneck it to an extent?extide - Thursday, August 13, 2015 - link
Yes, as thats only 2 lanes, vs M.2 or U.2 being 4 lanes.jay401 - Thursday, August 13, 2015 - link
"Density is improved by about 40% while production costs only increased slightly, so price per GB will be going down.""3.2 TB, 6.4TB"
Excellent. As soon as we have 6TB SSDs that are price competitive with their HDD brethren, I'm all in. Can't wait to get rid of mechanical drives for good! Price just has to be equitable, which shouldn't be hard once these processes have matured a year or so.