Last week, Intel officially launched their first Optane product, the SSD DC P4800X enterprise drive. This week, 3D XPoint memory comes to the client and consumer market in the form of the Intel Optane Memory product, a low-capacity M.2 NVMe SSD intended for use as a cache drive for systems using a mechanical hard drive for primary storage.

The Intel Optane Memory SSD uses one or two single-die packages of 3D XPoint non-volatile memory to provide capacities of 16GB or 32GB. The controller gets away with a much smaller package than most SSDs (especially PCIe SSD) since it only supports two PCIe 3.0 lanes and does not have an external DRAM interface. Because only two PCIe lanes are used by the drive, it is keyed to support M.2 type B and M slots. This keying is usually used for M.2 SATA SSDs while M.2 PCIe SSDs typically use only the M key position to support four PCIe lanes. The Optane Memory SSD will not function in a M.2 slot that provides only SATA connectivity. Contrary to some early leaks, the Optane Memory SSD uses the M.2 2280 card size instead of one of the shorter lengths. This makes for one of the least-crowded M.2 PCBs on the market even with all of the components on the top side.

The very low capacity of the Optane Memory drives limits their usability as traditional SSDs. Intel intends for the drive to be used with the caching capabilities of their Rapid Storage Technology drivers. Intel first introduced SSD caching with their Smart Response Technology in 2011. The basics of Optane Memory caching are mostly the same, but under the hood Intel has tweaked the caching algorithms to better suit 3D XPoint memory's performance and flexibility advantages over flash memory. Optane Memory caching is currently only supported on Windows 10 64-bit and only for the boot volume. Booting from a cached volume requires that the chipset's storage controller be in RAID mode rather than AHCI mode so that the cache drive will not be accessible as a standard NVMe drive and is instead remapped to only be accessible to Intel's drivers through the storage controller. This NVMe remapping feature was first added to the Skylake-generation 100-series chipsets, but boot firmware support will only be found on Kaby Lake-generation 200-series motherboards and Intel's drivers are expected to only permit Optane Memory caching with Kaby Lake processors.

Intel Optane Memory Specifications
Capacity 16 GB 32 GB
Form Factor M.2 2280 single-sided
Interface PCIe 3.0 x2 NVMe
Controller Intel unnamed
Memory 128Gb 20nm Intel 3D XPoint
Typical Read Latency 6 µs
Typical Write Latency 16 µs
Random Read (4 KB, QD4) 300k
Random Write (4 KB, QD4) 70k
Sequential Read (QD4) 1200 MB/s
Sequential Write (QD4) 280 MB/s
Endurance 100 GB/day
Power Consumption 3.5 W (active), 0.9-1.2 W (idle)
MSRP $44 $77
Release Date April 24

Intel has published some specifications for the Optane Memory drive's performance on its own. The performance specifications are the same for both capacities, suggesting that the controller has only a single channel interface to the 3D XPoint memory. The read performance is extremely good given the limitation of only one or two memory devices for the controller to work with, but the write throughput is quite limited. Read and write latency are very good thanks to the inherent performance advantage of 3D XPoint memory over flash. Endurance is rated at just 100GB of writes per day, for both 16GB and 32GB models. While this does correspond to 3-6 DWPD and is far higher than consumer-grade flash based SSDs, 3D XPoint memory was supposed to have vastly higher write endurance than flash and neither of the Optane products announced so far is specified for game-changing endurance. Power consumption is rated at 3.5W during active use, so heat shouldn't be a problem, but the idle power of 0.9-1.2W is a bit high for laptop use, especially given that there will also be a hard drive drawing power.

Intel's vision is for Optane Memory-equipped systems to offer a compelling performance advantage over hard drive-only systems for a price well below an all-flash configuration of equal capacity. The 16GB Optane Memory drive will retail for $44 while the 32GB version will be $77. As flash memory has declined in price over the years, it has gotten much easier to purchase SSDs that are large enough for ordinary use: 256GB-class SSDs start at around the same price as the 32GB Optane Memory drive, and 512GB-class drives are about the same as the combination of a 2TB hard drive and the 32GB Optane Memory. The Optane Memory products are squeezing into a relatively small niche for limited budgets that require a lot of storage and want the benefit of solid state performance without paying the full price of a boot SSD. Intel notes that Optane Memory caching can be used in front of hybrid drives and SATA SSDs, but the performance benefit will be smaller and these configurations are not expected to be common or cost effective.

The Optane Memory SSDs are now available for pre-order and are scheduled to ship on April 24. Pre-built systems equipped with Optane Memory should be available around the same time. Enthusiasts with large budgets will want to wait until later this year for Optane SSDs with sufficient capacity to use as primary storage. True DIMM-based 3D XPoint memory products are on the roadmap for next year.

Source: Intel

Comments Locked

127 Comments

View All Comments

  • Gothmoth - Tuesday, March 28, 2017 - link

    ok hybrid HDD´s are not that fast... but they help.

    personally i would not waste a precious M.2 port on a cache device like this.
    maybe when our mainboards have 4 or 6 M.2 ports. :)

    i have one fast 1 TB M.2 SSD for the OS and 5 data disk (HDD) in my system.
    i don´t care much how fast the data disks are. they are for storage only.

    i don´t see this optane SSD creating much interest for most users.
  • Denithor - Tuesday, March 28, 2017 - link

    ...except that this is configured to ONLY cache the boot drive. So, no.
  • IntelUser2000 - Monday, March 27, 2017 - link

    Actually if the performance works well enough I bet its going to be fairly popular.

    No, you may not get to see it. Why? Because you are an "enthusiast".

    Most markets are sold with HDDs. 80% in fact. And industry moves to newest generation fairly quickly. 30-40% of new computers sold are based on the latest generation. 80% of that is 25% to 35%. Meaning 25-35% of the new computers sold will benefit from Optane Memory.

    You'll see Optane Memory pre-configured systems coming fro vendors real soon. The pricing is reasonable too.
  • DanNeely - Monday, March 27, 2017 - link

    Unless it gives a much bigger performance boost than the PR slides/specsheets are implying I don't think we will. Tiny SSD caches with an HDD never really went anywhere because bargain market would rather have just an HDD and a slightly lower pricepoint, the enthusiast market was generally underwhelmed by the overall performance and either didn't bother or manually split media/etc off of a midsized SSD to an HDD. The problem for Optane+SSD is that SSDs are good enough in a way that HDDs weren't; and the price of a 32GB optane is more than enough to get you from a 128 to 256GB SSD and most of the way from 256 to 512GB. For most consumers the bigger SSD will be a better option. The rarity of dual SSD m.2 slots on laptops, combined with the larger 2280 form factor and idle power level also mean it's mostly going to be a desktop part in this version; meaning you've eliminated a big chunk of the market already.
  • CaedenV - Tuesday, March 28, 2017 - link

    Yes, and they use HDDs simply because they are cheap. An OEM bulk 500GB HDD costs them $30 vs a 250GB OEM SSD that would cost them $50. The idea that they are going to 'spring' for an expensive cache layer for an additional $40-70 is ridiculous. I mean, for that price they could just get rid of the HDD entirely and just move to a SSD in the first place! Even a 1TB OEM SSD is only going to cost ~$110.

    But that is the whole point. Cheap laptops are cheap BECAUSE they use cheap parts. PC makers will continue to use HDDs on the low end, and SSDs in the mid to high end because nothing else makes any sense.
  • name99 - Monday, March 27, 2017 - link

    It's more interesting, I think, to consider what Intel is NOT shipping and to ask why.

    As far as I can tell (given the limited tech details Intel has released) what we have is something that's both less and more than Apple's Fusion technology.
    Less than Fusion in that the cache sizes are smaller (32 or 64GB, Apple provides 128GB although you can create your own fusion drive using ay SSD you like --- I created one from a 64GB FW2 external drive fused to the internal HD of an old iMac);
    more than Fusion in that the cache is apparently common to all drives --- which is nice but also constrains how aggressively the caching can perform.

    MORE interesting is that, again as far as I can tell, this caching role ONLY REQUIRES BLOCK WRITES. In other words it takes no advantage of that supposed improvement of 3D XPoint that it has byte read/writes. And the caching offered seems unlikely to provide any real value over similar caching done using an SSD (ie the Apple solution).

    Compare to the product that Intel COULD have shipped...
    Imagine an SSD that's a 3D-Xpoint hybrid. The 3D-XPoint memory is used to maintain the various metadata of the flash (so all the usage stats, the remap tables etc) --- basically uses where the byte granularity would be of substantial value, and would be used ONLY by the SSD controller so would not be relevant to file system and OS.
    This seems like it could be an SSD with a nice performance boost over standard SSDs, not just because the metadata updating is faster, but because it allows for the use of different, more flexible, flash block management algorithms.

    But we are not seeing such a product. Why? Everything Intel has done seems to suggest that SOMETHING about their story is fishy. Either that supposed byte granularity doesn't work (ie so many extra ECC bits are required that large sector granularities are the only feasible architecture) and/or the costs of the memory are so much higher than flash that they couldn't create a hybrid at a viable cost point.

    Either way, I see this not as the next great step in Optane, but as an attempt to try to get something, ANYTHING, to work with the technology, regardless of the fact that what they're selling solves a non-problem.
  • Billy Tallis - Monday, March 27, 2017 - link

    The caching is only of the boot volume, not all drives. But that limitation might be removed in future driver versions. It is a bit of an unfortunate limitation that it's just a storage cache and doesn't also get cleanly used as swap space and for hibernation, but those data blocks can land in the cache if they're detected as hot enough by the cache management algorithms.

    I do think once 3D XPoint is more openly available on the market as a component, we'll see some drives replace their DRAM with 3D XPoint and ditch the supercaps. It's probably a bit too soon to tell if the economics will work out right for that to be a viable product, but there will certainly be a lot of engineers interested in making that kind of prototype. In the short term, such a product is obviously more complicated to design (since 3D XPoint needs wear leveling) and it's no surprise that the first Optane devices to hit the market are simple NVMe SSDs.
  • garygech - Tuesday, March 28, 2017 - link

    Optane has to be integrated in some pattern at the 10 nm node with Cannon Lake and a system on a chip design. The entire goal has to be higher efficiency, lower latency, and more reliability. Intel demonstrated this with graphics on the chip, and this was very well well received in mobile platforms. I think the problem has been yields. The proof of principal is there. The mother board change is there. The software improvements are present. The yields at the 10 nm node have been low. This has been their hurdle, and seems to have delayed Surface Pro 5, which was their goal, a completely integrated Wintel Solution, mobile Surface Pro Power House. We will have to wait 6 months to see if they hit their benchmarks. If they hit 10 nm at high yield by September, my guess, they will be primed for a great Holiday Season. But the key is an integrated platform. Think how integrated the I-phone 7 has become. This is the premier mobile platform in the world right now based on profitability per unit.
  • name99 - Wednesday, March 29, 2017 - link

    While Intel talks about this and muses how best to enforce segmentation to screw enterprise, TSMC is actually DOING it ("intergrated Optane").
    Look at item number 6 on this URL:
    https://www.semiwiki.com/forum/content/6675-top-10...
    and the date (2H17).
    OK, so the storage capacities are presumably NOT at the the sorts of level Intel has in mind. But there's something to be said for actually SHIPPING, and improving as you go. TSMC has established a reputation as quietly shutting up until they are sure they can deliver, while every quarter Intel comes across more as the belligerent drunk, shouting at everyone "You think you can take me? You think you can take me? Let me tell you how awesome I am, and all the things I got planned."

    Would it be crazy to imagine that (in a side deal that's NOT in the options TSMC lists) Apple does something like get 512MB of MRAM or reRAM embedded on 2018 or 2019's A- SoC, and used to cache the FS metadata (catalog trees and suchlike, maybe also the FTL metadata) of the main flash storage? All the time while Intel is still talking about how great Optane is, one day, going to be.
    (Alternatively maybe you add this storage to Apple's custom flash controller, which I assume is on 28nm.)

    It's not clear to me the extent to which is yet possible. Everspin dies look way too large for this, but the MRAM density should (in principle) be more dense than DRAM, so I don't know if that's because they're using WAY older lithography as a way to hold down burn rate as they perfect the technology.
  • BrokenCrayons - Monday, March 27, 2017 - link

    I hope that second generation Optane will offer improved endurance, higher capacities, and lower prices. I'd consider using it as an OS/boot drive and then using a cheaper 3D TLC drive (or MLC depending on capacity needs versusu prices) as a storage disk, but I don't think I'd want to use Optane as cache for a HDD. Benchmarks that show clear performance improvements might change my mind, but I just don't think I need a HDD for bulk storage these days.

Log in

Don't have an account? Sign up now