Intel Officially Kicks Off Optane Launch with SSD DC P4800X

Subject: Storage | March 19, 2017 - 12:21 PM |
Tagged: XPoint, SSD DC P4800X, Optane Memory, Optane, Intel, client, 750GB, 3D XPoint, 375GB, 1.5TB

Intel brought us out to their Folsom campus last week for some in-depth product briefings. Much of our briefing is still under embargo, but the portion that officially lifts this morning is the SSD DC P4800X:

View Full Size

View Full Size

View Full Size

MSRP for the 375GB model is estimated at $1520 ($4/GB), which is rather spendy, but given that the product has shown it can effectively displace RAM in servers, we should be comparing the cost/GB with DRAM and not NAND. It should also be noted this is also nearly half the cost/GB of the X25-M at its launch. Capacities will go all the way up to 1.5TB, and U.2 form factor versions are also on the way.

For those wanting a bit more technical info, the P4800X uses a 7-channel controller, with the 375GB model having 4 dies per channel (28 total). Overprovisioning does not do for Optane what it did for NAND flash, as XPoint can be rewritten at the byte level and does not need to be programmed in (KB) pages and erased in larger (MB) blocks. The only extra space on Optane SSDs is for ECC, firmware, and a small spare area to map out any failed cells.

Those with a keen eye (and calculator) might have noted that the early TBW values only put the P4800X at 30 DWPD for a 3-year period. At the event, Intel confirmed that they anticipate the P4800X to qualify at that same 30 DWPD for a 5-year period by the time volume shipment occurs.

Read on for more about the SSD DC P4800X (and other upcoming products!)

View Full Size

Note that the P4800X is able to exceed 500,000 mixed random IOPS at a Queue Depth of only 12, while Intel's own P3700 can not even reach 100,000 at that same QD.

View Full Size

The predictions I made based on the previous P4800X spec sheet leak were mostly confirmed at the event. Latencies and QoS figures are extremely impressive across the board.

View Full Size

Intel brought up an interesting metric to show differences between Optane and their own flagship NAND SSD. The above test steadily increases what they call 'write pressure', which is simply an increasing random write load. Note what happens to the read latency of the P3700, while the P4800X stubbornly remains close to the axis throughout.

View Full Size

All of this great latency does not do any good unless there is an easy way to put it in a server, and Intel is doing their best to do just that. Intel Memory Drive is a hypervisor-like driver that can map P4800X's (multiple are possible) as if they were an extension to system DRAM.

View Full Size

In the above example, comparing a server with 768GB of DRAM to one with 128GB of DRAM combined with a pair of P4800X's, 80% of the transactions per second were possible (with 1/6th of the DRAM). More impressive was that matrix multiplication of the data saw a 1.1x *increase* in performance. This seems impossible, as Optane is still slower than DRAM, but the key here was that in the case of the DRAM-only configuration, half of the database was hanging off of the 'wrong' CPU (dual Xeon systems have half of the memory connected to each CPU, meaning accessing the other half requires additional hops to get to the CPU doing the work, adding latency to each transaction).

View Full Size

Many high-end enterprise vendors spoke at the event, and to say they are all excited about the possibilities enabled by Optane is an understatement. We will be getting some testing time with the P4800X shortly and will be reporting on our results very soon, but the P4800X is not the only story here!

View Full Size

We were briefed on (and even demoed hands on) Optane Memory at the event, but those NDA’s are not up yet, so I’ll have to direct you to previously published information on that for now.

There is one more thing that I’m allowed to mention here: An Optane Client SSD is coming! I can’t go into details yet, but I can give you a quick history lesson:

Need I say more?

Video News

March 19, 2017 | 01:39 PM - Posted by Moyeni (not verified)

So, Intel SSD 860 ? :p When will we get some info ?

March 20, 2017 | 05:20 AM - Posted by Anonymous (not verified)

Do not bother just buy SLC SSD.
What? Not available anymore? I wonder why? No one needs extremely fast and reliable non-volatile memory?

March 19, 2017 | 01:41 PM - Posted by Moyeni (not verified)

Or 860X ?

March 19, 2017 | 02:57 PM - Posted by Allyn Malventano

Soon, but not soon enough.

March 19, 2017 | 04:47 PM - Posted by Moyeni (not verified)

Awww... still, in H1 right ?

March 19, 2017 | 02:10 PM - Posted by Anonymous (not verified)

They are making that mistake again, the technical press is again not mentioning Micron’s competing brand of QuantX/XPoint. So Queue up the Daft commentards “AMD does not support Optane” or “Intel will not make available Optane for AMD’s benefit”, Etc!

They are even failing to mention any potential Micron/QuantX competition against any Intel/Optane SKUs at The Register, and the commentards are already showing that Daftness knows no boarders! Nothing says: Sponsored Article content masquerading as a “Impartial” article/review like failing to mention any competing product’s existence, even if that competition is still incoming at this point in time.

Do you think that Micron will magically disappear off the face of the earth with Intel remaining the only one getting at that XPoint market all to itself for very much longer! Look at Intel’s pricing on its Optane SKUs and Micron will be very busy working towards the same XPoint ends for any of Micron’s potential QuantX/XPoint customers, AMD/others. So do not think that Intel will refuse AMD any XPoint/Optane offerings and any XPoint profits that go with any XPoint sales!

Folks Micron is the co-creator of XPoint along with Intel, work that out in your heads and come to the obvious conclusion!

March 19, 2017 | 02:48 PM - Posted by Allyn Malventano

There is no mention of Micron / QuantX because they have not briefed us on anything beyond their initial joint XPoint announcement with Intel, and their subsequent branding as QuantX. We have still given them additional press when there was something to show. Micron was early to demonstrate actual performance benefits of XPoint (and they even did so while Intel was fumbling their demos of the same), but what matters now is shipping products, and to date, Micron has not announced any.

I don't go through every company responsible for the development of NAND flash in, say, a Samsung SSD review. That doesn't mean I was paid off by Samsung to write the article. It just means I'm focusing on what is relevant to the discussion. We have yet to see so much as a leaked spec sheet or roadmap of Micron's XPoint-driven offerings, so for all we know they may never have a retail product. I can only speculate if I have something to go on.

Also, what does any of this discussion have to do with AMD? Are they making storage products now? (no, this does not count).

March 19, 2017 | 03:22 PM - Posted by Anonymous (not verified)

Yes but XPoint is a new technology and the technology press is responsible for providing the proper technological background! Peter Bright at Ars Technica mentioned that XPoint was an Intel/Micron creation! There needs to be more history/background on any relatively VERY New and different Technology like XPoint that the reading public is not very familiar with!

There is only one other future competing XPoint product for the foreseeable future and Micron’s existence as a producer of the only other competing product warrants the same sort of focus in and by technology press as the x86 competition between Intel and AMD receives! This is not some generally available IP to an entire market of many producers as say your example of the SSD/NAND market/s many SKUs from many different producers tries to suggest!

It is unacceptable for the Technology Press at this early in the the XPoint introductory phase in and entirely new XPoint IP market for there to be no mention of XPoint’s co-creator and no mention of any future XPoint market competition between Intel and Micron!

It’s not only PCPer’s coverage on the XPoint Technology that warrants a better historical/market perspective with respect to any New Technology in its nascent stage of introduction, as that's the duty of the Technology press as a whole. Micron, in the context of this new XPoint market, is the Only other possessor of the XPoint IP and will be Intel’s only other competitor for quite some time!

March 19, 2017 | 03:34 PM - Posted by Anonymous (not verified)

P.S. With regards to any suspicion in the reading public's mind with regards to any "Sponsored" implications. It’s what is not mentioned that can raise suspicions more than what is said, and leaving out mention of the only other possible XPoint competitor for the forseeable future in the XPoint market for any analysis of Xpoint(QuantX or Optane) invites such suspicion! This holds true across the entire technology press and not just PCPer!

March 21, 2017 | 09:35 AM - Posted by Anonymous (not verified)

Hilarious you call out "commentards" then to go full "commentard" yourself. If Intel does a press briefing and talks about a bunch of products, why would PcPer (or anyone in the tech press) talk about a bunch of speculative products based on the same technology (dur, IMFT is a JV) that haven't even been talked about by their manufacturer? Maybe people who know what they're talking about prefer to talk about what they know rather than a bunch of speculative bullshit? If you want the BS, go to BS websites that trade on BS and let the big boys do their thing.

March 19, 2017 | 03:59 PM - Posted by Allyn Malventano

To put it bluntly, if Micron wants more coverage on QuantX, they are invited to brief the press accordingly. We have done more than our fair share covering them in prior XPoint posts. This post was an Optane post, not an 'every competing XPoint technology' post, nor is it required to be. If folks want to perceive it as sponsored content even though it is not, so be it.

March 19, 2017 | 04:47 PM - Posted by Anonymous (not verified)

And thus the online technology press will always be suspect!Because there is no statment required from Micron for the very existence of any competition in the XPoint marketplace to NOT be journalistically mentioned, and your and your publication's reputation is fruther damaged in the public's eyes!

You job(Technology Press) is as much to properly educate the public on new technology as it is to report on one or the other maker's latest news with respect to that technology!

I'd expect that Intel will be mentioned at every opportunity when Ryzen is discussed, ditto for XPoint and the Intel's and Micron's products befor and after they are to market!

The online press and a large bolder of rock salt is ther rule PCPer or any Other online source!

March 21, 2017 | 09:38 AM - Posted by Anonymous (not verified)

Allyn's job is not to write a comprehensive market review with every new piece of information released. Intel will be first to market. When Micron gets to market, I'm sure they will be compared to intel, but until then... deal with the fact that you don't know shit about what micron is even trying to do with XPoint.

Trying so hard to make everyone out to be a shill... whats the point?

March 21, 2017 | 12:30 PM - Posted by Jeremy Hellstrom

Our job is whatever that commentator tells us it is apparently.  They were just desperate to point out our employment is in jeopardy as we are not fulfilling terms of the contract they created and signed for us.

Or, a different way of responding.

March 19, 2017 | 07:01 PM - Posted by Anonymous (not verified)

What exactly does this "brief the press accordingly" entail?

Does it involve briefcases and brown envelopes, or threats of withholding review samples, or cuts in ad slots purchased.(This is an ONLINE Tecnology Reporting Industry wide and very endemic problem, the FTC/FCC properly regulates the over the air and NON internet cable TV broadcast industry)

The "technology press" is supposed to know the subject matter, and its competative landscape, and the history of the technology! Even More so the technology Press needs to know their readership and try to help make more sense out of any of the provided press release/ad copy than any product's producer is inclined to provide.

I never trust any one sided content without mention of what the competition, if any, is up to making and producing in the near future or even any RTM production to compete in any product offerings.

XPoint is here from Intel and coming from Micron and that's good all around for everyone concerned in the consumer and professional Server/Workstation/HPC markets! for CPUs/SOCs and GPU's(With their own on card NVM stores and without).

The Internet is the wild wild West, wild wild West...

March 19, 2017 | 11:12 PM - Posted by Anonymous (not verified)

Nothing would make me happier than if you painfully die

March 19, 2017 | 11:38 PM - Posted by Anonymous (not verified)

Ha ha ha! People painfully die all the time! And unregulated online news industries with endemic conflicts of interests issues will be regulated just as the over the air/print media ones where in the past! Because some politicians where against the internet from day one until they learned that they could exempt the "New Technology Shiny" from any of those old pesky regulations that came about from the past periods of excesses abusive marketing practices in the over the air Radio/TV and print media of the olden days!

Yes some folks like you get angry when the full glory of their royal scams are revealed from under the artificial civilized guise of this new technology revolution of the past 40 something odd years. This “Clean” technology industry is not oh so much different to the old gilded age unregulated monopoly Dirty Oil/Railroad Trusts of the late 19h century into the early 20th century. You must certainly be in the internet marketing business as you anger shows!

March 20, 2017 | 10:48 AM - Posted by Anonymous (not verified)

It has nothing to do with your ideas. It's your terrible writing. Just nonsense with run-on sentences and regurgitated crap you've been spewing for years. Everyone on the site gets it... you don't like Intel. Who cares anymore. I hope ARM takes over the world so you'll have less shit to complain about with your poorly written essays on each of this site's articles.

March 19, 2017 | 02:28 PM - Posted by Jeppe (not verified)

Is it bootable?

March 19, 2017 | 02:41 PM - Posted by serpico (not verified)

I dont know the answer to that, but just throught I'd point out enterprise probably doesn't care whether or not it's bootable.

March 19, 2017 | 02:50 PM - Posted by Allyn Malventano

Also a good point. The primary enterprise use case for these is not as a boot device.

March 19, 2017 | 02:49 PM - Posted by Allyn Malventano

This was asked specifically during the briefing. The P4800X is a fully NVMe compliant product. If the system can boot from an MVMe device, it should be able to boot from the P4800X.

March 19, 2017 | 05:03 PM - Posted by JohnGR

This looks great. I guess, we the consumers will not really care about using the Optane SSDs as memory, we already have more DRAM than we really need, but the rest of the specs make today typical SSDs look like mechanical hard drives.

March 19, 2017 | 09:17 PM - Posted by Anonymous (not verified)

So much for any DIMM mixed with DRAM or even HBM2/HBM# NVM mixed with DRAM usage with the endurance issues according to S/A! (1). It's great for random I/O/Fine grained above any NAND block not so fine grained access. But I do hope that Both Intel and Micron can get any future endurance levels up there for DIMM/HBM# usage in the future! The S/A article is very detailed but worth the read if you want a more detialed first estimate. I can not wait for the independent reviews on the RTM SKUs! And even Charlie needs to mention Micron, Hell he even has done some Micron related XPoint related reconnaissance articles with some estimates on XPoint/XPoint controller metrics figures.

"Intel officially introduces Xpoint with the DC P4800X SSD
Niche product because of endurance problems

Mar 19, 2017 by Charlie Demerjian"

March 20, 2017 | 05:43 AM - Posted by Allyn Malventano

That SemiAccurate article is disproven by anyone who cares to read the product spec sheet. 30 DWPD for 3 years - expected to be validated to 5 years by volume shipment, is plenty for what this product intends to achieve. Second gen will be in DIMM form factor and accessed via DRAM protocols. The 'detail' in Charlie's article is simply not based on facts - it is based on sensationalism, which in my opinion does nothing useful for the reader but spread bad info.

March 20, 2017 | 02:35 AM - Posted by Dark_wizzie

I'll take 5!

March 20, 2017 | 05:05 AM - Posted by Anonymous (not verified)

"product has shown it can effectively displace RAM in servers"
You are funny. Let me fix it for you:
"product has shown it can effectively displace UNUSED RAM in servers"

March 20, 2017 | 05:36 AM - Posted by Allyn Malventano

You think enterprises spend tens of thousands of dollars on DRAM just to let it sit there unused? You're funny.

March 20, 2017 | 09:22 AM - Posted by Anonymous (not verified)

80 % + 1% times matches RAM does not sound RAM replacement to me. More like: your customer satisfaction rating will drop 19% percent.

Flash is much more dense then RAM so it makes sense. SLC flash (or 3D SLC flash named Optane by marketing) does not anymore.

RAM that wears out within seconds? Come on.

March 20, 2017 | 05:53 AM - Posted by Anonymous (not verified)

Only man-child gamers have too much RAM.

I have a server with maxed out RAM. And I hardly have any users.

March 20, 2017 | 09:51 AM - Posted by Anonymous (not verified)

Please take IT hardware scaling course. Crying here will not help.

March 20, 2017 | 11:30 AM - Posted by Anonymous (not verified)

It cost way to much. 375 gig for only $1520.00 I already bought the 1.2tb 750 ssd from intel. I got lucky and only yes I said only. $850.00. I saved some, but I will never pay $1520.00 for a 375 gig ssd!!

March 21, 2017 | 09:41 AM - Posted by Anonymous (not verified)

Thats not the point of this product. If you don't see the value of it, then you aren't the target market and should move along.

Don't forget, SSD's cost a similar amount per GB on release, and this is an enterprise validated part.

March 20, 2017 | 02:59 PM - Posted by Danny (not verified)

Is Intel Optane DC P4800X SSD designed for enterprise user only? What about for gamers?

March 20, 2017 | 07:16 PM - Posted by Anonymous (not verified)

64GB consumer DDR4 RAM costs ~400USD. Games fit well within 64GB. I just saved you 1000USD.

March 21, 2017 | 12:42 AM - Posted by Anonymous (not verified)

Allyn can dump gallons of water on me
for repeating myself, repeating myself.
Given USB 3.1's intelligent use of a
faster clock and 128b/132b jumbo frames,
here's what I would like Intel to do:
(a) mount Optane on the 2.5" form factor (cuz we like it)
(b) increase the clock to 12G, like current SAS
(c) provide a jumper to bump clock to 16G for PCIe 4.0
(d) support the 128b/130b jumbo frame of PCIe 3.0
(e) start with an Add-In Controller to drive the latter
(f) announce plans to do the same with PCIe 4.0 chipsets
(g) support modern RAID modes with (e) and (f).
OK, Allyn, start emergency spillway flows!
p.s. almost forgot:
(h) above 2.5" Optane SSD supports NVMe and U.2 cabling.
[end of rant]

March 21, 2017 | 12:50 AM - Posted by Anonymous (not verified)

Update: I just saw "U.2 375GB Q2" in a graphic above
under "Form Factors and Capacities". April 1 is Q2, yes?
Or, is that April Fool's Day? :-)

March 21, 2017 | 01:03 AM - Posted by Anonymous (not verified)

Patrick Kennedy answers one of my questions here:

"The first implementations that will be publicly available are the PCIe devices either in AIC or U.2 2.5″ form factors. ...
"The first 3D XPoint device from the company will be the Intel Optane SSD DC4800X and is available starting today [3/19/2017] in the 375GB PCIe AIC form factor."

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

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