Intel Optane SSDs May Not Be On Shelves Until 2018

Megalith

24-bit/48kHz
Staff member
Joined
Aug 20, 2006
Messages
13,000
There is still plenty of mystery surrounding Optane, like when you’ll be able to buy one or which segment Intel will target first (servers, probably). I wonder if QuantX (Micron’s 3D Xpoint effort) will see further delays as well.

Intel CEO Brian Krzanich said sample Optane products will ship to more testers next year, and that "it’s really a 2018 ramp for that product," according to a transcript of an October earnings call, posted on Seeking Alpha. As production ramps up next year, the cost of SSDs will come down, Krzanich said. Prototype SSDs are already used in Lenovo's cloud servers, where enterprise applications can be remotely tested. It can sometimes take months until hardware is qualified for use as an end product.
 
Well of course they're delayed. They gotta fabricate an overpriced 4-8TB cash grab before their new tech forces further price drops to existing MLC and TLC tech.
 
I don't know if this tech is going to immediately take off. Even with traditional SSD's, performance doesn't seem bottle necked by storage anymore because many companies don't code their software to take advantage of fast storage to begin with. Many build with slow storage in mind and design their apps in ways to avoid storage.

I see this technology taking off in the future where flash and memory are merged into one technology. RAM and storage becoming one, further eliminating complexity and the number of components needed as SoC continues to evolve.
 
I don't know if this tech is going to immediately take off. Even with traditional SSD's, performance doesn't seem bottle necked by storage anymore because many companies don't code their software to take advantage of fast storage to begin with. Many build with slow storage in mind and design their apps in ways to avoid storage.

I see this technology taking off in the future where flash and memory are merged into one technology. RAM and storage becoming one, further eliminating complexity and the number of components needed as SoC continues to evolve.

I can think of a few applications where this is going to matter a lot. But if you're not working with servers servicing hundreds to thousands of users at once or high-end compute, yes it would seem unnecessary. Technology like this makes new software solutions possible that couldn't exist otherwise and I'd wager that most of the best ones will be thought up after the tech is available.
 
This could end up in high end workstations that are IO starved and to simplify OEM construction.
 
I don't know if this tech is going to immediately take off. Even with traditional SSD's, performance doesn't seem bottle necked by storage anymore because many companies don't code their software to take advantage of fast storage to begin with. Many build with slow storage in mind and design their apps in ways to avoid storage.

I see this technology taking off in the future where flash and memory are merged into one technology. RAM and storage becoming one, further eliminating complexity and the number of components needed as SoC continues to evolve.
They're releasing Optane 3D Xpoint in a DDR4 form factor NVDIMM, in addition to the NVMe SSDs. A few months ago, the rumor was that the max capacity would be 512GB for a NVDIMM module with speeds similar to DDR4 (except non-volatile). Kaby Lake was rumored to be the first architecture to support Optane NVDIMM modules and older systems would support the NVMe SSDs. Intel isn't the only one though, many manufacturers have produced NVDIMM flash nands as early as 2014 and have been improving their write endurance. The difference is 3D Xpoint and other 3D nand technologies will have a much higher density.

http://www.tomshardware.com/news/intel-3d-xpoint-kaby-lake,31966.html
http://www.tomshardware.com/news/intel-3d-xpoint-picture-nvdimm,30890.html
 
I said it long ago, datacenters are lined up in a big queue for this. And the end user is by nature the last to get it.
 
I don't know if this tech is going to immediately take off. Even with traditional SSD's, performance doesn't seem bottle necked by storage anymore because many companies don't code their software to take advantage of fast storage to begin with. Many build with slow storage in mind and design their apps in ways to avoid storage.

I see this technology taking off in the future where flash and memory are merged into one technology. RAM and storage becoming one, further eliminating complexity and the number of components needed as SoC continues to evolve.

I work for a medium sized company and we've just moved our SQL workloads onto NVMe solid state storage. It has made a huge difference in performance. Our Devs and DBAs can now restore databases in 5 minutes instead of 30, and reports that used to take 10 hours to run now take only one. Storage performance is a massive deal for enterprise workloads.
 
I work for a medium sized company and we've just moved our SQL workloads onto NVMe solid state storage. It has made a huge difference in performance. Our Devs and DBAs can now restore databases in 5 minutes instead of 30, and reports that used to take 10 hours to run now take only one. Storage performance is a massive deal for enterprise workloads.
I was speaking specifically to general system performance. There's always going to be certain technologies that really push one aspect.
 
is 3D Xpoint going to be implement in M.2 SSD only? Meaning you have to have Windows 10 to recognize 3D Xpoint SSD? Or will Intel makes a version that win 7 can recognize?

Kind of like the 540S series vs. the 600 series INtel SSD
 
is 3D Xpoint going to be implement in M.2 SSD only? Meaning you have to have Windows 10 to recognize 3D Xpoint SSD? Or will Intel makes a version that win 7 can recognize?

Kind of like the 540S series vs. the 600 series INtel SSD
I don't see why any PCI-E interface would be excluded. M.2 is just a different way to connect to the same bus.
 
Back
Top