QLC NAND Flash Based Intel SSD 660p Could Lower Prices of PCIe x4 NVMe SSDs

[btarunr]

Intel debuted its 3D QLC NAND flash memory on new SSD DC series 2.5-inch U.2 PCIe drives. Its technology partner Micron, too gave its 3D QLC an enterprise debut with the 5120 ION. The first client-segment debut from the IMFlash combine could be the Intel SSD 660p series of M.2 NVMe SSDs. Slotted between the 700p and the 600p, the new 660p implements homebrew 64-layer QLC NAND flash memory, and a new controller. It comes in sizes of 512 GB, 1 TB, and 2 TB.

The best part about the 660p is its performance numbers. The drive takes advantage of PCI-Express 3.0 x4, and offers (at least on paper), performance numbers identical to those of the pricier 700p. The drives read at speeds of up to 1800 MB/s, with up to 1100 MB/s writes. The 600p, in comparison, capped out at 560 MB/s sequential writes, while the 700p is only slightly higher, at 1200 MB/s. Random access speeds are up to 150,000 IOPS (both reads and writes). QLC pays off rich dividends to consumers by way of price/GB. The 660p 512 GB is expected to be priced at 113.90€ (0.22€/GB), the 1 TB variant at 197.75€ (0.20€/GB), and the 2 TB variant at 391.43€ (0.20€/GB). Not bad for launch prices, considering these are PCIe NVMe drives priced competitively with SATA SSDs.



View at TechPowerUp Main Site

 

[dj-electric]

This closed-to-SATA pricing is pretty much forced.
DRAMless and drives like the ADATA SX8200 (using 760p's hardware, ironically enough) leave them no choice but to go under .30$ per GB

 

[Deeveo]

Now if they added a 4TB version for around 7-800€ that would be nice. Samsung 860 EVO 4TB can be found around 900€ now (Amazon.de).

 

[CheapMeat]

Curious as to what the SATA based QLC drives will cost. Hoping to get a bunch for my bulk storage.

 

[Fx]

Curious as to what the SATA based QLC drives will cost. Hoping to get a bunch for my bulk storage.

Hard drives are still the way to go for "bulk" storage.

 

[Owen1982]

And not one word about warranty length or TBW or any reliability stats/comments/tests regarding their QLC flash.

 

[TheLostSwede]

And not one word about warranty length or TBW or any reliability stats/comments/tests regarding their QLC flash.

It's a freakin' product leak, do you seriously expect full product details in a leak?!?!?

 

[Hood]

I'm curious about the endurance of QLC. P/E cycles for SLC are around 100,000, MLC around 10,000 and TLC around 1,000. Each generation results in an order of magnitude worse endurance. For QLC we’ve yet to see whether the P/E cycle will be in the 100 range or whether manufacturers have managed to make the technology more reliable somehow. Low 3-digit write endurance could make these drives relatively worthless, except for long term storage of permanent files that require no writes, such as movies, video footage, music, and photos. Anything that require writes, like an OS, application, or game, would quickly wear out the drive (maybe a year?)
Reading has no effect on endurance, so once you fill up that movie or music drive, it should last forever, as long as you don't ever change anything (be sure you get all the tags/cover art perfect before writing). But the price would have to be close to HDDs (2-3¢ /GB) to be feasible, not 20¢/GB as these are. This looks more promising for extremely cheap SATA drives. All this hinges on the actual write endurance of QLC flash cells, of course.
Also, the long term archive persistence (retention) is unknown - will these QLC cells still hold correct digital values after a few months, much less 10 or 20 years down the road? Or will the voltages tend to drift, like the problems already encountered with TLC (reads become progressively slower the longer data is stored). Samsung dealt with this by using an algorithm that periodically re-writes all the files, so that they never have a chance to deteriorate - not ideal, since write endurance is limited on TLC, so this work-around isn't even feasible on QLC.
I suppose we'll have to wait for the accelerated endurance tests that hammer the drive with write/erase cycles. If Intel has conducted these kind of tests, they're not talking.

 

[Owen1982]

Yes fair point, too soon for full details. Thinking out loud

 

[CheapMeat]

I'm curious about the endurance of QLC. P/E cycles for SLC are around 100,000, MLC around 10,000 and TLC around 1,000. Each generation results in an order of magnitude worse endurance. For QLC we’ve yet to see whether the P/E cycle will be in the 100 range or whether manufacturers have managed to make the technology more reliable somehow

The P/E cycle of QLC is expected to be 1,000, not 100. The same as planar TLC. Plenty of articles and news about this for the past 3 years stating so. It's no longer considered just a WORM drive (write once, read many). At 60+ layers, like upcoming 96 layer QLC, over provisioning is much easier & economical as well. Heck, Samsung is doing 8TB and 16TB QLC M.3 drives for 1U servers. It's how they're getting close to 1 Petabyte per server. They would not be doing so if 100 cycles.

Also those archiving data would be using various integrity schemes (they're not putting it on just 1 drive) and/or using tape or other avenues for long term storage (10-20+ years). Also consider that 3D Xpoint (Optane) has endurance higher than even SLC NAND. And then there's Samsung's Z-NAND, which treats MLC & TLC NAND like SLC with some special sauce & controller; so higher endurance and inherent qualities.

Hard drives are still the way to go for "bulk" storage.

Not for me. I'm going QLC NAND. I don't even care about 3.5" drives.

 

[bonehead123]

word:

sammypro

 

[Blueberries]

Large drives lead to laziness and poor maintenance (i.e., 100GB+ of unaccesed in years files or applications that haven't been deleted because the user hasn't needed space).

I find it surprising SATA drives, most particularly HDD, have stayed afloat. 512GB should be overkill for most people so long as they're not hoarders.

Are HDDs even viable for enterprise use these days?