Stories
Slash Boxes
Comments

SoylentNews is people

SoylentNews is powered by your submissions, so send in your scoop. Only 15 submissions in the queue.
posted by Fnord666 on Sunday September 29 2019, @08:27PM   Printer-friendly
from the do-the-bits-protest-their-crowded-conditions? dept.

Arthur T Knackerbracket has found the following story:

Wednesday, Intel announced it's joining Toshiba in the PLC (Penta-Level Cell, meaning 5 bits stored per individual NAND cell) club. Intel has not yet commercialized the technology, so you can't go and buy a PLC SSD yet—but we can expect the technology will lead eventually to higher-capacity and cheaper solid state drives.

Intel also differentiates itself from competitors by sticking with the floating-gate cell design used in early SLC devices, instead of the less expensive charge-trap design the rest of the industry has shifted to. It's unclear to casual researchers which technology is actually better from a technical perspective, but Intel argues that the floating gates can be manufactured at a higher density, meaning it can pack more cells into the same physical area.

Unfortunately, while PLC SSDs will likely be bigger and cheaper, they'll probably also be slower. Modern SSDs mostly use TLC storage with a small layer of SLC write cache. As long as you don't write too much data too fast, your SSD writes will seem as blazingly fast as your reads—for example, Samsung's consumer drives are rated for up to 520MB/sec. But that's only as long as you keep inside the relatively small SLC cache layer; once you've filled that and must write directly to the main media in real time, things slow down enormously.

[...] With sequential write speeds to QLC media already decreasing to or below that of conventional hard drives, PLC seems likely to be a niche player that will compete far more with NAS and datacenter drives than it does with laptop and desktop SSDs aimed at high performance. Sequential throughput isn't everything, of course—and PLC media should still offer much higher IOPS in challenging random-access workloads than conventional disks can. But it's probably not going to be a good solution in anything but truly massive-capacity drives, which can use higher parallelism (think "invisible RAID0") to offset the invididually-slow characteristics of PLC cells.

-- submitted from IRC


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 1, Interesting) by Anonymous Coward on Sunday September 29 2019, @08:59PM (7 children)

    by Anonymous Coward on Sunday September 29 2019, @08:59PM (#900517)

    Reducing robustness 10x and speed several times from SLC to MLC saved 50% silicon. Understandable.
    Reducing robustness 10x and speed several times from MLC to TLC saved 33% silicon. Painful but still somewhat understandable.
    Reducing robustness 10x and speed several times from TLC to QLC saved 25% silicon. Starts to smell of sadism.
    Reducing robustness 10x and speed several times from QLC to PLC would save 20% silicon. Looks downright crazy, doesn't it?

    Either the idea is to train the marks that any out-of-cloud data storage is unreliable as hell; or the current crop of execs are unable to perform simple arithmetics.

    Starting Score:    0  points
    Moderation   +1  
       Interesting=1, Total=1
    Extra 'Interesting' Modifier   0  

    Total Score:   1  
  • (Score: 2) by takyon on Sunday September 29 2019, @09:06PM (6 children)

    by takyon (881) <takyonNO@SPAMsoylentnews.org> on Sunday September 29 2019, @09:06PM (#900520) Journal

    TLC and QLC supposedly have about 1/3 the write cycles of the previous type instead of 1/10:

    https://www.architecting.it/blog/qlc-nand/ [architecting.it]

    P/E cycles for SLC were around 100,000, MLC around 10,000 and TLC around 1,000, although this figure has been improved by vendors. Each generation results in an order of magnitude worse endurance. For QLC we were expecting the figures to be around the 100 range. However, manufacturers have improved the resilience and we now see around 3,000 P/E cycles for TLC and 1,000 for QLC (figures taken from Micron Reviewers’ Day, August 2018).

    --
    [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
    • (Score: 1, Touché) by Anonymous Coward on Sunday September 29 2019, @10:21PM (5 children)

      by Anonymous Coward on Sunday September 29 2019, @10:21PM (#900561)

      We all know the honest and reliable nature of advertising materials, don't we?

      • (Score: 2) by takyon on Sunday September 29 2019, @10:35PM (4 children)

        by takyon (881) <takyonNO@SPAMsoylentnews.org> on Sunday September 29 2019, @10:35PM (#900565) Journal

        If QLC NAND endurance was bad to the point of being unusable, we would have heard about it by now.

        If you make the drive big enough (e.g. 4 TB), some people will be able to use it just fine even if it only has 10 P/E cycles. The write speed issue only comes into play if you can fill the SLC or DRAM cache. The major advantage of SSDs over hard drives isn't sequential speeds, it's random read IOPS.

        QLC NAND can be used by casual phone/laptop/desktop users, as well as companies like Facebook that asked for the technology [tomshardware.com] and know the tradeoffs.

        The only thing that probably can't be waved away is unpowered data retention. So don't shove your external 3D PLC SSD into a desk and leave it there for 2 years.

        --
        [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
        • (Score: 1) by fustakrakich on Sunday September 29 2019, @10:58PM

          by fustakrakich (6150) on Sunday September 29 2019, @10:58PM (#900583) Journal

          The major advantage of SSDs over hard drives isn't sequential speeds, it's random read IOPS.

          And the lack of moving parts.

          --
          La politica e i criminali sono la stessa cosa..
        • (Score: 0) by Anonymous Coward on Sunday September 29 2019, @11:35PM (2 children)

          by Anonymous Coward on Sunday September 29 2019, @11:35PM (#900602)

          The write speed issue only comes into play if you can fill the SLC or DRAM cache.

          Which, in our days of 4K video and broadband, is rather easy to achieve in domestic environment.

          The major advantage of SSDs over hard drives isn't sequential speeds, it's random read IOPS.

          Which, in our days of multiple GBs of main memory being used for disk cache, is entirely marginal in said domestic environment.

          • (Score: 2) by takyon on Monday September 30 2019, @12:13AM (1 child)

            by takyon (881) <takyonNO@SPAMsoylentnews.org> on Monday September 30 2019, @12:13AM (#900617) Journal

            Which, in our days of 4K video and broadband, is rather easy to achieve in domestic environment.

            Most people who watch 4K video stream it instead of downloading it.

            Here's the Samsung 860 QVO using QLC NAND: https://www.anandtech.com/show/13633/the-samsung-860-qvo-ssd-review/2 [anandtech.com]

            The SLC cache size is up to 42 GB for the 1 TB model, or 78 GB for the 2 TB and 4 TB models. That seems pretty good, and maybe we will see larger caches with larger consumer SSDs.

            Which, in our days of multiple GBs of main memory being used for disk cache, is entirely marginal in said domestic environment.

            How much memory? If you're right, you shouldn't have a problem with spinning disk or a smaller, faster SSD.

            I suspect that boot and application load times are still going to be better with a PLC SSD than spinning disk.

            --
            [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
            • (Score: 0) by Anonymous Coward on Monday September 30 2019, @02:13AM

              by Anonymous Coward on Monday September 30 2019, @02:13AM (#900650)

              The SLC cache size is up to

              I suspect even small children these days do know, how "up to" translates from marketspeak.

              How much memory?

              About 3.5 GB buff/cache as top says, out of 16 GB installed.

              If you're right, you shouldn't have a problem with spinning disk or a smaller, faster SSD.

              Naturally, a spinning disk is happily spinning in said system.

              I suspect that boot and application load times are still going to be better with a PLC SSD than spinning disk.

              Some maybe five or seven seconds less each boot, and a couple seconds when something huge gets launched the first time, could I guess be designated "better" - for a purely nominal value of "better". Still, how long would those seconds have to accumulate, to cover even the time needed to set up a SSD+HDD configuration? Or backups to external HDD for SSD-only system.
              Incidentally, a whole-partition restore from a backup is exactly the sort of load that TLC/QLC/PLC SSDs are destined to horribly suck at.