Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Friday December 04 2015, @05:35PM   Printer-friendly
from the stocking-stuffer-reference-items dept.

In my experience, one of the highest-impact upgrades you can perform to increase Raspberry Pi performance is to buy the fastest possible microSD card—especially for applications where you need to do a lot of random reads and writes.

There is an order-of-magnitude difference between most cheap cards and the slightly-more-expensive ones (even if both are rated as being in the same class)—especially in small-block random I/O performance. As an example, if you use a normal, cheap microSD card for your database server, normal database operations can literally be 100x slower than if you used a standard microSD card.

Because of this, I went and purchased over a dozen different cards and have been putting them through their paces. Here are the results of those efforts...

Visit TFA for the full table. The overall winner seems to be OWC Envoy SSD (USB), with hdparm buffered: 34.13 MB/s; dd write: 34.4 MB/s; 4K rand read: 7.06 MB/s; 4K rand write: 8.20 MB/s


takyon: The value winner in the article is the Samsung Evo+ 32 GB (purchased for $9.99 from Best Buy) with decent/passable speeds.

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This discussion has been archived.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: 5, Insightful) by JoeMerchant on Friday December 04 2015, @05:56PM

    by JoeMerchant (3937) on Friday December 04 2015, @05:56PM (#271879)

    Interesting, actionable knowledge.

    --
    🌻🌻 [google.com]
    • (Score: 3, Informative) by Phoenix666 on Friday December 04 2015, @06:02PM

      by Phoenix666 (552) on Friday December 04 2015, @06:02PM (#271883) Journal

      Thanks for the feedback. There aren't many like this that show up in the usual places like ArsTechnica or phys.org--if you have additional places to look please mention them or submit articles from them. It would be great to diversify away from traditional outlets.

      --
      Washington DC delenda est.
      • (Score: 4, Informative) by takyon on Friday December 04 2015, @06:27PM

        by takyon (881) <reversethis-{gro ... s} {ta} {noykat}> on Friday December 04 2015, @06:27PM (#271894) Journal

        I would replace phys.org with Futurity [futurity.org].

        Instead of linking directly to Futurity you can easily find the university websites the articles are sourced from. Phys.org is scummy by comparison and makes it difficult to find the original source.

        In addition, the DOI and link to the scientific study is included at the bottom of the article where applicable (almost always).

        The only problem is that you may need to Inspect Element to get the non-capslocked form of the Futurity headline, if you intend to use that.

        Here is an example story:

        http://www.futurity.org/orcas-whales-ships-speed-1062552-2/ [futurity.org]
        http://www.washington.edu/news/2015/12/02/vessel-speed-biggest-factor-in-noise-affecting-killer-whales/ [washington.edu]
        DOI: 10.1371/journal.pone.0140119
        http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0140119 [plos.org]

        Every scientific paper published should have a unique DOI identifier, which can be very useful when searching databases for the paper.

        --
        [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
        • (Score: 0) by Anonymous Coward on Friday December 04 2015, @10:43PM

          by Anonymous Coward on Friday December 04 2015, @10:43PM (#271981)

          > The only problem is that you may need to Inspect Element to get the non-capslocked form of the Futurity headline, if you intend to use that.

          Copy-paste fixes it automagically for me. Which is to be expected since you can't paste a fontface into soylent.

      • (Score: 3, Informative) by Hairyfeet on Saturday December 05 2015, @01:06PM

        by Hairyfeet (75) <{bassbeast1968} {at} {gmail.com}> on Saturday December 05 2015, @01:06PM (#272148) Journal

        Try Daily Rotation [dailyrotation.com] which is where I got every article I ever submitted, they even have a handy sort where you can just to filter by the latest trending topics like Encryption, Open Source, Apple/Google/MSFT, tablets, phones, etc.

        If you want a "one stop shop" for nerdy tech heavy articles? they are the place to go, in fact you could probably keep this site very well fed with all the latest tech news simply by grabbing stories from Daily Rotation. Hell just looking at their top 10 headlines you have stories about solving the mystery of leaking arsenic in ground water,a story on let's encrypt, magnetic fields near the Milky Way black hole, its about as nerdy and tech heavy as one can get.

        --
        ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
        • (Score: 2) by Phoenix666 on Saturday December 05 2015, @01:48PM

          by Phoenix666 (552) on Saturday December 05 2015, @01:48PM (#272158) Journal

          Thanks, HairyFeet.

          --
          Washington DC delenda est.
    • (Score: 2) by frojack on Friday December 04 2015, @07:15PM

      by frojack (1554) on Friday December 04 2015, @07:15PM (#271914) Journal

      Yes, but hardly news.

      Seems like every tech rag has compared brands over the years, and this comparison stands out ONLY due to sneaking Raspberry Pi into the title. It would have been better had he at least mentioned the Class of each card, as well as representative prices.

      --
      No, you are mistaken. I've always had this sig.
      • (Score: 2) by richtopia on Friday December 04 2015, @07:35PM

        by richtopia (3160) on Friday December 04 2015, @07:35PM (#271917) Homepage Journal

        In the name of the card they listed the class; all of them but one had C10 to designate class 10

        • (Score: 2) by frojack on Friday December 04 2015, @07:40PM

          by frojack (1554) on Friday December 04 2015, @07:40PM (#271919) Journal

          Perhaps if you followed the links inside the article but simple inspection of TFA reveals that what you claim simply isn't true. Only three cards had a class number in TFA.

          --
          No, you are mistaken. I've always had this sig.
      • (Score: 2) by JoeMerchant on Friday December 04 2015, @08:42PM

        by JoeMerchant (3937) on Friday December 04 2015, @08:42PM (#271941)

        While the testing done in the TFA may not be the most rigorous or scientifically reproduceable, they resemble what I would do if I had the time and are probably more applicable to my real-world performance concerns than the rigorous "fair" tests that usually get published.

        And, "RaspberryPi in the title" is kind of relevant - I'm a lot more likely to buy and equip a few Raspberry Pis in the coming months than I am to make decisions about super-computers, esoteric materials, high dollar systems, etc.

        --
        🌻🌻 [google.com]
        • (Score: 2) by frojack on Friday December 04 2015, @09:24PM

          by frojack (1554) on Friday December 04 2015, @09:24PM (#271955) Journal

          Most people are far more likely to read such a comparison-review with regard to a smartphone or camera or tablet.

          In relative terms, Raspberry Pis are non-existent.

          Disclosure: I have an old Pi running cups as a print server. Perfectly adequate for any printer that does not need anything more than a ppd for a driver. It only has a class 4 card in it. Maybe someday I'll put in something faster, but realistically, its fast enough.

          Oh, look, the stash on the shelf has three Class 10 Ultras still in the bubble-pack. (I tend to buy these when they go on sale even if I have no immediate need.) Maybe I'll upgrade that old Pi. But probably not, since it took me a couple years to find any use for it in the first place.

          So much of what people do with PIs is "Just because I can" sort of projects.

          --
          No, you are mistaken. I've always had this sig.
          • (Score: 2) by JoeMerchant on Saturday December 05 2015, @03:48PM

            by JoeMerchant (3937) on Saturday December 05 2015, @03:48PM (#272186)

            My Pi usage is mostly headless server stuff, but I do run an X desktop on it and I actually develop some stuff in Qt Creator directly on the target Pi, so performance is ... marginal, and the difference between a Class 4 and a Class 10 card is relevant to me during development, not so much in deployment where a Class 4 would be fine.

            I've seen vague references in product reviews stating the obvious "not all Class 10 cards are created equally" and have always wondered just how significant those differences are and whether or not I should care. In a practical sense, I too have 3 Class 10 cards still unopened in the bubble pack, purchased to have handy the next time I want to do a "Hold My Beer while I try this new configuration" experiment. But, I do have friends and colleagues who are actively developing new Pi projects, and it's nice to be able to show them what I've got and speak a little more knowledgeably about what the impact of their SD card selection really is.

            Mostly, I will send them a link to the article, rather than getting into the discussion directly.

            --
            🌻🌻 [google.com]
  • (Score: 3, Interesting) by richtopia on Friday December 04 2015, @06:04PM

    by richtopia (3160) on Friday December 04 2015, @06:04PM (#271884) Homepage Journal

    This article is really interesting. I am curious how a single vendor varies between card and card.

    I don't buy flash often, but I have read many articles similar to this for USB drives. It is tricky to really identify the winners or more importantly the losers.

    • (Score: 2, Interesting) by Anonymous Coward on Friday December 04 2015, @08:24PM

      by Anonymous Coward on Friday December 04 2015, @08:24PM (#271930)

      It can be a significant issue when using sd cards in production with more complicated gadgets. For example, if you have a device which stores its fault log in eeprom, and it keeps writing to the fault log recorder errors which arise from the sd card being unable to handle the writes properly, it can blow your eeprom up as well.

      When you're doing serious field work, you never want to skimp on your removable storage.

      --Ethanol-fueled from the beach

  • (Score: 3, Insightful) by wonkey_monkey on Friday December 04 2015, @06:28PM

    by wonkey_monkey (279) on Friday December 04 2015, @06:28PM (#271895) Homepage

    Visit TFA for the full table. The overall winner seems to be OWC Envoy SSD (USB)

    Well, yes, I'm not surprised at an USB SSD drive winning a MicroSD card comparison test, if you let one compete...

    I think it, and the Sandisk USB flash drive, was just in there to give a bit of context for the SD card speeds.

    --
    systemd is Roko's Basilisk
    • (Score: 1) by legont on Friday December 04 2015, @10:18PM

      by legont (4179) on Friday December 04 2015, @10:18PM (#271968)

      My issue with Pi is not speed, but reliability, therefore I want to move root partition to USB attached SSD. The question is if Pi will provide enough power. Any recommendations out there? No, I don't want self powered USB hub.

      --
      "Wealth is the relentless enemy of understanding" - John Kenneth Galbraith.
  • (Score: 2) by Snotnose on Saturday December 05 2015, @12:30AM

    by Snotnose (1623) on Saturday December 05 2015, @12:30AM (#272025)

    Some 10-15 years ago I had to write an SDCC driver. Details are fuzzy, but for version 1.0 you had flowcharts on how to read/write/erase/etc. Violating the flowchart resulted in a timeout instead of an error more often than not. Then cards larger than what 1.0 could support came out and the flowcharts had to be updated. How, you ask? Early in the process you sent the SDCC card a command that 1.0 didn't recognize but 2.0 did. Then waited to see if it timed out, if it did you had a 1.0 chip. Whoops, some vendors of 1.0 chips added a status bit that, when set, indicated WTF; but they returned that error fairly quickly. Then cards larger than what 2.0 supported came out. Again, add a command that neither 1.0 nor 2.0 recognized, but 3.0 did. Wait for a timeout and act accordingly. This is in addition to lots of fuzzy timing requirements, and some states had exits that weren't accounted for in the flowchart.

    Now consider each timeout is several milliseconds long, and it wasn't uncommon to get 3-4 timeouts when starting up, things got real slow real fast. Didn't help that my first driver was for 1.0, then modified for 2.0, then modified for 3.0, etc. Would have been better to rewrite it to assume 3.0 and, if that failed, revert to the slower earlier versions.

    I've written maybe 50 device drivers over the last 30 years, the SDCC driver was one of the more challenging ones. I gained a lot of respect for what Microsoft had to put up with, all my other drivers had the chip I was talking to soldered to the board.

    --
    Why shouldn't we judge a book by it's cover? It's got the author, title, and a summary of what the book's about.
  • (Score: 2) by darkfeline on Saturday December 05 2015, @05:09AM

    by darkfeline (1030) on Saturday December 05 2015, @05:09AM (#272088) Homepage

    Well, that's kind of a pointless optimization, isn't it? There aren't very many cases where you would want to increase your Raspberry Pi's IO performance to and from an SD card.

    I mean, first of all if you're doing IO intensive operations you wouldn't want to do that on the SD card, which will die quite quickly from all those writes. You would want to use the SD card solely for booting and move everything else to a proper external drive, HDD or SSD.

    Hell, TFS even says the winner is an SSD. Kind of puts the whole thing into perspective.

    --
    Join the SDF Public Access UNIX System today!