Forgot your password?
Close
typodupeerror
Math Data Storage

Backblaze Hosts 314 Trillion Digits of Pi Online (nerds.xyz) 67

Posted by BeauHD from the impressive-datasets dept.
BrianFagioli shares a report from NERDS.xyz: Cloud storage company Backblaze has partnered with StorageReview to make a massive dataset containing 314 trillion digits of Pi publicly accessible. The digits were calculated by StorageReview in December 2025 after months of heavy computation designed to stress modern hardware. The dataset now hosted in the cloud weighs in at over 130TB, while the full working dataset used during the calculation reached about 2.1PB when intermediate checkpoints were included. The report notes that the Pi digits have been broken into roughly 200GB chunks to make it more practical for researchers or enthusiasts to download.

Here's what StorageReview founder Brian Beeler said about the project: "Pushing [Pi] to 314 trillion digits was far more than a headline number. It was a sustained, months-long computational challenge that stressed every layer of modern infrastructure, from high core-count CPUs to massive high-speed storage, and it gave us valuable insight into how extreme, real-world workloads behave at scale. Making this dataset available in the Backblaze cloud takes the project a step further by opening access to one of the largest raw outputs ever generated in a single-system calculation. Hosting multi-petabyte files for the broader community is no small feat, and we appreciate Backblaze stepping up to ensure researchers, developers, and enthusiasts can explore and build on this record-setting achievement."
This discussion has been archived. No new comments can be posted.

Backblaze Hosts 314 Trillion Digits of Pi Online More

Backblaze Hosts 314 Trillion Digits of Pi Online

Comments Filter:
  • Isn't this a very linear process, one core should be plenty for the calculation itself? Or a better question why is and how is a parallel process even needed other than for storing the results to disk?

  • Confirmed (Score:5, Funny)

    by battingly ( 5065477 ) on Friday March 13, 2026 @03:11PM (#66039706)

    I just checked it on my TI calculator and confirmed the digits are correct.

  • I use the last digits of pi as my standard pin code.

  • 130TB? (Score:3)

    by thegarbz ( 1787294 ) on Friday March 13, 2026 @03:20PM (#66039724)

    Jesus just zip the damn thing! Do you even compress bro? /s

    • You can certainly gain something by compressing ASCII down to binary. That could get you a factor of 10/256 theoretically. However, the digits of pi are more-or-less random, not easy to compress further with any algorithm, not even the one zip uses.

      • Whoops, make that log(10)/log(256), per the post from 'Sloppy' below.

      • However, the digits of pi are more-or-less random, not easy to compress further with any algorithm, not even the one zip uses.

        I thought signing off my post with "/s" was pre-emptively woosh enough but some people just live with wind in their hair.

  • Spoiler alert (Score:4, Funny)

    by Kamineko ( 851857 ) on Friday March 13, 2026 @03:27PM (#66039742)

    Spoilers: it's three and a bit.

  • How sure are we that it's accurate? Has the data been validated? If they are stressing the hardware then couldn't a bit get flipped at some point and throw off the whole things?

  • You bastards! (Score:3)

    by UnknowingFool ( 672806 ) on Friday March 13, 2026 @03:40PM (#66039766)
    Now I have to change my luggage combination again. Thanks for nothing!
  • Does anyone know off the top of their head the 314,000,000,000,001th digit of pi?

    • 3.

      Prove me wrong.

    • /*
      * Computation of the n'th decimal digit of \pi with very little memory.
      * Written by Fabrice Bellard on January 8, 1997.
      *
      * We use a slightly modified version of the method described by Simon
      * Plouffe in "On the Computation of the n'th decimal digit of various
      * transcendental numbers" (November

    • Re: (Score:2)

      by rbrander ( 73222 )

      I checked TFA, but it just said "314 Trillion". I'm still wondering if the actual number of digits they stopped at was 314,592,653,589,793 ...if guys like that start a little joke with "314", you'd think they'd go all the way.

  • Decent compression (Score:3)

    by Sloppy ( 14984 ) on Friday March 13, 2026 @04:19PM (#66039860) Homepage Journal

    >>> import math
    >>> 314*math.log(10)/math.log(256)
    130.38567772432899

    130TB for 314T decimal digits sounds about right!

    I'd just use a plain arithmetic encoder (where all 10 symbols are equally weighted) with no dictionary pass (since the next digit can't be predicted from any or all of the previous digits).

  • Simple (Score:2)

    by PPH ( 736903 )

    The digits are: 0, 1, 2, 3, 4, 5 ,6 ,7 ,8, 9.

    Duplicate and arrange as needed.

    • Re: (Score:2)

      by robi5 ( 1261542 )

      Why, just because we have 10 fingers? No, it's 0, 1. Or perhaps 0, 1, 2, 3, 4, 5, 6.
      The conspiracy theorists of mathematicians are fascinated by the digits of pi. But invariably, in an arbitrary number base: 10.

  • This is just flat-out waste, anything beyond about 90 digits for pi has no real-world use case. Unless you want to look up some random subset a trillion or so digits in to use for encryption or a random number seed?

    • The summary says it was used as a stressor. Wikipedia: "y-cruncher can also be used for stress-tests, as performed computations are sensitive to RAM errors and the program can automatically detect such errors.[1][2]; "The technical challenge does not (any longer) lie in the calculation itself, but in providing an environment that enables a comparatively efficient execution.[11]" https://en.wikipedia.org/wiki/... [wikipedia.org]

      Also about the previous record from Google: "This achievement is a testament to how much faster G

  • I believe that after about 7 digits, the rest really doesn't matter. For practical calculations like calculating the trajectory of a spaceship through the universe, it would only be a few mm off at about 9 digits. So, with no practical applications, this seems kind of a waste of time, and compute power. But I admit it is kind of geeky and cool.

    • Agree. Initial measurements when setting off would have margins of error too, making the mm of difference less important. Corrections on the fly have to be factored in.

  • Has anyone converted it to base11

  • Useless... (Score:3)

    by LordHighExecutioner ( 4245243 ) on Friday March 13, 2026 @05:16PM (#66040006)
    ...with this method [wikipedia.org] you can calculate whichever digit of pi you want with an hand-cranked calculator.

  • ...why?

  • I always round Pi once I get to 100 trillion digits.

    This is 3.14 times too much.

  • ridiculous precision (Score:3)

    by Tom ( 822 ) on Friday March 13, 2026 @05:55PM (#66040104) Homepage Journal

    If the math is correct, that means we can calculate the circumference of the known universe to a precision much, much smaller than the Plank length.

    So in other words: No, there isn't any practical application for this, not now, not in the forseable future, and probably not before the heat death of the universe.

    • No, there isn't any practical application for this

      The point of calculating pi to this degree was never about practical applications. Practical application cease being practical at precision you can hand calculate. The point of this was an attempt to confirm a mathematical theory that pi is truly infinitely irrational without a repeating pattern.

      • The point of this was an attempt to confirm a mathematical theory that pi is truly infinitely irrational without a repeating pattern.

        You'll never achieve that by calculating a finite number of digits of pi, no matter how many. The fact that pi is irrational was settled long ago, by deductive reasoning, not calculating.

        I'd say the point of calculating pi to this degree was to demonstrate the computational system, and to achieve bragging rights, even if only briefly. Someone else will calculate more digits soon enough.

        • Re: (Score:2)

          by Tom ( 822 )

          Someone else will calculate more digits soon enough.

          Really soon now, we'll figure out that the entire purpose of Earth was always to calculate Pi. Douglas Adams was right, he just lied about the actual number. :-)

    • If the math is correct, that means we can calculate the circumference of the known universe to a precision much, much smaller than the Plank length.

      The Planck length is about 1.616 x 10^-35 m. The circumference of the known universe is about 2.77 x 10^27 m. The ratio is about 5.834 x 10^-63.

      So, calculating the circumference of the universe to the precision of the planck length would require about 64 digits of pi. We knew pi to that many digits by 1706, when 100 digits were known.

    • Re: (Score:2)

      by bn-7bc ( 909819 )
      Have wi found a deffinitive length for tha plank length yet? If not how van you possibly know that?

      • Re: (Score:2)

        by bn-7bc ( 909819 )
        And ofc the plank length yet was posted a few comments prlow me ( out eatlier) thus rendering my atemyed snatk vooud, an slasdot still ,acks a delrtr buron in 2026 so mod plz remove both this its emidiate parrent to keep this trewad free of redundant comments

  • That's great and all. You do you.

    > 130TB of Pi digits
    > Backblaze says making the dataset available in its B2 cloud storage platform allows researchers, developers, and curious tech enthusiasts to explore the results themselves.
    .

    Any ideas for any (serious) examples of what someone might do with this? I imagine anybody really serious about number theory of pi has already access to something better and perhaps bigger....but I don't know. Help me understand the point of this.

  • Hmmm (Score:2)

    by allo ( 1728082 )

    Are you sure that digit 386,761 is correct?

  • They stopped 159 billion or so digits from truly being awesome.

  • That 2/1,000,000,000,000 the digits.

    3.14159

  • Is that binary digits or decimal, there is a bit off difference in storrade requirements.at any rate when whould me ever need that kind of precision for pi, we would need any other values involved in the calculation to defend a single constant ocupying multible GB of memory, what am I missing here ?
  • " Two brothers in a cramped Manhattan apartment build a homemade supercomputer to chase billions of digits of pi—and a glimpse of order." (1992) One of the all time great articles.

    The Mountains of Pi [newyorker.com]

  • The observable universe is about 47 billion light years. If one consider a sphere with that diameter and with the Earth in the centre, how long distance would ±1 on the last digit in this list of digits change?

  • Is it base 2? Or base 10? How is it useful if I need a digit in another base? Why would someone be interested in a specific digit of pi in a specific base?

  • It's easy, if you sort them first before compression. I have roughly 100 trillion of each number 0-9, and a single decimal point.

Slashdot Top Deals

The universe is all a spin-off of the Big Bang.

Close