Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>Linear congruential generators all suffer from a problem identified by George Marsaglia. "Marsaglia's Theorem" says that k-tuples (vectors of length k) will fall on a bounded number of hyperplanes. The bound is <code>m**(1/k)</code>, where k is the size of the tuple and m is the number used for the modulus of the generator. Thus, if the modulus is <code>(2**31 - 1)</code> and you're looking at sets of 3, a 3-d plot will show the points falling on no more than the cube root of <code>(2**31 - 1)</code>, or about 1290 planes, when viewed from the right orientation.</p> <p>All LCG's are subject to Marsaglia's Theorem. A "good" one performs at or close to the upper bound, a bad one falls well short of the upper bound. That's what the spectral test is effectively measuring, and that's what you were seeing in your Wikipedia link - RANDU, the LCG from hell, produces triplets that fall in a mere 15 planes.</p> <p>Apple's carbon library generator uses 16807 as its multiplier and <code>(2**31 - 1)</code> as its modulus. As LCG's go, it's not really all that bad. Hence your plot didn't show the same extremes RANDU has. However, if you want decent quality random numbers don't use an LCG.</p> <p><strong>Addendum</strong></p> <p>I've gone ahead and cranked a billion numbers from the Apple rand() function, but only printed the ones where both values of the pair were less than 2 million, i.e., the bottom left corner of your plot. Sure enough, they fall on lines. You just need to really zoom in to see it because of the density of lines.</p> <p>Old George was a clever fella!</p> <p><img src="https://i.stack.imgur.com/IFXJ5.png" alt="Marsaglia at work"></p>
    singulars
    1. This table or related slice is empty.
    1. POHow can I make OSX's rand() fail the spectral test?
      singulars
      1. PTQuestion
    1. PTAnswer
    1. USpjs
    1. USpjs
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. POHow can I make OSX's rand() fail the spectral test?
      singulars
      1. PTQuestion
    1. This table or related slice is empty.
    1. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    3. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    1. COThanks! I had a hard time understanding how the lines could be vertical, but then I `sort -n`'ed the data and found that the lines are in fact not perfectly vertical but slightly tilted.
      singulars
      1. PO
      1. USlindelof
    2. COYou're welcome! Good call using `sort -n`. Yeah, the lines look vertical until you realize that the spacing between "adjacent" lines is over 100K, so horizontal shifting by 100's or even 1000's is barely going to be noticeable.
      singulars
      1. PO
      1. USpjs
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload