do 128 bit numbers work with gradeup & gradedown?

General APL language issues

do 128 bit numbers work with gradeup & gradedown?

Postby tclviii-dyalog on Mon Jun 24, 2019 4:59 pm

Roger Hui, (isnt it always Roger?), had this slick one liner utilizing gradeup that gave the Nth lexical permutation of iota X ...

... but Roger then mentioned something about 64 bit floats only having 53 bits of integer precision.

i'm guessin' that means X is limited to iota 18

does grade up work on the new 128 bit numbers? (and how many bits of integer precision does a 128 bit number have?)

thanks for taking the time & best regards,
TonyC

P.S. this has to do with a wacky idea that i have about speeding up Monte Carlo when using gamma rather than guassian variates
tclviii-dyalog
 
Posts: 8
Joined: Tue Mar 02, 2010 6:04 pm

Re: do 128 bit numbers work with gradeup & gradedown?

Postby AndyS|Dyalog on Tue Jun 25, 2019 8:20 am

Hi Tony

I am informed that 128-bit numbers have 34 decimal digit accuracy - 64-bit floats have 16/17 decimal digit accuracy.

All the primitives that support 64-bit floats support 128-bit decimal values .. just bear in mind that they are significantly slower. And we strongly recommend not switching between ⎕FR←645 and ⎕FR←1287 - there is a rather large "However" section in the ⎕FR documentation - http://help.dyalog.com/17.0/Content/Language/System%20Functions/fr.htm which you should read !
User avatar
AndyS|Dyalog
 
Posts: 220
Joined: Tue May 12, 2009 6:06 pm


Return to Language

Who is online

Users browsing this forum: No registered users and 1 guest