Hello @KReiser

too many possible problems in math you can solve.

okay it's the lower half actually.

Just checking out the room ? @KReiser

Clicking on the notification brought me back here. Please do not tag me again - I do not wish to receive notifications from this chatroom at this time.

Another reason I don't like the documentation is there are usually approximations to the made functions, some of which can possibly run faster than the premade ones a=vector(#a,i,#a-i+1) is an equivalent to Vecrev on a predeclared vector a for example. This works as counting starts at 1 in PARI GP unlike most other languages.

Hello @ThomasShelby

Hi @RoddyMacPhee

Just checking out the room ?

@MrPie you thought I used parforvec cool before, you actually use it with a parfor and recreate forpart in parallel : parfor(n=nmin,nmax,X=parvector(n,i,[amin,amax]),c,parforvec(x=c,vecsum(x)==k,d,if(d,seq))) it's better to concat and return later though. and parallel only works faster if you code it right and have enough threads and memory.

parforpart(m,l,p,q)=my(nmin=q[1],nmax=q[2],amin=p[1],amax=p[2],seq=l,k=m);parfor(n=nmin,nmax,X=parvector(n,i,[amin,amax]),c,parforvec(x=c,vecsum(x)==k,d,if(d,eval(seq)))) is probably closer.

doh found my error... doesn't check of they are in order...

@MrPie