Sum of the elements in an array

[Gquadro]

Hi guys,

I would like to do this operation with the elements in my 1D array.

Given array:

a

b

c

d

what i would obtain in a new array:

a

a+b

a+b+c

a+b+c+d

could you show me how to do?

[aeastet]

I would something like this:

[altenbach]

Here's what I usually do

[mcduff]

I use this, but save it as a VIM. (Note it will not catch overflow errors for integer types).

mcduff

[altenbach]

Well, since the question is a bit vague, here's an alternative interpretation 😄

(... and yes I know there are much more efficient ways ...;)

(well, the subject says "sum", but who knows? ;))

[terickson]

[mcduff]

Here's a method that operates completely in-place, which can be turned into a VIM also.

---

@terickson wrote:

---

Neat implementation, but slow for big arrays.

mcduff

[altenbach]

---

@mcduff wrote:

Here's a method that operates completely in-place, which can be turned into a VIM also.

---

Don't underestimate the compiler. Mine (above) has the same number of array buffer allocations. 😉

If you really want to use the IPE, I think having an extra scalar (below) probably compensates for your operating on two array elements with each iteration. Of course you have one fewer iterations, which could make a difference for extremely small arrays.

Not tested, just guessing of course. 😄

[mcduff]

---

@altenbach wrote:

---

---

Don't underestimate the compiler. Mine (above) has the same number of array buffer allocations. 😉

---

+1. This is why I said the VI operates totally in-place. The screen shots below are LabVIEW 2019, using the "Profile Buffer Allocations Tool"

Profile Tool only records 1 buffer

Using the VI from the earlier post, which is a bit faster, I see this

Profile Tool records two buffers

It's possible the first buffer in the second screen shot is erased immediately after execution; this is @rolfk territory, so in effect 1 buffer. But at least according to NI's tools, there are 2 buffer allocations.

---

@altenbach wrote:
If you really want to use the IPE, I think having an extra scalar (below) probably compensates for your operating on two array elements with each iteration. Of course you have one fewer iterations, which could make a difference for extremely small arrays.

Not tested, just guessing of course. 😄

Besides complex numbers I thought you liked to do things in parallel. 🙂

mcduff

---

[thols]

I like these threads. Its like when someone asks a in forum to photoshop him holding the sun. (or, no its not quite there yet).