Performing a shift in IDL with GPULib

A couple of users recently asked about doing a shift of a matrix using GPULib. While there is no interface that mimics 'shift' at the moment, all the functionality for shifting arrays is there. For example, you can use the following to shift an array along the x-direction:

IDL> x = findgen(5, 5)
IDL> print, x
0.00000 1.00000 2.00000 3.00000 4.00000
5.00000 6.00000 7.00000 8.00000 9.00000
10.0000 11.0000 12.0000 13.0000 14.0000
15.0000 16.0000 17.0000 18.0000 19.0000
20.0000 21.0000 22.0000 23.0000 24.0000

IDL> a = gpuPutArr(x)
IDL> b = gpuFltarr(5, 5)

; here comes the actual shift:
; we first perform b[0:3, *] = a[1:*, *]
; see the documentation for 'gpuSubArr' for an
; explanation of the arguments.
IDL> gpuSubArr, a, [1, -1], -1, b, [0, 3], -1

; and then b[4, * ] = a[0,*]
IDL> gpuSubArr, a, 0, -1, b, 4, -1

IDL> res = gpuGetArr(b)
IDL> print, res

1.00000 2.00000 3.00000 4.00000 0.00000
6.00000 7.00000 8.00000 9.00000 5.00000
11.0000 12.0000 13.0000 14.0000 10.0000
16.0000 17.0000 18.0000 19.0000 15.0000
21.0000 22.0000 23.0000 24.0000 20.0000

Shifts in other directions can be implemented similarly. This is not the fastest possible shift, but at least it should allow you to perform the shift on the GPU, rather than transferring it back to the CPU.

As always, if you have questions about this or any other use of GPULib, feel free to leave comments here, or email us at support@txcorp.com.