Looking at the source I see that m&cd"0 parses m for each call. That
could be factored out, & I'll put it on my work list.
3x performance improvement is statistically significant, but are we
talking about enough time that it really matters? The overheads for
function call and argument copying are pretty small.
Henry Rich
On 8/23/2020 10:19 PM, Kevin Li wrote:
Hi all,
I have a C function that computes a curvature, but for our purposes it's
simply a function whose C signature:
double f(double x);
I compiled this to a *.so (I am on Linux) and I could successfully call the
shared library via cd. However, there seems to be a
non-negligible overhead in calling through the DLL, when I call f over an
array. Specifically, I have:
f =: 'path.to.library.so f > d d'&cd"0
So I could do, e.g.
f i:10
Alternatively, I could define another function in the C file:
void f_list(double *x, long len) {
for (long i = 0; i < len; i++) x[i] = f[i];
}
and expose this in the *.so. Then I could define another verb
fs =: 'path.to.library.so f_list > n *d x&cd
and call that instead. I find that fs is usually 2-3 times faster than f,
presumably because of the overhead in calling into the shared library.
However, fs has two disadvantages:
1. It modifies the array in place, which makes my code less functional
2. It makes it harder to work with atoms (I have to convert an atom into a
1 dimensional array first).
Is there a way to get the best of both worlds?
Thanks!
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
--
This email has been checked for viruses by AVG.
https://www.avg.com
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
