This can be recreated using a short function module, but it is complicated slightly by the fact that the LOC function operates in row major order while the VECH function operates in column major order, and so the need to transpose the matrix. The code below should do the trick, but I have no idea how efficient or inefficient this might be with your large matrices.
proc iml;
start vechalf(x);
a=nrow(x);
b=a#a-1;
return(t(x)[loc(mod(0:b,a)>=int((0:b)/a))]);
finish;
x=shape(10:25,4,4);
print x;
y=vechalf(x);
print y;
quit;