Hallo!
but looking at the headers for SIMD instructions, it seems I have to set up a dsp-perform routine (or something like it) for the plus_perf_simd instruction
look at e.g. volctl~ by Tim: you have to make one perform routine without simd (volctl_perform) and one for simd (volctl_perf_simd)
Then in the volctl_dsp you look (at runtime) if the processor can make SSE:
if(SIMD_CHECK2(n,sp[0]->s_vec,sp[1]->s_vec)) dsp_add(volctl_perf_simd, 4, x, sp[0]->s_vec, sp[1]->s_vec, n);
if so, you can use the simd perform routine ...
LG Georg