Neither the high bit-depth SSE code nor any of the AVX2/AVX-512 implementations appears to have the same issue as far as I can see. SGR is also unaffected.