q8_KV_r8: don't use nrc_y = 16 on Zen4

This is faster - 350 t/s. Why?
Much better than the 290 t/s we had before, but still slower
than the 370 t/s for q8_k_r8.

Author: Iwan Kawrakow

ggml/src/iqk/iqk_mul_mat.cpp

@@ -9367,9 +9367,9 @@ bool MulMat::prepare(int typeA, int typeB, int ne00, MulMat& mm, int Ny) {
             mm.funcs[5] = mul_mat_q8_KV_r8_q8_KV<6>;
             mm.funcs[6] = mul_mat_q8_KV_r8_q8_KV<7>;
             mm.funcs[7] = mul_mat_q8_KV_r8_q8_KV<8>;
-#ifdef HAVE_FANCY_SIMD
-            mm.func16 = mul_mat_q8_KV_r8_q8_KV<16>;
-#endif
+//#ifdef HAVE_FANCY_SIMD
+//            mm.func16 = mul_mat_q8_KV_r8_q8_KV<16>;
+//#endif
             expected_typeB = GGML_TYPE_Q8_KV;
             break;
         case GGML_TYPE_IQ4_K_R4: