Switch matrix parallel to task stealing, type and wrap are refacored, thread execution core has not yet

StgSharp.Native/sn-src/m.fma_tile.x86.c

@@ -7,34 +7,8 @@
 
 #define _FMA_SELECT(i) _MM_SHUFFLE(i, i, i, i)
 
-/*
-[MethodImpl(MethodImplOptions.AggressiveInlining)]
-internal static unsafe MatrixKernel<T>* GetKernelAddressUnsafe<T>(
-                                        MatrixKernel<T>* buffer,
-                                        int colLength,
-                                        int x,
-                                        int y
-) where T : unmanaged, INumber<T>
-{
-    return buffer + (colLength * x + y);
-}
-*/
-
 #define GET_INDEX(col_len, x, y) (col_len * x + y)
 
-/*
-for (int k = 0; k < commonK; k += 1) {
-    NativeIntrinsic.Context
-                       .mat[f64].kernel_fma(
-                        (MatrixPanel*)GetKernelAddressUnsafe((MatrixKernel<double>*)leftBuffer,
-                                                             ansHeight, k, j),
-                        (MatrixPanel*)GetKernelAddressUnsafe((MatrixKernel<double>*)rightBuffer,
-                                                             commonK, j, k),
-                        (MatrixPanel*)GetKernelAddressUnsafe((MatrixKernel<double>*)ansBuffer,
-                                                             ansHeight, i, j));
-}
-*/
-
 // float
 
 DECLARE_KERTILE_PROC_LEFT_RIGHT_ANS(float, sse, , fma)