True of this tensor is dense and row-major

LZ4_compress_generic() : * inlined, to ensure branches are decided at compilation time; * takes care of src == (NULL, 0) * and forward the rest

Couple utility methods, they workaround the lack of bit fields in C++

COM Server methods

manipulating listbox items

On nVidia 1080Ti that approach is much slower, by a factor of 2.4 On AMD Cezanne that approach is faster by a factor of 0.69, i.e. 30% faster. Dunno w

Detach without Release(), set this pointer to nullptr

inline static void ggml_vec_scale_f32(const int n, float * y, const float v) { for (int i = 0; i < n; ++i) y[i] *= v; }

Pointer to the start of the buffer, aligned by memory page = 4 kilobytes

for single-selection listboxes

for entire listbox