After much time wasted with struggling to get VNC working on Ubuntu that stands for without monitor box - I've compiled the below draft notes on how I finally managed to configure the damn VNC. This should be stated, that I've come a long journey from systemd/init.d configuration scripts, through various VNC daemons (tightvnc, tigervnc, xvnc) and experienced lot of issues (VNC starting before X, VNC not having Xauthority, VNC not being able to open display :0, and so on) .

Finally, this below setup got me working with VNC.

This file serves a BKM to get better performance on CPU for PyTorch, mostly focusing on inference or deployment. Chinese version available here.

layout refers to how data is organized in a tensor. PyTorch default layout is NCHW, from optimization perspective, MKL-DNN library (renamed as DNNL recently) may choose a different layout, sometimes refered to as internal layout or primitive layout. This is actually a normal technique for acceleration libraries, common knowledge is that NHWC runs faster than NCHW for convolution, changing the default NCHW to NHWC is called a reorder. MKL-DNN may choose different internal layouts based on the input pattern and the algorithm selected, e.g. nChw16c, a.k.a. reorder a 4-dim tensor into 5-dim by chop down dimension C by 16, for vectorization purpose (AVX512 instruction length is 16x32 bit).

By default on CPU, conv2d will ru

This is a short post that explains how to write a high-performance matrix multiplication program on modern processors. In this tutorial I will use a single core of the Skylake-client CPU with AVX2, but the principles in this post also apply to other processors with different instruction sets (such as AVX512).

Matrix multiplication is a mathematical operation that defines the product of