TheXYZLab Blog

Sign in Subscribe

Latest — Aug 11, 2023

Managing shell commands with Python

Let's admit this, shell scripts sometimes can be hard to write. Especially if you got 1000s shell commands to run and would like to restrict the resources they can use on with a 10-core CPU. Whether it be test programs or simulations, it would be nice if we

More issues

Paper Reading Notes #03: meProp: Sparsified Back Propagation for Accelerated Deep Learning with Reduced Overfitting

The paper "meProp: Sparsified Back Propagation for Accelerated Deep Learning with Reduced Overfitting" from ICML 2017 by researchers at Peking University. This paper presents a technique to speed up machine learning model training by gradient sparsification. The key idea is sparsifying backpropagation gradients by retaining only the top-k

Paper Reading Notes #02: Towards Fully Sparse Training: Information Restoration with Spatial Similarity

Pruning is a popular technique for reducing the size of deep neural networks without sacrificing accuracy. However, traditional pruning methods can be computationally expensive and lack to hardware support. In this paper, Towards Fully Sparse Training: Information Restoration with Spatial Similarity, the authors propose a new approach to structured pruning

Paper Reading Notes #01: Attention Is All You Need

This is a new series of my notes on paper reading, covering various areas in computer architecture, algorithms, and machine learning. The first paper is Attention Is All You Need from Google that introduce self-attention mechanism and Transformer to NLP tasks. The motivation for the Google folks is that RNN

Running ARM FVP with Linux

This post begins with a goal to compile and run some assembly programs in AA64 on ARMv8-A architecture for students in a microcontroller course. My prior solution to this was failing so I have came up with the idea to run the ARM FVP (Fixed Virtual Platform) with Linux and

C++ STL notes

Vector 1. reserve() and resize() are different 1. reserve() change the capacity of the vector can hold, has no effect on the current items/does not change the size of the vector 1. Need to use push_back() to add elements instead of using [] access 2. resize() change the actual

OpenGL Learning Note

Before all 1. OpenGL Site 2. OpenGL Wiki OpenGL drawing process 1. Read obj file from vertices vector 2. Vertex Buffer Object (VBO) 1. Generate vertex buffer object 2. Copy the vertices vector to the VBO (reside in GPU memory) 1. First need to bind the buffer using glBindBuffer 2.

SST Learning Note

Before all Key interfaces 1. SST::Component: basic class for simulation objects 2. SST::Event: interactions between components or recurring events like component clock ticks 1. Event class can have handler with form of void func(Event* ev). The handler will invoke a given member function whenever triggered. 2. The

FPGA Log #01: Include header files for simulation

After finish writing my SystemVerilog code and being able to compile them successfully in Quartus Prime, I decided to launch the simulation to verify whether my design was sound. However, during the simulation compilation phase, I encounter the following issue: # vlog -reportprogress 300 -sv -work work "+incdir+/home/user/

FPGA Log #00: Setting up the environment

I currently have a Altera Cyclone IV FPGA development board and since it is winter break, I decided to code some SystemVerilog modules on it to refresh my RTL design skill. Install Quartus Prime Lite First we need to install the Quartus Prime Lite, which is a free to use