Francesco149

buttery smooth vte/libvte with the 120fps patches

vte/libvte is extremely choppy because it's hard-capped to 40fps. all terminals based on it suffer from this, so I decided to make these quick n dirty patches for various vte versions that bump the cap up to 120fps

here's a 120fps video demostration https://streamable.com/j8i0n

also a 60fps clip on twitter if your browser can't handle 120 https://twitter.com/roriicon/status/1113186464332951559

wilzbach

int foo(int a, int b)
{
  return (a % 3235272) % b;
}

int bar(int a)
{
  return foo(a, 88);
}

nadavrot

High-Performance Matrix Multiplication

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).

Intro

Matrix multiplication is a mathematical operation that defines the product of

xplorld

array -> contiguous subarray with max(f(arr))

套路 - 2 max

从左往右扫，维护 2 个 max,

• 一个是 max(sub(arr[0:n]))
• 一个是 max(sub(arr[0:n], including arr[n]))

for n in arr:

klzgrad

为什么流量可以进行分类

这里的“流量”一般定义为中间人观测到的一组由（时间，方向，包大小）元数据组成的序列 [Cai2014]。其源头是应用层的读写操作，经过传输层协议的变换（分片、协议状态机、加密等），流量序列产生一定变化。但是这种变化非常有限，因为流量的发生过程本质是确定性的，随机因素较小，因此对于特定环境中的特定应用（浏览器访问 google.com）各种流量特征体现出相当大的一致性和独特性，这就使“从流量特征识别应用”的监督学习问题成为可能。虽然有若干不利因素使得确定性下降，例如多层次上软件多版本的排列组合爆炸、有状态的缓存、流水线和连接复用、用户随机行为，但是因为版本的幂律分布、应用层读写操作间的依赖关系、流量特征和检测算法的改进等原因，分类依然具有相当的可行性。

分类的对象：流量应用分类与网站指纹攻击

根据分类的对象产生了两个相近但是不同的研究领域。从流量特征中分类应用类型的被称为流量分类（traffic classification），从流量特征中分类所访问网站或者网页的被称为网站指纹（website fingerprinting）。以机器学习的方法而论前者是比后者更弱但本质相同的一个问题。

这两类攻击的威胁类型不同。流量分类威胁的是可用性，如果GFW检出流量是隧道应用然后进行封锁，则破坏了可用性。而网站指纹威胁的是匿名性和隐私，如果从隐秘流量中检出是谁在访问哪个网站，则破坏了匿名性，丝绸之路就是这样被FBI破获的。

qfgaohao

#include <stdio.h>


void deviceQuery ()
{
  cudaDeviceProp prop;
  int nDevices=0, i;
  cudaError_t ierr;

  ierr = cudaGetDeviceCount(&nDevices);

owulveryck

About

This is a proof of concept of a portable operating system a-la-docker.

It is a single binary that contains an image of an OS in SquashFS and the binary to run it.

a make will build the binary with an embedded alpine image. make havre-xenial will build a binary with an embedded ubuntu image.

What does the binary do?

htfy96

In this article we compared different behavior of static, inline and static inline free functions in compiled binary. All the following test was done under g++ 7.1.1 on Linux amd64, ELF64.

Test sources

header.hpp

#pragma once

inline int only_inline() { return 42; }
static int only_static() { return 42; }

antirez

# LVDB - LLOOGG Memory DB
# Copyriht (C) 2009 Salvatore Sanfilippo <antirez@gmail.com>
# All Rights Reserved

# TODO
# - cron with cleanup of timedout clients, automatic dump
# - the dump should use array startsearch to write it line by line
#   and may just use gets to read element by element and load the whole state.
# - 'help','stopserver','saveandstopserver','save','load','reset','keys' commands.
# - ttl with milliseconds resolution 'ttl a 1000'. Check ttl in dump!

SaveTheRbtz

#!/bin/bash -ue
# Sets up mq+$shed combo for given interface.

if [ $# -ne 2 ] || [ -z "$1" ] || [ -z "$2" ]; then
    echo "usage: $0 interface scheduler" >&2
    echo "  e.g: $0 eth0 fq"             >&2
    echo "  e.g: $0 eth0 pfifo_fast"     >&2
    exit 64  # EX_USAGE
fi