chadbrewbaker

FROM amazonlinux:2

RUN yum install -y wget java-11-amazon-corretto openssl
RUN wget https://artifacts.elastic.co/downloads/enterprise-search/enterprise-search-8.2.3.rpm
RUN wget https://artifacts.elastic.co/downloads/elasticsearch/elasticsearch-8.2.3-x86_64.rpm
RUN wget https://artifacts.elastic.co/downloads/kibana/kibana-8.2.3-x86_64.rpm
RUN export JAVA_HOME=/usr/lib/jvm/java-11-amazon-corretto.x86_64
RUN rpm -ivh ./*.rpm
RUN systemctl daemon-reload
RUN systemctl enable elasticsearch.service

chadbrewbaker

Perfchat 26 June 2022

• stress-ng - slides
• perf
• Jim Keller - The Future of Compute
• Introduction to lldb for rust
• std::string performance issues in Chrome

DRAM performance

• row/columns of capacitors that need recharged. Dense but power draw is significant compared to SRAM.
• Rows usually 8kb in size. Moved to row buffers. Request columns from that row to read to CPU.

chadbrewbaker

• Vectorized and performance-portable Quicksort
• SIMDGiraffe: Visualizing SIMD Instructions
• How fast are Linux pipes anyway?
• Using the mold linker for fun and 3x-8x link time speedups.
• Why don't we have a SIMD cookbook?
• SIMD in JDK 17
• Agner Fog optimization manuals

chadbrewbaker

prog = """ 
#include <stdio.h>
#include <stddef.h>
#include <stdint.h>
#include <stdlib.h> 
#include <unistd.h>

  int main() {
  size_t buf_size = 1 << 18; // 256KiB
  char* buf = (char*) malloc(buf_size);

chadbrewbaker

29 May 2022 Perfchat Show Notes

News

• AMD announced Ryzen 7000, Zen4 - https://www.anandtech.com/show/17399/amd-ryzen-7000-announced-zen4-pcie5-ddr5-am5-coming-fall
• Blog article: Compilers may generate branches for straight-line code - https://kristerw.github.io/2022/05/24/branchless/
• Paper: Intel Labs developed a quantum toolchain - https://arxiv.org/abs/2202.11142

Discussion

chadbrewbaker

Perfchat 15 May 2022 Show Notes

News by Denis

• AMD enhanced their Performance Monitoring Unit. https://www.phoronix.com/scan.php?page=news_item&px=AMD-Zen-4-IBS-Linux - better logging of L3 cache misses to help Kernel on page caching.
• rdpmc - a small C++ header-only library to count CPU level events. https://github.com/rodgarrison/rdpmc - x86 only. More lightweight than PAPI
• Tachyum launches their Cloud/AI/HPC chip called Prodigy. https://www.hpcwire.com/off-the-wire/tachyum-launches-prodigy-universal-processor/ - how to run QEMU simulator
• Mysterious memset, a blog post about pointer aliasing preventing compiler optimizations. https://vector-of-bool.github.io/2022/05/11/char8-memset.html

General Disussion

• Mark's blog post

chadbrewbaker

• Brendan Gregg goes to intel
• Guidelines for writing efficient code by Codee
• Augury, a new microarchitectural attack against Apple CPUs
• “Software Engineering at Google” O’Reilly FREE book
• Optimization blueprint for using Large Pages
• New low latency JVM
• Minotaur - new SIMD superoptimizer by Regher's student
• clang-tidy for static performance linting
• windows tracing tools

chadbrewbaker

git clone https://github.com/curl/curl
cd curl
./configure --with-secure-transport
./scripts/coverage.sh
# https://opensource.apple.com/source/Security/Security-59754.80.3/ for Apple native TLS

chadbrewbaker

AWS Pi Day 2022

Registration Link

Twitch Link

# Intelligent data tiering
aws s3api put-bucket-intelligent-tiering-configuration

chadbrewbaker

• Michael Larabel - MichaelLarabel.com
• Phoronix
• Started in 2004 - first test suite published in 2007.
• Other business models like specCPU
• Phoronix lastest ARM server results
• Still getting Linux kernel regressions but more rare.
• Asrock motherboard is very reliable/afordable. Supermicro, Asus, Gigabyte also good to look at.
• Measuring memory copy time to GPU - making sure it isn't confused with GPU processing times.
• What makes a good benchmark? Core count scaling is more important than it was. More than 64 core scaling is still challenging.
• Top compiler/kernel improvements in the past 20 years - propper PCIe power configuration (20 watt improvement) - better optimization configurations for gcc/llvm defaults - Intel's P stat