Qemu 是一個模擬不同 Instruction Set 很好用的工具，鑑於網路上對於 Qemu 的教學都非常的分散，本議程將整理講者研究 Dynamic Binary Translation (DBT) 的心得，並分享 Qemu 如何實作該技術。此議程將帶領聽眾入門 Dynamic Binary Translation，並涵蓋如何 Translate Binary to QEMU TCG-IR、Compile TCG-IR to another ISA on host machine、Handle block chaining、Emulate behavior of source program (e.g. code execution, handle trap, interrupt, system call, etc)。

** 不會涵蓋 Full System Emulation **

• Video on YouTube
• Slides
• Collaborative notes on HackMD

LLVM 是一個十分模組化的編譯器系統，提供了自有的中間表示式，我們可以輕易地在 LLVM 新增一個指令集架構。RISC-V 則是目前最火紅的開放指令集架構，各式各樣的 toolchain 與 OS 都正在如火如荼的移植中，當然 LLVM 也不例外，以 30 個 patch 為基礎教你如何移植 RISC-V 到 LLVM 後端，不只是 Compiler，其中還會包含一些有關 Assembler、Disassembler 和 Linker 的基本知識，帶著大家一路從 LLVM 中間表示式到最終生成 binary 檔，一步一步揭開編譯器神秘的面紗。

• Video on YouTube
• Slides
• Collaborative notes on HackMD

While the Intel x86-64 architecture is undisputedly market leader in the
server space, several vendors have started introducing ARM64 boards.
This presentation examines the suitability of ARM64 server boards for
network servers. In particular, we look at the workload of a
moderate-size ccTLD DNS zone (.dk) and how it would perform on ARM64
running FreeBSD.

We consider the viability of the ARM64 platform from performance and
performance/power perspectives. While ARM64 is definitely slower than
Intel on many workloads, it performs at least as well or better than
Intel on workloads that are interesting to the internet community.
Notably DNS is a very appropriate workload for ARM64.

• Video on YouTube
• Slides
• Collaborative notes on HackMD

What is nds32
Prerequisite to port an arch to Linux mainline
What should you port for your arch
Rebase to the latest kernel codes
Developing cycle
Ready to be merged to linux-next
Get your signed
Get an account of kernel.org
Send the pull request to Linus

• Video on YouTube
• Slides
• Collaborative notes on HackMD

Hundreds of AI chips are releasing in the near future, the latest figures indicate 34 IC and IP vendors will provide various AI chips and deep learning accelerator (DLA) ASICs in 2018. These all reflect the urgent need for an open compiler to support different AI chips.

We foresaw the trend and developed the compiler ONNC. Based on ONNX, ONNC is an efficient way to connect all current AI chips, especially DLA ASICs, with ONNX.

Open Neural Network Exchange Format (ONNX) is a standard for representing deep learning models that enables models to be transferred between frameworks. We introduce ONNC that supports ONNX format and mainstream AI frameworks such as Caffe and Tensorflow. ONNC’s dominant advantage to current AI frameworks is that it provides direct support to DLA ASIC chips by ability to describe variants of performance cost models of hardware and by general optimization passes. DLA ASIC vendors can reuse these optimization passes by describing its special performance cost model in ONNC. ONNX and ONNC together help DLA ASIC vendors support various AI frameworks within a short time, improves DLA’s performance and shortens developing schedule.

ONNX aims to guarantee interoperatability for all tools, and ONNC further guarantees executability for all devices.

• Collaborative notes on HackMD

原先只是大學課堂的一個點子，卻因為它的開放理念而成為世界關注的新型態指令集架構，在 2017 年先後完成初步的移植，進入 FreeBSD 與 Linux 的上游，RISC-V 就此橫空出世。
講者 Alan (高魁良) 來自晶心科技，是國內站在這一浪潮尖端的半導體 IP 公司的軟體工程師。他將帶來的是第一手的移植經驗：從 ftrace 與 perf 的技術本身到 Linux 核心貢獻的眉眉角角，再從與上游維護者的互動到公司面對開源系統的愛恨情仇；小小的 commit 卻大有文章。

• Video on YouTube
• Slides
• Collaborative notes on HackMD

所有 CPU 等級裝置都包含了一個尚未被喚醒的功能, 向量暫存器, 目前全世界所有的開源只有 OpenCV 達到 30% 的使用率, 本次演講將會針對 Intel SSE/AVX 及 Arm NEON 兩大平台, 如何超越 compiler O3 等級的優化 (release mode), 發揮向量暫存器 100% 的極限, 將會比現存的 O3 等級優化再快 4 倍, 對於 AI 人工智慧, blockchain 演算法類, 將會是重要的里程碑

• Video on YouTube
• Slides
• Collaborative notes on HackMD

隨著 Web Application 的快速發展，WebAssembly 組合語言的規格也被提出，讓網站可以用組合語言增加更多的可能性。
我覺得在網站上和桌面能有共通的組合語言是很酷的事情
想和大家分享 WebAssembly ，以及我自幹 WebAssembly 虛擬機器時的發現和心得

• Video on YouTube
• Slides
• Collaborative notes on HackMD