Chips and Cheese
AMD’s CDNA 3 Compute Architecture
#ChipAndCheese

Telegraph | source
(author: clamchowder)

Telegraph

AMD’s CDNA 3 Compute Architecture

AMD has a long history of vying for GPU compute market share. Ever since Nvidia got first dibs with their Tesla architecture, AMD has been playing catch up. Terascale 3 moved from VLIW5 to VLIW4 to improve execution unit utilization in compute workloads.…

Chips and Cheese
Cortex A57, Nintendo Switch’s CPU
#ChipAndCheese

Telegraph | source
(author: clamchowder)

Telegraph

Cortex A57, Nintendo Switch’s CPU

In the early 2010s, Arm’s 32-bit cores had established themselves in cell phones and tablets. But rising memory capacities in those devices meant Arm would have to go 64-bit sooner or later. On top of that, Arm had server ambitions – and 64-bit support was…

Daniel Lemire's blog
Measuring the size of the cache line empirically

Telegraph | source

Telegraph

Measuring the size of the cache line empirically

Our computers do not read or write memory in units of bits or even bytes. Rather memory is accessed in small blocks of memory called “cache line”. For a given system, the cache line size is usually fixed and small (e.g.,  16 to 256 bytes). All Intel/AMD x64…

#今日看了什么
https://dl.acm.org/doi/10.1145/3613424.3614272

ACM Conferences

Clockhands: Rename-free Instruction Set Architecture for Out-of-order Processors | Proceedings of the 56th Annual IEEE/ACM International…

今天是真的A350（上次被换了777）
以及南航这一排双发宽体看上去还是挺震撼的

#推荐书籍
有英文电子版

https://item.m.jd.com/product/13634267.html

Daniel Lemire's blog
Fast Buffer-to-String conversion in JavaScript with a Lookup Table

Telegraph | source

Telegraph

Fast Buffer-to-String conversion in JavaScript with a Lookup…

When programming in a JavaScript environment such as Node.js, you might recover raw data from the network and need to convert the bytes into strings. In a system such as Node.js, you may represent such raw bytes using a Buffer instance. You can conveniently…

杰哥的{运维，编程，调板子}小笔记

VIPT 与缓存大小和页表大小的关系

VIPT（Virtual Index Physical Tag）是 L1 数据缓存常用的技术，利用了虚拟地址和物理地址的 Index 相同的特性，得以优化 L1 数据缓存的读取。但是 VIPT 的使用，与页表大小和 L1 数据缓存大小都有关系。这篇博客探讨一下，VIPT 技术背后的一些问题。

source

杰哥的{运维，编程，调板子}小笔记

反向代理的 Partial Transfer 问题

反向代理已经是无处不在，但是如果反向代理没有根据使用场景调优，或者出现了一些异常，可能会带来不好的用户体验，并且现象十分奇怪，例如访问某 GitLab 实例的时候，偶尔会出现页面加载不完整的情况。

这些问题困扰了我们很久，到最后才发现，原来问题在反向代理上。下面就来回顾一下事情的经过。

source

杰哥的{运维，编程，调板子}小笔记

包管理器打包命令速查

随着 Linux 使用逐渐深入，开始尝试参与到一些发行版/包管理器的维护当中。在此记录一下打包相关命令，方便自己速查。

source

Daniel Lemire's blog
How fast can you validate UTF-8 strings in JavaScript?

Telegraph | source

Telegraph

How fast can you validate UTF-8 strings in JavaScript?

When you recover textual content from the disk or from the network, you may expect it to be a Unicode string in UTF-8. It is the most common format. Unfortunately, not all sequences of bytes are valid UTF-8 and accepting invalid UTF-8 without validating it…

MiaoTony's Box

#今天又看了啥 #paper #arxiv #academic #tools arxiv-extensions Add two links to arxiv page: HTML version of arxiv & AI chat for arxiv - html version from ar5iv.org - html version from arxiv-vanity.com - AI Chat from arxiw.org https://greasyfork.org/zh-CN/scripts/481023…

#今天又看了啥 #paper #arxiv #academic #tools
trick:
将任意 arxiv.org 论文页面的域名中的
- x -> 5 (ar5iv.org), Convert from TeX to HTML with LaTeXML
- v -> w (arxiw.org), AI Chat with paper

#今天又看了啥 #paper #arxiv #academic #tools
arxiv-extensions
Add two links to arxiv page: HTML version of arxiv & AI chat for arxiv
- html version from ar5iv.org
- html version from arxiv-vanity.com
- AI Chat from arxiw.org

https://greasyfork.org/zh-CN/scripts/481023-arxiv-extensions

greasyfork.org

arxiv-extensions

HTML version of arxiv & AI chat for arxiv

#今日看了什么
https://github.com/zeroasiccorp/switchboard

GitHub

GitHub - zeroasiccorp/switchboard: Communication framework for RTL simulation and emulation.

Communication framework for RTL simulation and emulation. - zeroasiccorp/switchboard

Chips and Cheese
GCN, AMD’s GPU Architecture Modernization
#ChipAndCheese

Telegraph | source
(author: clamchowder)

Telegraph

GCN, AMD’s GPU Architecture Modernization

AMD’s Terascale architecture became very competitive as it matured with the HD 5000 and 6000 series. But using GPUs for general purpose compute was trending around the 2010s and AMD didn’t want to miss out. Terascale’s SIMD engines were distant relatives…

#今日看了什么
https://community.arm.com/arm-community-blogs/b/infrastructure-solutions-blog/posts/arm-neoverse-v1-top-down-methodology-for-performance-analysis

Arm

Neoverse V1 Performance Analysis whitepaper

The Arm Neoverse V1 Performance Analysis Methodology whitepaper is now available to help you optimize your application code for V1-based production systems.

#今日看了什么
https://dl.acm.org/doi/10.1145/3624062.3624233

ACM Other conferences

An Empirical Comparison of the RISC-V and AArch64 Instruction Sets | Proceedings of the SC '23 Workshops of the International Conference…

#今日看了什么
https://github.com/llvm/llvm-project/pull/70762

GitHub

[clang] Add support for new loop attribute [[clang::code_align()]] by smanna12 · Pull Request #70762 · llvm/llvm-project

This patch adds support for new loop attribute: [[clang::code_align(N)]].
This attribute applies to a loop and specifies the byte alignment for a loop.
The attribute accepts a positive integer cons...

Arch Linux: Recent news updates
Bugtracker migration to GitLab completed

We are happy to announce that the migration of the bugtracker to GitLab is done! 🥳

Thanks to everyone who has helped during the migration!

This means the issue tracker and merge requests on the GitLab package repos are now enabled.

The old bugtracker will subsequently be closed down. For archiving reasons there will be a static copy so that links (for example the randomly picked Task #56716) are still stable, migrated bugs have a closing comment pointing to the new URL on GitLab.

Packaging bugs are now opened on the repo hosting the corresponding packaging sources, the "Add a new Bug" button on the package page on archlinux.org will automatically direct you to the correct place to open the issue. The workflow afterwards is mostly the same, first our Bug Wranglers will have a look at the issues and triage them, and then they will be handed over to the respective Package Maintainers to fix. A list of all issues can be found here.

If you do not have an account for GitLab already (which authenticates against our SSO service), please write us a mail with your desired username to accountsupport@archlinux.org as advised in the banner.

source
(author: Christian Heusel)