There were no shortage of AI accelerators at Hot Chips 2024. One of those was FuriosaAI’s RNGD, pronounced Renegade. It’s a 150W TDP chip with 48 GB of HBM3, and aims to handle inference at lower power than GPUs. RNGD is physically implemented on a PCIe card…

Mainframes still play a vital role in today, providing extremely high uptime and low latency for financial transactions.

This is our second video interview from Hot Chips 2024 and this time I got to sit down with Susan Eickoff and Christian Jacobi from IBM to talk about their brand new Telum II processor and Spyre AI cards from the development process to the architecture of…

We sometimes have to work a large quantity of floating-point numbers. This volume can be detrimental to performance. Thus we often want to compress these numbers. Large-language models routinely do so. A sensible approach is to convert them to brain floating…

For today’s article we have another video interview for you folks, this time with Arik Gihon from Intel where we talk about Lunar Lake. Before we get into the video and the transcript, I would just like to give a massive thank you to all of you who have donated…

2024/8/30，是一个看似平常的周五下午。Windowd Update 告诉我安装了几个无关痛痒的小更新，考虑到已经快两个月没有重启电脑了，我就顺手点了重启。万万没想到，这是一个长达 25 小时的噩梦的开始。 TL;DR：Hyper-V 虚拟交换机的未知异常导致虚拟机无法正常枚举 PCI 设备，因此 WSL2 无法启动。解决方法是从注册表中删除虚拟所有交换机并重启。 问题起源 我的工作站是两年前配的 13900K + NVMe，当然重启很快，然而 WSL 却没一直能起来，表现为 Windows Terminal…

Ampere Computing focuses on developing cloud-native processors, and aims to serve many users while providing security, privacy, and consistent performance, without relying on process node scaling. Features that increase performance at the expense of consistency…

Last year at Hot Chips 2023, Tesla introduced their Dojo supercomputer. For Tesla, machine learning is focused on automotive applications like self driving cars. Training deals with video, which can demand a lot of IO bandwidth. As an example, the size of…