Microarch Club - 0: Hello World

Episode Date: February 12, 2024

In this introductory episode, I detail my mission and goals for the podcast. Subscribe for new episodes every other Wednesday!(00:00) - Intro (03:07) - Wrap Up Show Notes:"A Three Year Bet on Chip Design"More: https://microarch.club/episodes/0

Hey folks, my name is Dan Mangum, and this is a short trailer for a new podcast I'm launching, The MicroArch Club. I've already recorded the first handful of episodes, and I am planning on a release cadence of every other Wednesday, starting with the first episode this Wednesday. I wanted to record this trailer for two reasons. First, it is a low-cost way to test out some of the podcast infrastructure I have set up before releasing the first full episode. But second, and most important, I wanted to talk about the mission for this show and why I think adding to the already heavily saturated world of podcasts is worthwhile. Last summer, I put out a blog post on DanielMangum.com

entitled A Three-Year Bet on Chip Design. I'll link it in the show notes. It is a short post, and there's nothing incredibly remarkable about it, but I essentially describe how I plan to develop deeper expertise in processor design by the time I turn 30 years old. I had just turned 27. My background is in software, and my current day job is primarily as a software engineer, but I've always been interested in hardware, and I spend a lot of my free time working on digital logic design and studying processor architecture. As part of this goal, I started reaching out to

folks, seeing if they would be willing to talk to me about their experience in the industry. This included individuals who are currently doing impactful work in the field, as well as those who had made large contributions in the past. I wanted to know what it looked like to go from nothing to a microprocessor and gain a better understanding of how the lessons of the past inform how we architect processors today. One thing that became immediately clear to me was that processor design is as much an art as it is a science. When making design decisions, the importance of sampling common workloads and evaluating benchmarks is critically important, but understanding how a set of design decisions are going to interact and how that will impact the success of the product requires iteration, creativity, and a clear vision. Another thing that became readily apparent to me in these conversations was that many

of the attributes of processors today are linked to concepts developed through academic research and product development from half a century ago. Many folks working in the industry today reference designs from the past that were either successful or not as motivation for decisions in their modern designs. Sometimes I would have the privilege of talking to one of these people, then talking to the person who developed the concept or process that inspired them. This gave me a unique view into the sequence of events that caused us to arrive where we are today, and that insight is not something you can get without talking to folks who are in the decision-making room. Throughout my career in engineering, I have always been a big proponent of writing design

documents. For folks who may not be familiar with the process, design documents are a practice of reaching consensus on engineering architecture decisions by authoring technical descriptions then going through a review and ratification exercise. They not only are effective at ensuring that multiple people are on the same page, but also provide a detailed history of decisions that have been made and motivation for them, which is an invaluable resource when making future decisions. There have been a few times where I've joined an organization

that has not previously written design documents, and as part of the process of starting the practice, we have written a set of retroactive documents. I think that is a pretty close approximation of what I am hoping the MicroArch Club will be, writing the retroactive design documents for processor design. In order to do so, we are going to need to talk to a lot of folks who were in or are in the room where decisions are made. If you're interested in hearing about the art, science, and history of processor design, come join me on this journey.