71K views · Aug 17, 2023 · 1:35:12 · 1 HN recommender · 4 mentions · discussed 2023–2025
Chandler Carruth's CppNow 2023 talk on modernizing compiler design for the Carbon toolchain, using data-oriented design and modern hardware assumptions. It covers lexing, parsing and compact data structures aimed at compiler performance.
Why Hacker News points people here: Recommended as a recent take on rethinking compiler architecture, and cited in discussions of flattened ASTs and fast lexers.
“I recommend a watch of this for a recent take on trying to change how we think of these.”HN comment, 2023
“I recommend a watch of this for a recent take on trying to change how we think of these.”
Recent take on rethinking compiler architecture, an open research field.
No but it is endemic to how we think about compilers up to now, and especially when the rust compiler arch was massively developed, more than 10 years ago. I recommend a watch of this for a recent take on trying to change how we think of these. It is an open research field this video
Cache the position of quotes/apostrophes, etc. Optimizing for your edge case would require everyone writing and reading code to conform to this extra thing, which seems completely unnecessary. Machines are pretty fast. this video Modernizing Compiler Design for Carbon Toolchain - Chandler Carruth - CppNow 2023
“Chandler discusses it in this video though!”
Chandler Carruth discusses flattened AST layout and its benefits.
Chandler discusses it in this video though! this video You get some traversals for free with this layout (preorder, reverse post order). Can search for subtrees with string searching algorithms or more complex things with regex.
“In this talk about the Carbon language, Chandler Carruth shows and explains some goals/challenges regarding performance”
Carbon talk on compiler performance goals and challenges.
Is lexing ever a bottleneck though? Even if you push for lexing and parsing 10M lines/second [1], I'd argue that semantic analysis and codegen (for AOT-compiled languages) will dominate the timings. That said, there's no reason not to squeeze every bit of performance out of it! [1]: In this talk about the Carbon language, Chandler Carruth shows and explains some goals/challenges regarding performance: this video
