6 usability improvements in GCC 15 https://lobste.rs/s/bnntrp #compilers
https://developers.redhat.com/articles/2025/04/10/6-usability-improvements-gcc-15

Global Value Numbering Using Random Interpretation https://lobste.rs/s/w65rcu #pdf #compilers
https://www2.eecs.berkeley.edu/Pubs/TechRpts/2003/6359.html

GSoC 2024: The 1001 thresholds in LLVM https://lobste.rs/s/yphejx #compilers #performance
https://blog.llvm.org/posts/2024-08-31-the-1001-thresholds-in-llvm/

Doing the Prospero-Challenge in RPython https://lobste.rs/s/5nedly #compilers #performance #plt
https://www.pypy.org/posts/2025/04/prospero-in-rpython.html

I think one doesn't want to do PRE if any form of if-conversion will happen later.

A surprising enum size optimization in the Rust compiler

https://jpfennell.com/posts/enum-type-size/

Discussions: https://discu.eu/q/https://jpfennell.com/posts/enum-type-size/

How Does C Code Become Assembly https://lobste.rs/s/fe9byf #assembly #c #compilers
https://www.xta0.me/2021/07/13/MIT-6172-5.html

A surprising enum size optimization in the Rust compiler https://lobste.rs/s/w3jjb2 #compilers #rust
https://jpfennell.com/posts/enum-type-size/

"Verified" "Compilation" of "Python" with Knuckledragger, GCC, and Ghidra

https://www.philipzucker.com/knuckle_C_pcode/

Discussions: https://discu.eu/q/https://www.philipzucker.com/knuckle_C_pcode/

Okay as could be predicted turns out the very easy and obvious solution does not work. A friend pointed me to a mailing list discussion where it was noted that QBE, harec's backend, currently lacks support for generating position-independent code (PIC), which is required to produce a shared library for fairly obvious reasons.

If anyone following me happens to know how to work with compiler backends and is inclined to climb the mountain to implement PIC support, you would be saving many distro maintainers many headaches and possibly mitigating a significant amount of future security issues by making it easier to distribute the patches.

"Verified" "Compilation" of "Python" with Knuckledragger, GCC, and Ghidra by @sandmouth https://lobste.rs/s/b9kbg3 #assembly #compilers #formalmethods #python #reversing
https://www.philipzucker.com/knuckle_C_pcode/

Making a Brainf*** JIT Compiler in Rust https://lobste.rs/s/vrbtjw #video #compilers
https://www.youtube.com/watch?v=3aRdIijtGrM

When should a compiler expand garbage collection barriers? https://lobste.rs/s/o2tsdx #compilers
https://robcasloz.github.io/blog/2024/02/14/when-should-a-compiler-expand-garbage-collection-barriers.html

JIT calculators finale https://lobste.rs/s/ri3gqs #assembly #compilers #rust
https://ochagavia.nl/blog/jit-calculators-finale/

Compilers: Incrementally and Extensibly

https://okmij.org/ftp/tagless-final/Compiler/index.html

Compilers: Incrementally and Extensibly https://lobste.rs/s/sbuuzu #compilers #ml #plt
https://okmij.org/ftp/tagless-final/Compiler/index.html

where are the proofs!

https://web.stanford.edu/class/archive/cs/cs143/cs143.1128/

Discussions: https://discu.eu/q/https://web.stanford.edu/class/archive/cs/cs143/cs143.1128/

A Simple Reply https://lobste.rs/s/ofklws #compilers
https://github.com/SeaOfNodes/Simple/blob/main/ASimpleReply.md

This seems a very cool and interesting evolution for a stable and productive language, focused on real needs in the industry. The focus on making sure it all integrates well and works with legacy code is astounding. Very nice. Ocaml future seems great.

--

Making OCaml Safe for Performance Engineering | Lobsters

https://lobste.rs/s/ujvpdn/making_ocaml_safe_for_performance

Making OCaml Safe for Performance Engineering https://lobste.rs/s/ujvpdn #slides #video #compilers #ml #performance
https://youtube.com/watch?v=g3qd4zpm1LA