May 2026 Community Meeting: mojoBLAS and raylib Mojo bindings

In the May community meeting, Shanker (Photon on Discord), a PhD student in theoretical particle physics, walks through his work porting BLAS routines to Mojo: level 1 through level 3 operations, the cache and SIMD optimizations he applied to close the gap with OpenBLAS, and benchmark results showing Mojo BLAS pulling ahead at larger vector sizes. Vladimir (Kivy Code) follows with a close look at his Raylib bindings for Mojo, covering the multi-layer architecture that keeps unsafe FFI calls away from the end user, the ARM ABI edge cases that cost him days of debugging, and the auto-generation approach Raylib's own parser makes possible. He closes with a live particle physics demo. The Q&A touches on consumer GPU support across AMD RDNA, Apple Silicon, and NVIDIA Blackwell SM120, how MAX compares to vLLM for local and scaled inference, HuggingFace model architecture support, and the future of where clauses in parametric structs in Mojo. 00:55 MojoBLAS 25:35 Raylib Mojo bindings 45:24 Q&A Resources Mojo BLAS: https://forum.modular.com/t/mojoblas-... Raylib bindings for Mojo: https://forum.modular.com/t/mojo-rayl... Join our community 🤝 Forum - https://forum.modular.com/ GitHub - https://github.com/modular X (aka Twitter) - https://x.com/modular LinkedIn -   / modular-ai   Reddit -   / modularai   #gpu #programming #ml