Personal Notes

• Collectives with Zero-SMs — Oct 02, 2026
• CUTE + NCCL Python Bindings for All! — Sep 23, 2026
• Producer-Consumer Warp Specialization for TMA Operations — Sep 08, 2026
• Hyena Models, Convolutions, & Signal Processing — Sep 07, 2026
• NCCL's Device API — Aug 04, 2026
• FFT & Long Convolutions — Jul 28, 2026
• Notes on FlashMoE's Kernel Design — Jul 26, 2026
• The NVSHMEM API — Jul 20, 2026
• Some Interesting Multi-GPU Problems Pt. 1 — Jul 17, 2026
• My Personal Favorite FlashAttention (+ KV-Caching) Visualization — Jul 15, 2026
• Annotated SSMs — Jul 11, 2026
• Explaining DeepEP Dispatch (Tourist's Guide to Expert-Parallel Libraries, Pt. 1) — Jun 30, 2026
• Demystifying NCCL — Jun 23, 2026