al-folio

Breaking the OOM Barrier: High-Rank Multi-Dimensional Sparse Einsum via Compressed Sparse Fiber (CSF) on GPU

19 min read   ·   October 28, 2026

2026   ·   cuda   einsum   sparse-tensors   csf   high-performance-computing   ·   machine-learning   hpc

High-Performance Context Management in Yaucl

An in-depth analysis of the single-threaded adaptive Persistent LinkedHashMap view implemented in the yaucl library.

20 min read   ·   October 13, 2026

2026   ·   compilers   pl-theory   c++   performance   data-structures   ·   algorithms-and-data-structures

Deep-Dive Tutorial: Programmatic Control Flow (`prim::If` & `prim::Loop`) in LibTorch JIT C++

22 min read   ·   October 10, 2026

2026   ·   libtorch   pytorch   torchscript   jit   ·   machine-learning   c++

LibTorch for Memory Predictability and Runtime Safety: Enter MetaTensor

11 min read   ·   September 29, 2026

2026   ·   tensorflow   compilers   openxla   cpp   libtorch   distributed

Expressing Tensorflow 2.x operations in plain Mathematical Notation

Documenting some of the Tensorflow 2.x operations using Mathematical notation

12 min read   ·   September 29, 2026

2026   ·   Tensorflow,   TensorAlgebra,   LinearAlgebra   ·   logds