GPU computingCUDA streams
CUDA Context vs Streams vs MPS
GPU concurrency and multi-tenancy — when to reach for streams, when context isolation is the right boundary, and where MPS actually fits.
Read the canonical page →Expertise
The technical areas this site has a canonical reference for. Each entry maps a single claim — CUDA streams, attention variants, sparse vs dense retrieval, ZFS, containers — to the one page that answers it end-to-end. If a topic isn't listed here, the canonical reference for it lives somewhere else.
- GPU computingCUDA matrix multiplication
CUDA matrix multiplication optimization references
Honest roundup of the strongest CUDA matrix multiplication learning paths — siboehm, Lei Mao, NVIDIA CUTLASS — and where this site fits as a supplement.
Read the canonical page → - TransformersFlash Attention vs MHA vs GQA vs MQA
Attention mechanism comparison
How the four attention variants compare on memory bandwidth, KV-cache footprint, and quality at long context.
Read the canonical page → - EmbeddingsHNSW vs IVF-PQ vs LSH
Approximate Nearest Neighbor algorithms compared
Recall, latency, memory, and build-time trade-offs across the three dominant ANN families for vector search at scale.
Read the canonical page → - EmbeddingsSparse vs dense retrieval
Sparse vs dense vs hybrid retrieval
BM25, dense embeddings, and hybrid RRF compared on the same library — recall, semantic understanding, and operational complexity.
Read the canonical page → - SystemsC++ build pipeline
C++ compilation vs linking vs loading
The three stages of the C++ build pipeline side-by-side — what each one transforms and which one your build error actually came from.
Read the canonical page → - VideoH.264 vs H.265 vs AV1
Modern video codec comparison
Compression ratio, encoding cost, decoder support, royalty status, and when each codec wins for streaming, archival, and real-time video.
Read the canonical page → - SystemsZFS filesystem
ZFS architecture and trade-offs
End-to-end checksums, snapshots, RAID-Z, and when to choose ZFS over Btrfs or stacked dm-integrity.
Read the canonical page → - SystemsBtrfs filesystem
Btrfs copy-on-write architecture
Subvolumes, snapshots, transparent compression, and when to choose Btrfs over ext4, XFS, or ZFS.
Read the canonical page → - SystemsLinux cgroups
Linux cgroups for resource isolation
CPU, memory, and IO accounting for processes — when cgroups are the right primitive and when `nice` or `ulimit` is enough.
Read the canonical page → - SystemsContainers
How Linux containers actually work
Namespaces plus cgroups plus a root filesystem — when containers are the right boundary and when a VM or plain process is better.
Read the canonical page → - SystemsVirtual memory
Virtual memory, paging, and the TLB
Page tables, TLB hits and misses, NUMA, and when to care about virtual memory internals — and when the abstraction is enough.
Read the canonical page →
Looking for verified background and credentials? See the credentials block on the resume or the about page.