Historical Writing · 2008 – 2025

Archive

Selected essays on GPU computing, mobile deep learning, MapReduce, search engines, and software engineering recovered from memkite.com, medium.com, amundtveit.com, atbrox.com, amundblog.blogspot.com, and LinkedIn.

2025

2017

2016

2015

Dec 28

Announcing DeepLearningKit — Open Source Deep Learning Framework for Apple's iOS, OS X and tvOS

Launching DeepLearningKit, an open-source deep learning framework written in Swift and Metal for running pretrained Convolutional Neural Networks on iPhone, iPad, Mac, and Apple TV.

medium.com3 min
Sep 09

Swift and Metal GPU Programming on tvOS for the New Apple TV

Running custom Metal compute shaders from Swift on Apple's tvOS simulator and hardware on launch day.

memkite.com5 min
Jun 10

Example of Swift and Metal GPU Programming on OS X 10.11 El Capitan

Porting Swift and Metal GPGPU vector computation to OS X 10.11 El Capitan right after Apple brought Metal from iOS to the Mac.

medium.com2 min
Apr 17

Deeplearning.University — Bibliographies from LISA Labs (Yoshua Bengio's Lab)

Adding curated deep learning bibliographies from Yoshua Bengio's LISA Lab at the University of Montreal to Deeplearning.University.

memkite.com1 min
Feb 01

Example of Sharing Memory Between GPU and CPU with Swift and Metal for iOS 8

Demonstrating zero-copy shared CPU/GPU memory buffers with Swift and Apple Metal on iOS 8 devices — an early precursor to unified memory on Apple Silicon.

medium.com2 min

2014

Dec 18

GPGPU Performance of Swift/Metal vs Accelerate on iPhone 6 & 5S, iPad Air and iPad Mini

Benchmarking Swift + Metal GPU compute against Apple's Accelerate framework and pure Swift for Sigmoid evaluation on iPhone 5S, iPhone 6, iPad Air, and iPad Mini.

memkite.com2 min
Dec 15

Data-Parallel Programming with Metal and Swift for iPhone/iPad GPU

An 8-step walkthrough for writing general-purpose GPU compute kernels in Metal Shading Language and invoking them from Swift on iOS 8.

memkite.com3 min
Sep 19

Recent Experiences from Being on Hacker News' First Page with Startup Blog Posts

Lessons and traffic observations from landing multiple Memkite technical blog posts on the Hacker News front page.

medium.com1 min
Apr 22

Mobile Eats the Cloud

Why 64-bit mobile CPUs, mobile GPUs, and fast local flash storage shift computation back from the cloud to on-device supercomputers.

memkite.com3 min
Apr 01

Technical Feasibility of Building Hitchhiker's Guide to the Galaxy (Offline Web Search)

Back-of-the-envelope storage, memory, and RRAM feasibility analysis of fitting offline web search (Wikipedia, books, and web indexes) onto a pocket device.

memkite.com6 min

2011

2010

2009

2008

Dec 03

Cinpy — Writing and JIT-Compiling C Inside Python

Using Cinpy and TinyCC to embed and compile C functions directly inside Python with ctypes for dramatic inner-loop speedups.

amundblog.blogspot.com2 min
Nov 29

Tools for Accelerating Python

A 2008 field guide to speeding up Python programs — from parallelizing with Hadoop, MPI, and Stackless to JIT/compilers (Psyco, Cython, Shedskin) and native C/Fortran/Assembly extensions.

amundblog.blogspot.com1 min
Oct 18

Example of Microsimulation with Stackless Python

Simulating 100,000 concurrent robot agents playing musical chairs on a 10,000×10,000 grid using Stackless Python microthreads (tasklets).

amundblog.blogspot.com2 min
Mar 21

Test-Driven Parsing with Python, Dparser and Doctest

Building GLR grammars incrementally in Python using Dparser paired with inline doctest specifications.

amundblog.blogspot.com1 min
Jan 13

Future Number of Programming Languages — Singularity or Infinity?

Why monolithic do-everything languages collapse under complexity, and why small, Python-like declarative domain-specific languages (DSLs) are the better path.

amundblog.blogspot.com2 min
Jan 06

Increase Automation of Test-Driven Development?

Early 2008 thoughts on automating both sides of the TDD loop with program synthesis, machine learning, and automated corner-case test generation.

amundblog.blogspot.com2 min