Published inAI Mind·Aug 24🔥 FibyTree vs. 🐍 Set and SortedSetCouple of days ago, I wrote an article about FibyTree, a data structure I designed and implemented in Mojo. Today I would like to reveal some benchmarks. Words of caution. This benchmarks are preliminary. At current point in time Mojo code runs only in a Jupyter Notebook, so I run…Python9 min readPython9 min read
Published inAI Mind·Aug 19A high level introduction to FibyTreeFibyTree is an optionally fully balanced and complete implicit binary search tree, which has the same semantics as a sorted set. I did a reference implementation of this data structure in Mojo, but fear not if you have no idea about Mojo, this article is very high level. …Data Structures6 min readData Structures6 min read
May 28Simple CSV parser in MojoParsing 400MB per second with less than 80 lines of 🔥 — So what is so hard about parsing CSV, isn’t it just .split("\n") and .split(",")? It can be, if you don’t have to comply with the RFC 4180. RFC 4180 states, that a field can be placed between quotes (") and then it is legal to have new lines and commas…Csv7 min readCsv7 min read
May 18Counting chars with SIMD in MojoMojo is a very young (actually a work in progress) programming language designed and developed by a new company called Modular. Here is a blurb from their website: Mojo combines the usability of Python with the performance of C, unlocking unparalleled programmability of AI hardware and extensibility of AI models. …Mojo9 min readMojo9 min read
Published inNerd For Tech·Jun 29, 2021Progressive indexEnable random accessibility for a stream of asymmetric data points — When we store multiple data points together in one stream, we need to make sure that the stream is deconstruct-able. A trivial case being symmetric data points. What do I mean by symmetric? I mean that the size (in bytes) of each data point is equal. If we have a…Data Serialization6 min readData Serialization6 min read
Dec 15, 2020Data structure vs. Data baseDoes your mobile app really need a DB? — Data bases are essential when it comes to multi user applications. They help us store huge amount of data generated by, or related to, huge amount of users. They let us query for a sub set of data, relevant to our business cases. …Mobile App Development3 min readMobile App Development3 min read
Jul 24, 2020Packaging 93K LevelsA story of custom binary serialization for a game made with Unity3D — This year I helped a customer develop and ship a word puzzle game for iOS and Android. One of the challenges in this project was, the game should ship localized in 31 languages. This means every language gets its own set of levels. The customer wanted to start with 3K…Unity4 min readUnity4 min read
Published inThe Startup·Jun 8, 2020Binary Search vs. Eytzinger OrderYou probably know what binary search is but WTF is Eytzinger order? You might have heard of a data structure called binary heap. …Computer Science4 min readComputer Science4 min read
May 10, 2020Why don’t we use DTOs in FlatBuffersDTOs (Data Transfer Objects) are classes which people use to represent and store data before they serialise and send the data over the wire. So the primarily purpose of a DTO is to be able to turn an object graph into a buffer (array of bytes). [Application state] --> [DTOs]…Flatbuffers4 min readFlatbuffers4 min read
Apr 25, 2020Stop using JSON in productionIf you need a self describing object notation? Check out CBOR (Concise Binary Object Representation): CBOR Edit descriptioncbor.me Paste your precious JSON into the playground and see how much bytes you are wasting with JSON. If you are using JSON + JSON-Schema? You are really doing something wrong. The main benefit of JSON is that the payload is self describing. The main benefit of Schemas is that it…Json1 min readJson1 min read
