Featured
Elementary Functions and NOT Following the IEEE 754 Floating-Point Standard
Automated Approximation: The Three Step Problem
PhD Summary
The Problem With Abstractions and Libraries
Ruminations on software complexity
16-bit Floating-point Error and Activation Function Tolerance
Sometimes Activation Functions Only Vaguely Matter
Improving the Performance of TensorFlow Activation Functions
What can we learn from GPU Frame Captures: Stardew Valley
What can we learn from GPU Frame Captures: Europa Universalis 4
C++17: The new problem with 'auto'
When Approximations Are More Accurate And Better Performing. Part 2
When Approximations Are More Accurate And Better Performing. Part 1
UNDERSTANDING ERROR AND APPROXIMATION: 7. Relaxation and Regularisation
What if we want Hash Collisions?
When can we approximate results?
Storing Nearby Answers in a Function Level Cache
Defining Approximate Type Systems To Allow for Controlled Precision
Approximate methods for handling error in multi-threaded solutions
Bjarne Stroustrup: No Littering
Build time - Run time distribution in Clang
Updated Shader Analyser With AMD IL support