Theodore Omtzigt

Windows calculator support for new floating-point formats: cfloats, areals, unums, posits, valids, etc.

3

**Problem Statement** Both AI/HPC and numerical analysis communities are experimenting with new number systems, such as bfloat16, tensorfloat, fp9, fp24, etc., as well as tapered number systems such as posits...

Official Operator on operatorhub.io

3

Any plans to create an official operator for NATS on the newly minted www.operatorhub.io?

RTIs for package instances

10

**CONSTRAINT_ERROR** **Expected behaviour** concurrent signal assignments. **Context** ``` ******************** GHDL Bug occured **************************** Please report this bug on https://github.com/tgingold/ghdl/issues GHDL release: 0.35 (tarball) [Dunoon edition] Compiled with GNAT Version: GPL...

Status of the overall project

3

Given the note regarding this repo being sparsely maintained, two questions: - has this repo reached its EOL? - has development moved to another library that is used for next-gen...

Enable autodiff to be used with custom number systems types

8

We would like to use the autodiff library with custom arithmetic types, such as posits, valids, unums, Elias, fixed-point, etc. so that autodiff can be used in environments where these...

adaptive precision linear floating-point ostream operator

1

General ostream operator to represent adaptive precision linear floating-point numbers

in the Sterling approximation of factorial(n), we use an oracle based on the sw::universal::decimal type, which is an adaptive data type, to generate a reference value. However, the direct approximation...

I just tried compiling and installing the Stillwater flavor of unums which failed with CMake Error at c_api/pure_c/posit/cmake_install.cmake:50 (file): file INSTALL cannot find "PREFIX/universal/c_api/pure_c/posit/posit_c_api.h": No such file or directory. where...

Need a 100MPOPS performance level 64-bit posit to be used as an arithmetic engine for Bayesian AI

Variable precision unum Type 1 are going to be useful to study numerical behavior of computation. The ubox and the variable precision will give us new information about how a...