https://github.com/igor-240340/float32avrtest
https://github.com/igor-240340/float32avrtest
Last synced: about 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/igor-240340/float32avrtest
- Owner: igor-240340
- Created: 2024-07-13T11:48:59.000Z (10 months ago)
- Default Branch: en
- Last Pushed: 2024-11-27T18:00:17.000Z (6 months ago)
- Last Synced: 2025-02-04T22:29:56.762Z (3 months ago)
- Language: C++
- Size: 144 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Float32AVRTest
Original Russian version of README.md is [here](https://github.com/igor-240340/Float32AVRTest/tree/dev) in dev branch.
This repository contains reference examples of floating-point calculations. The results from these examples will be compared against a custom software floating-point emulation for AVR from [this repo](https://github.com/igor-240340/Float32AVR).
The idea is to use the hardware implementation of IEEE 754 on a desktop as the reference. We calculate reference results and compare them with those returned by our custom software emulation.
The numbering of examples is cross-referenced: example number 1 in this repository corresponds to example number 1 in Float32AVR and to example number 1 in the mind map analyzing possible cases (the mind map files with extension `*.xmind` can be found in docs directory in the repo of the main project Hardware Calculator From Scratch [here](https://github.com/igor-240340/HardwareCalculatorFromScratch/tree/en/docs/Basic%20Research%20Notes)).
One caveat: Float32AVR does not support `inf`, `nan`, or denormalized numbers. Therefore:
- For examples where the hardware floating-point implementation returns `inf` or `nan`, Float32AVR will throw an exception.
- For cases where a denormalized number is returned, Float32AVR will return zero.