@dosyago/rainsum

1.1.1 • Public • Published

Rain

This repository houses the Rainbow and Rainstorm hash functions, developed by Cris at DOSYAGO and licensed under Apache-2.0. The 64-bit variants have passed all tests in the SMHasher3 suite. Results can be found in the results/ subdirectory.

The code contains a reference implementation in C++, a port to WASM, and a Makefile for building everything.

Algorithm Speed Hash Size Purpose Core Mixing Function Security
Rainbow 5.79 GiB/sec 64 to 256 bits General-purpose non-cryptographic hashing Multiplication, subtraction/addition, rotation, XOR Not designed for cryptographic security
Rainstorm 1.91 GiB/sec (at 4 rounds, tuneable) 64 to 512 bits Potential cryptographic hashing Addition/subtraction, rotation, XOR No formal security analysis yet

Usage options

JavaScript channel:

  1. As a Node.JS library, via:
npm i @dosyago/rainsum@latest

Then you can use as:

import {rainbowHash, rainstormHash} from '@dosyago/rainsum';

const seed = 0x0;
const hash = await rainbowHash(256, seed, 'Hello there!');
const intendedCryptoHash = await rainstormHash(512, seed, Buffer.from('Hello there!'));
  1. As a NPM global binary, via:
npm i -g @dosyago/rainsum@latest
jsrsum --help

Native channel:

First, make everything:

make clean && make
sudo make install # if you wish 

Then you can use either:

  1. As object code, or a C library or by using the C++ files in ./src/, specifically importing "tool.h" and using "rainsum.cpp"

  2. As a command line tool:

rainsum file.txt
# will output a hash using the default values: Algorithm: rainbow, Hash size in bits: 256
# see --help for other arguments
rainsum --help
Usage: rainsum [OPTIONS] [INFILE]
Calculate a Rainbow or Rainstorm hash.

Options:
  -m, --mode [digest|stream]        Specifies the mode, where:
                                    digest mode (the default) gives a fixed length hash in hex, or
                                    stream mode gives a variable length binary feedback output
  -a, --algorithm [bow|storm]       Specify the hash algorithm to use. Default: storm
  -s, --size [64-256|64-512]        Specify the bit size of the hash. Default: 256
  -o, --output-file FILE            Output file for the hash or stream
  -t, --test-vectors                Calculate the hash of the standard test vectors
  -l, --output-length HASHES        Set the output length in hash iterations (stream only)
  -v, --version                     Print out the version
  --seed                            Seed value (64-bit number or string). If string is used,
                                    it is hashed with Rainstorm to a 64-bit number

Note: The jsrsum NPM binary (which uses wasm, and is installable via npm i -g @dosyago/rainsum) offers the same API and command-line options as the native C++ binary, but is slower.

Table of Contents

Assets

This repository produces:

  • rainsum CLI tool and rainstorm.o and rainbow.o object files
  • wasm binary
  • JavaScript CLI tool (API compat with C++ binary, but slower)
  • NPM package with jsrsum global (the JavaScript CLI tool), plus an importable ES Module library exporting async rainbowHash(hashSize, seed, input) and async rainstormHash(hashSize, seed, input) functions taking inputs as string or Buffer, backed by the wasm binary.
  • A collection of scripts:
    • ./scripts/bench.mjs: benchmark the speed between the JS WASM and C++ implementations
    • ./scripts/verify.mjs: verify the correctness of the implementations
    • And other scripts.

Building

Build with make:

make clean && make

Optionall to install rainsum into an executable path location, run:

make install

which may require sudo privileges.

Benchmark

Rain hash functions C++ vs Node/WASM benchmark:

Test Input & Size (bytes)         Run        C++ Version       WASM Version          Fastest
input1 (10 bytes)                  1         7,310,292 ns    122,337,000 ns     16.00x (C++ wins!)
input2 (100 bytes)                 1         5,250,875 ns    113,674,500 ns     21.00x (C++ wins!)
input3 (1,000 bytes)               1         5,341,625 ns    112,402,958 ns     21.00x (C++ wins!)
input4 (10,000 bytes)              1         4,912,209 ns    113,795,750 ns     23.00x (C++ wins!)
input5 (100,000 bytes)             1         5,247,208 ns    112,086,708 ns     21.00x (C++ wins!)
input6 (1,000,000 bytes)           1         5,717,125 ns    112,697,792 ns     19.00x (C++ wins!)
input7 (10,000,000 bytes)          1        10,700,834 ns    119,352,417 ns     11.00x (C++ wins!)
input8 (100,000,000 bytes)         1        52,803,417 ns    191,939,500 ns      3.00x (C++ wins!)
input1 (10 bytes)                  2         3,671,250 ns    107,211,375 ns     29.00x (C++ wins!)
input2 (100 bytes)                 2         5,072,792 ns    112,407,625 ns     22.00x (C++ wins!)
input3 (1,000 bytes)               2         4,668,833 ns    111,602,583 ns     23.00x (C++ wins!)
input4 (10,000 bytes)              2         4,660,792 ns    112,008,292 ns     24.00x (C++ wins!)
input5 (100,000 bytes)             2         4,675,875 ns    112,914,708 ns     24.00x (C++ wins!)
input6 (1,000,000 bytes)           2         5,334,500 ns    114,160,917 ns     21.00x (C++ wins!)
input7 (10,000,000 bytes)          2        11,232,792 ns    119,210,000 ns     10.00x (C++ wins!)
input8 (100,000,000 bytes)         2        52,097,042 ns    181,990,375 ns      3.00x (C++ wins!)
input1 (10 bytes)                  3         3,726,791 ns    106,394,125 ns     28.00x (C++ wins!)
input2 (100 bytes)                 3         4,990,500 ns    113,719,291 ns     22.00x (C++ wins!)
input3 (1,000 bytes)               3         4,882,792 ns    113,991,917 ns     23.00x (C++ wins!)
input4 (10,000 bytes)              3         5,054,458 ns    112,881,500 ns     22.00x (C++ wins!)
input5 (100,000 bytes)             3         5,180,958 ns    112,406,250 ns     21.00x (C++ wins!)
input6 (1,000,000 bytes)           3         5,779,833 ns    113,486,834 ns     19.00x (C++ wins!)
input7 (10,000,000 bytes)          3        10,747,875 ns    119,093,209 ns     11.00x (C++ wins!)
input8 (100,000,000 bytes)         3        52,637,666 ns    170,915,917 ns      3.00x (C++ wins!)

Repo structure

  .
  |-- LICENSE.txt
  |-- Makefile
  |-- PAPER.md
  |-- README.md
  |-- docs
  |   |-- app.js
  |   |-- index.html
  |   |-- rain.cjs
  |   `-- rain.wasm
  |-- js
  |   |-- lib
  |   |   `-- api.mjs
  |   |-- package-lock.json
  |   |-- package.json
  |   |-- rainsum.mjs
  |   |-- test.mjs
  |   `-- wasm
  |       |-- rain.cjs
  |       `-- rain.wasm
  |-- rain
  |   |-- bin
  |   |   `-- rainsum
  |   `-- obj
  |       |-- rainbow.d
  |       |-- rainbow.o
  |       |-- rainstorm.d
  |       |-- rainstorm.o
  |       |-- rainsum.d
  |       `-- rainsum.o
  |-- rainsum -> rain/bin/rainsum
  |-- results
  |   |-- dieharder
  |   |   |-- README.md
  |   |   |-- rainbow-256.txt
  |   |   |-- rainbow-64-infinite.txt
  |   |   |-- rainstorm-256.txt
  |   |   `-- rainstorm-64-infinite.txt
  |   `-- smhasher3
  |       |-- rainbow-128.txt
  |       |-- rainbow-256.txt
  |       |-- rainbow.txt
  |       |-- rainstorm-128.txt
  |       |-- rainstorm-256.txt
  |       `-- rainstorm.txt
  |-- scripts
  |   |-- 1srain.sh
  |   |-- bench.mjs
  |   |-- blockchain.sh
  |   |-- build.sh
  |   |-- chain.mjs
  |   |-- debug.sh
  |   |-- package-lock.json
  |   |-- package.json
  |   |-- testjs.sh
  |   |-- vectors.sh
  |   `-- verify.sh
  |-- src
  |   |-- common.h
  |   |-- cxxopts.hpp
  |   |-- rainbow.cpp
  |   |-- rainstorm.cpp
  |   |-- rainsum.cpp
  |   `-- tool.h
  `-- verification
      `-- vectors.txt
14 directories, 52 files

Rainbow

Rainbow is a fast hash function. It's intended for general-purpose, non-cryptographic hashing. The core mixing function utilizes multiplication, subtraction/addition, rotation, and XOR.

Rainstorm - Unvetted for Security

Rainstorm is a slower hash function with a tunable-round feature. It's designed with cryptographic hashing in mind, but it hasn't been formally analyzed for security, so we provide no guarantees. The core mixing function uses addition/subtraction, rotation, and XOR.

Rainstorm's round number is adjustable, potentially offering additional security. However, please note that this is hypothetical until rigorous security analysis is completed.

Note on Cryptographic Intent

While Rainstorm's design reflects cryptographic hashing principles, it has not been formally analyzed and thus, cannot be considered 'secure.' We strongly encourage those interested to conduct an analysis and offer feedback.

Great! I see that the program rainsum.cpp is for calculating a Rainbow or Rainstorm hash. The software can operate in two modes, "digest" or "stream", which affects how it outputs the hash. It also allows the user to select different hash algorithms and sizes, specify the input and output files, use predefined test vectors, and set the output length in hashes.

Genesis

The fundamental concept for the mixing functions derived from Discohash, but has been significantly developed and extended. The overall architecture and processing flow of the hash were inspired by existing hash functions.

License

This repository and content is licensed under Apache-2.0 unless otherwise noted. It's copyright © Cris and The Dosyago Corporation 2023. All rights reserved.

Rainsum Field Manual

1. Introduction

Rainsum is a powerful command-line utility for calculating Rainbow or Rainstorm hashes of input data. This tool can operate in two modes: "digest" or "stream". In "digest" mode, Rainsum outputs a fixed-length hash in hex, while in "stream" mode it produces variable-length binary output.

Rainsum also offers multiple hashing algorithms, sizes, and various configurable options to cater to a wide range of use cases.

JavaScript WASM Version

There is also a JavaScript WASM version, consistent with the C++ version, and 8 - 16 times slower on small and medium inputs (100 bytes to 10MiB), and 2 - 3 times slower on large inputs (100MiB and up), at js/rainsum.mjs. This JavaScript version of rainsum can be used mostly like the C++ version, so the below guide and instrutions suffice essentially for both.

2. Basic Usage

2.1 Command Structure

The basic command structure of Rainsum is as follows:

rainsum [OPTIONS] [INFILE]

Here, [OPTIONS] is a list of options that modify the behavior of Rainsum and [INFILE] is the input file to hash. If no file is specified, Rainsum reads from standard input.

2.2 Options

Here are the options that you can use with Rainsum:

  • -m, --mode [digest|stream]: Specifies the mode. Default is digest.
  • -a, --algorithm [bow|storm]: Specify the hash algorithm to use. Default is storm.
  • -s, --size [64-256|64-512]: Specify the bit size of the hash. Default is 256.
  • -o, --output-file FILE: Specifies the output file for the hash or stream.
  • -t, --test-vectors: Calculates the hash of the standard test vectors.
  • -l, --output-length HASHES: Sets the output length in hash iterations (stream only).
  • --seed: Seed value (64-bit number or string). If a string is used, it is hashed with Rainstorm to a 64-bit number.
  • -h, --help: Prints usage information.
  • -v, --version: Prints out the version of the software.

3. Modes of Operation

3.1 Digest Mode

In digest mode, Rainsum calculates a fixed-length hash of the input data and outputs the result in hexadecimal form.

For example, to calculate a 256-bit Rainstorm hash of input.txt and output it to output.txt, you would use:

rainsum -m digest -a storm -s 256 -o output.txt input.txt

3.2 Stream Mode

In stream mode, Rainsum calculates a hash of the input data and then uses that hash as input to the next iteration of the hash function, repeating this process for a specified number of iterations. The result is a stream of binary data.

For example, to generate a 512-bit Rainstorm hash stream of input.txt for 1000000 iterations and output it to output.txt, you would use:

rainsum -m stream -a storm -s 512 -l 1000000 -o output.txt input.txt

4. Hash Algorithms and Sizes

Rainsum supports the following hash algorithms:

  • bow: Rainbow hash
  • storm: Rainstorm hash

And these sizes (in bits):

  • Rainbow: 64, 128, 256
  • Rainstorm: 64, 128, 256, 512

5. Test Vectors

Rainsum includes a set of predefined test vectors that you can use to verify the correct operation of the hash functions. To use these test vectors, include the -t or --test-vectors option in your command.

6. Seed Values

You can provide a seed value for the hash function using the --seed option followed by a 64-bit number or a string. If a string is used, Rainsum will hash it with Rainstorm to generate a 64-bit number.

7. Help and Version Information

Use -h or --help to print usage information. Use -v or --version to print the version of the software.

8. Compilation

Rainsum is written in C++. Make sure to have a modern C++ compiler and appropriate libraries (like filesystem, iostream, etc.) installed to compile the code. A makefile or build system setup might be required depending on your specific project configuration.

9. Conclusion

Rainsum provides a powerful and flexible command-line interface for calculating Rainbow and Rainstorm hashes. Whether you're working on a small project or a large-scale system, Rainsum offers the features and options you need to get the job done.

Developer Information

Stability

The hashes' stability may change over time, as we might modify constants, mixing specifics, and more as we gather insights. Should such changes alter the hashes' output, we will denote the changes with new version numbers.

Test vectors

The current test vectors for Rainstorm and Rainbow are:

Rainbow v1.0.4 Test Vectors

./rainsum -a bow --test-vectors:

b735f3165b474cf1a824a63ba18c7d087353e778b6d38bd1c26f7b027c6980d9 ""
c7343ac7ee1e4990b55227b0182c41e9a6bbc295a17e2194d4e0081124657c3c "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789"
53efdb8f046dba30523e9004fce7d194fdf6c79a59d6e3687907652e38e53123 "The quick brown fox jumps over the lazy dog"
95a3515641473aa3726bcc5f454c658bfc9b714736f3ffa8b347807775c2078e "The quick brown fox jumps over the lazy cog"
f27c10f32ae243afea08dfb15e0c86c0b601792d1cd195ca651fe5394c56f200 "The quick brown fox jumps over the lazy dog."
e21780122142956ff99d560069a123b75d014f0b110d307d9b23d79f58ebeb29 "After the rainstorm comes the rainbow."
a46a9e5cba400ed3e1deec852fb0667e8acbbcfeb71cf0f3a1901396aaae6e19 "@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@"

Rainstorm v0.0.5 Test Vectors

./rainsum --test-vectors:

e3ea5f8885f7bb16468d08c578f0e7cc15febd31c27e323a79ef87c35756ce1e ""
9e07ce365903116b62ac3ac0a033167853853074313f443d5b372f0225eede50 "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789"
f88600f4b65211a95c6817d0840e0fc2d422883ddf310f29fa8d4cbfda962626 "The quick brown fox jumps over the lazy dog"
ec05208dd1fbf47b9539a761af723612eaa810762ab7a77b715fcfb3bf44f04a "The quick brown fox jumps over the lazy cog"
822578f80d46184a674a6069486b4594053490de8ddf343cc1706418e527bec8 "The quick brown fox jumps over the lazy dog."
410427b981efa6ef884cd1f3d812c880bc7a37abc7450dd62803a4098f28d0f1 "After the rainstorm comes the rainbow."
47b5d8cb1df8d81ed23689936d2edaa7bd5c48f5bc463600a4d7a56342ac80b9 "@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@"

Building and Installing

You can build and install the rainsum utility using the provided Makefile. First, clone the repository and change into the root directory of the project:

git clone https://github.com/dosyago/rain
cd rain

Then, build the utility with make:

make

If you want to install rainsum globally, so it can be run from any directory, use the make install command:

make install

This command might require administrator privileges, depending on your system's configuration. If you encounter a permission error, try using sudo:

sudo make install

After installation, you can run rainsum from any directory:

rainsum --help

See the Field Manual for more information on usage.

Contributions

We warmly welcome any analysis, along with faster implementations or suggested modifications. Collaboration is highly encouraged!

Readme

Keywords

Package Sidebar

Install

npm i @dosyago/rainsum

Weekly Downloads

14

Version

1.1.1

License

Apache-2.0

Unpacked Size

205 kB

Total Files

7

Last publish

Collaborators

  • dosy