DerivX

V0.2.7-Beta Build 20210816

© 2021-2021 Xu Rendong. All Rights Reserved.

Project Summary

Derivatives Pricing Engine.

Install

C++:

Python:

pip install derivx

JavaScript:

npm install derivx

Usage

C++:

Python:

import derivx
print(derivx.Version())

For more usages please refer to test files in derivx_py folder.

DerivX is not dependent on numpy, pandas and matplotlib, but if you want to run tests, you'd better install them with:

pip install numpy pandas matplotlib

JavaScript:

const derivx = require('derivx')
console.log(derivx.Version())

For more usages please refer to test files in derivx_js folder.

DerivX is not dependent on numjs, but if you want to run tests, you'd better install it with:

npm install git+https://github.com/nicolaspanel/numjs.git

Benchmark

Environment:

• CPU: Intel Core i7-4790 4C/8T 3.60GHz
• RAM: 16.0GB
• SYS: Windows 10 Pro 64-bit 20H2 19042.1110

Parameters:

• rand_rows: 500000
• rand_cols: 250
• rand_seed: { 0, 1, 2, 3, 4, 5, 6, 7 }
• dual_smooth: true
• runs_size: 1000000
• runs_step: 244
• run_from: 0
• run_days: 1

C++:

• InitRand: 461ms - 550ms
• InitPath: 1575ms - 1961ms

    Autocall Snowball:

• CalcCoupon: 468ms - 564ms (there is still 200ms optimization space)
• CalcPayoff: 965ms - 983ms per price
• CalcDelta: 1932ms - 1983ms per price
• CalcGamma: 2898ms - 2918ms per price
• CalcVega: 2047ms - 2073ms per price
• CalcTheta: 1636ms - 1647ms per price (as run_days increases this tends to CalcPayoff's time consumption)
• CalcRho: 2106ms - 2161ms per price
• The difference between CalcCoupon and CalcPayoff shows that different prices and barriers will cause different time consumption.

Contact Information

QQ: 277195007, WeChat: xrd_ustc, E-mail: xrd@ustc.edu