Clad — Automatic Differentiation Using Clang and LLVM

Vassilev, Vassil; Vassilev, Martin; Penev, Alexander; Moneta, L.; Ilieva, V

doi:10.1088/1742-6596/608/1/012055

Cited by 15 publications

(13 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Кроме того, это налагает определённые ограничения на используемый компилятор и затрудняет перенос на GPU. Хотя применение инструментов на основе source-to-source преобразований исходного кода при помощи специального компилятора выглядит перспективным направлением с практической точки зрения [100] и позволяет надеяться на внедрение HMC в расчётные системы в будущем. Кроме того, в работах [94] и [96] отмечается, что при наличии высокочастотных карт нормалей или карт смещений (имитация микрорельефа поверхности) дифференцирование не будет корректно работать.…”

Section: типы динамических системunclassified

The current state of the methods for calculating global illumination in tasks of realistic computer graphics

Frolov¹,

Voloboi²,

Ершов³

et al. 2021

Proceedings of ISP RAS

View full text Add to dashboard Cite

Modern realistic computer graphics are based on light transport simulation. In this case, one of the main and difficult to calculate tasks is to calculate the global illumination, i.e. distribution of light in a virtual scene, taking into account multiple reflections and scattering of light and all kinds of its interaction with objects in the scene. Hundreds of publications and describing dozens of methods are devoted to this problem. In this state-of-the-art review, we would like not only to list and briefly describe these methods, but also to give some “map” of existing works, which will allow the reader to navigate, understand their advantages and disadvantages, and, thereby, choose a right method for themselves. Particular attention is paid to such characteristics of the methods as robustness and universality in relation to the used mathematical models, the transparency of the method verification, the possibility of efficient implementation on the GPU, as well as restrictions imposed on the scene or illumination phenomena. In contrast to the existing survey papers, not only the efficiency of the methods is analyzed, but also their limitations and the complexity of software implementation. In addition, we provide the results of our own numerical experiments with various methods that serve as illustrations for the conclusions.

show abstract

Section: типы динамических системunclassified

The current state of the methods for calculating global illumination in tasks of realistic computer graphics

Frolov¹,

Voloboi²,

Ершов³

et al. 2021

Proceedings of ISP RAS

View full text Add to dashboard Cite

show abstract

“…Automatic differentiation in ROOT is based on Clad [5,6]. Clad is a source transformation AD tool for C++.…”

Section: Architecture and Implementationmentioning

confidence: 99%

“…Clad operates on Clang AST (abstract syntax tree) by analyzing the AST of the original function and generating the AST of the derivative. Clad provides two API functions: clad::differentiate for forward mode and clad::gradient for reverse mode, which can be used directly in the source code to mark a function for differentiation (see [6] for more details on usage and code examples).…”

Section: Architecture and Implementationmentioning

confidence: 99%

See 1 more Smart Citation

Automatic Differentiation in ROOT

2020

View full text Add to dashboard Cite

In mathematics and computer algebra, automatic differentiation (AD) is a set of techniques to evaluate the derivative of a function specified by a computer program. AD exploits the fact that every computer program, no matter how complicated, executes a sequence of elementary arithmetic operations (addition, subtraction, multiplication, division, etc.), elementary functions (exp, log, sin, cos, etc.) and control flow statements. AD takes source code of a function as input and produces source code of the derived function. By applying the chain rule repeatedly to these operations, derivatives of arbitrary order can be computed automatically, accurately to working precision, and using at most a small constant factor more arithmetic operations than the original program. This paper presents AD techniques available in ROOT, supported by Cling, to produce derivatives of arbitrary C/C++ functions through implementing source code transformation and employing the chain rule of differential calculus in both forward mode and reverse mode. We explain its current integration for gradient computation in TFormula. We demonstrate the correctness and performance improvements in ROOT’s fitting algorithms.

show abstract

“…There is also a burgeoning research activity in AD for domain-specific and emerging languages. Examples include autodiff for Halide (Ragan-Kelley et al, 2013), JuliaDiff for Julia (Bezanson, Edelman, Karpinski, & Shah, 2017), and Clad as AD tool using Clang and LLVM (Vassilev, Vassilev, Penev, Moneta, & Ilieva, 2015). Tangent (van Merriënboer, Wiltschko, & Moldovan, 2017) can be used for the AD of code generated with the scripting language Python.…”

Section: High-level Implementations Of Admentioning

confidence: 99%

An introduction to algorithmic differentiation

Gebremedhin

Walther

2019

WIREs Data Min & Knowl

View full text Add to dashboard Cite

Algorithmic differentiation (AD), also known as automatic differentiation, is a technology for accurate and efficient evaluation of derivatives of a function given as a computer model. The evaluations of such models are essential building blocks in numerous scientific computing and data analysis applications, including optimization, parameter identification, sensitivity analysis, uncertainty quantification, nonlinear equation solving, and integration of differential equations. We provide an introduction to AD and present its basic ideas and techniques, some of its most important results, the implementation paradigms it relies on, the connection it has to other domains including machine learning and parallel computing, and a few of the major open problems in the area. Topics we discuss include: forward mode and reverse mode of AD, higher‐order derivatives, operator overloading and source transformation, sparsity exploitation, checkpointing, cross‐country mode, and differentiating iterative processes. This article is categorized under: Algorithmic Development > Scalable Statistical Methods Technologies > Data Preprocessing

show abstract

Clad — Automatic Differentiation Using Clang and LLVM

Cited by 15 publications

References 12 publications

The current state of the methods for calculating global illumination in tasks of realistic computer graphics

The current state of the methods for calculating global illumination in tasks of realistic computer graphics

Automatic Differentiation in ROOT

An introduction to algorithmic differentiation

Contact Info

Product

Resources

About