Supercompilation by evaluation

Jones, Simon Peyton

doi:10.1145/1863523.1863540

Cited by 19 publications

(8 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The backward mode (< and ) resembles weakest-precondition reasoning (Dijkstra 1975;Nori et al 2014), pre-image computation (Toronto, McCarthy, and Horn 2015), and constraint propagation (Saraswat, Rinard, and Panangaden 1991;Gupta, Jagadeesan, and Panangaden 1999). The forward mode ( > and ) resembles lazy evaluation (Launchbury 1993), in particular lazy partial evaluation (Jørgensen 1992;Fischer et al 2008;Mitchell 2010;Bolingbroke and Peyton Jones 2010). Our laziness postpones nondeterminism (Fischer, Kiselyov, and Shan 2011) in the measure monad, an extension of the probability monad (Giry 1982;Ramsey and Pfeffer 2002).…”

Section: Related Workmentioning

confidence: 99%

Exact Bayesian inference by symbolic disintegration

ShanChung-chieh¹,

RamseyNorman²

2017

SIGPLAN Not.

View full text Add to dashboard Cite

Bayesian inference, of posterior knowledge from prior knowledge and observed evidence, is typically defined by Bayes's rule, which says the posterior multiplied by the probability of an observation equals a joint probability. But the observation of a continuous quantity usually has probability zero, in which case Bayes's rule says only that the unknown times zero is zero. To infer a posterior distribution from a zero-probability observation, the statistical notion of disintegration tells us to specify the observation as an expression rather than a predicate, but does not tell us how to compute the posterior. We present the first method of computing a disintegration from a probabilistic program and an expression of a quantity to be observed, even when the observation has probability zero. Because the method produces an exact posterior term and preserves a semantics in which monadic terms denote measures, it composes with other inference methods in a modular way-without sacrificing accuracy or performance.

show abstract

Section: Related Workmentioning

confidence: 99%

Exact Bayesian inference by symbolic disintegration

ShanChung-chieh¹,

RamseyNorman²

2017

SIGPLAN Not.

View full text Add to dashboard Cite

show abstract

“…We suffer from the same problem that Bolingbroke and Peyton Jones [3] report: the supercompiler can sometimes prevent GHC from applying other important optimizations such as unboxing of arithmetics. Since that is a slightly different algorithm it is not exactly the same programs that get staggering increases in memory allocations.…”

Section: Measurementsmentioning

confidence: 99%

“…Avoiding name capture can cause significant churn and make the supercompiler dreadfully slow. Bolingbroke and Peyton Jones [3] report that for a particular example 42% of their supercompilation time is spent on managing names and renaming.…”

Section: Name Capturementioning

confidence: 99%

See 1 more Smart Citation

Taming code explosion in supercompilation

Jönsson

Nordlander

2011

Proceedings of the 20th ACM SIGPLAN Workshop on Partial Evaluation and Program Manipulation

View full text Add to dashboard Cite

Supercompilation algorithms can perform great optimizations but sometimes suffer from the problem of code explosion. This results in huge binaries which might hurt the performance on a modern processor. We present a supercompilation algorithm that is fast enough to speculatively supercompile expressions and discard the result if it turned out bad. This allows us to supercompile large parts of the imaginary and spectral parts of nofib in a matter of seconds while keeping the binary size increase below 5%.

show abstract

“…As mentioned in §5.2 we must manually unfold loops over regions and rectangles as GHC avoids inlining the definitions of recursive functions. The nice way to fix this would be some form of supercompilation [4,16]. Support for supercompilation in GHC is currently being developed, though still in an early stage.…”

Section: Manual Unwinding Of Recursive Functionsmentioning

confidence: 99%

Efficient parallel stencil convolution in Haskell

LippmeierBen¹,

Keller²

2011

SIGPLAN Not.

View full text Add to dashboard Cite

Stencil convolution is a fundamental building block of many scientific and image processing algorithms. We present a declarative approach to writing such convolutions in Haskell that is both efficient at runtime and implicitly parallel. To achieve this we extend our prior work on the Repa array library with two new features: partitioned and cursored arrays. Combined with careful management of the interaction between GHC and its back-end code generator LLVM, we achieve performance comparable to the standard OpenCV library.

show abstract

Supercompilation by evaluation

Cited by 19 publications

References 23 publications

Exact Bayesian inference by symbolic disintegration

Exact Bayesian inference by symbolic disintegration

Taming code explosion in supercompilation

Efficient parallel stencil convolution in Haskell

Contact Info

Product

Resources

About