Functional and Constraint Logic Programming

Question

1. What contributions have the authors mentioned in the paper "Memoizing a monadic mixin dsl" ?

2. How can combinators introduce their own monad transformers?

3. What is the common way to obtain a code generator?

4. What is the function that assigns the value of e to v?

Accepted Answer

The authors study the use of mixins in Haskell for a modular DSL for search heuristics used in systematic solvers for combinatorial problems, that generate optimized C++ code from a high-level specification.. The authors show how to apply memoization techniques to tackle performance issues and code explosion due to the highly recursive nature of the mixins.. Memoization of monadic components causes further complications for us to deal with.

Accepted Answer

Through existential types and the monad zipper, all combinators can introduce their own monad transformersto keep their own state throughout the code generation, without affecting any other transformers.

Accepted Answer

A code generator is obtained through mixin composition, simply using (◦):search2 :: Gen Identity search2 = fix (printM ◦ baseM )In the components the authors have seen so far, the monad type parameter m has not been used.

Accepted Answer

The combinator let(v, e, s) introduces a new variable v, initialized to the value of expression e, in the subsearch s, while assign(v, e) assigns the value of e to v and succeeds.

Accepted Answer

The simplest form of a search heuristic is obtained by applying the fixpoint combinator to a base component:fix :: Mixin a → a fix m = m (fix m)search1 :: Gen Identity search1 = fix baseMAdvice Component

Accepted Answer

When running a MST transformer, the enclosing Gen’s height parameter is passed to rStateT , using that as identifier for the layer.

Accepted Answer

newtype MST s m a = MST {runMST :: ReaderT Int (StateT s m) a }For convenience, MST is made an instance of MonadState, so switching from StateT to MST does not require any changes to the code interacting with it.

Accepted Answer

Performance is not affected by the increased number of function calls — perhaps compensated by the improved cache usage of smaller code.

Accepted Answer

For now the authors use Strings as Keys.newtype Key = String newtype Table = Map Key Stmt initMemoState = emptyThe authors capture the two essential operations of MT in a type class, which allows us to lift the operations through other monad transformers.

Accepted Answer

The system was running Ubuntu 10.10 64-bit, with GCC 4.4.4, Gecode 3.3.1 and Minizinc 1.3.1.10 Social golfer problem, CSPlib problem 10 11 Golomb rulers, CSPlib problem 6generated code, typically using unsafePerformIO .

Accepted Answer

Int →MST s m a → m a rStateT s height m =do let action = runReaderT (runMST m) height putCtx height (show s) result ← evalStateT action s clearCtx height return resultSo far the authors have only solved the first performance problem, repeated generation of code.

Accepted Answer

To implement this, the Table type is extended:type MemoContext = Map Int String type Key = (MemoContext ,String)data Table = Table {context :: MemoContext , memoMap :: Map Key Stmt }initMemoState = Table {context = empty , memoMap = empty }MemoContext is represented as a map from integers to strings.

Accepted Answer

|While Expr Stmt | ...A number of convenient abbreviations facilitate building this AST, e.g.,(#) = liftM ◦ (;) if ′ = liftM2 ◦ IfThenElseIn their modular code generation approach, AST fragments from different sources are combined.

Accepted Answer

The authors preserve the sharing obtained through memoization in the backend, by depositing the memoized code fragment in a C++ function that is called from multiple sites.

Accepted Answer

The C++ AST for a search heuristic is produced by a code generator:data Gen m = Gen {initG :: m Stmt , bodyG :: m Stmt , addG :: m Stmt , tryG :: m Stmt , resultG :: m Stmt , failG :: m Stmt , height ::

Functional and Constraint Logic Programming

Chat with Paper

AI Agents for this Paper

Most frequently asked questions

1. What contributions have the authors mentioned in the paper "Memoizing a monadic mixin dsl" ?

2. How can combinators introduce their own monad transformers?

3. What is the common way to obtain a code generator?

4. What is the function that assigns the value of e to v?

5. What is the simplest form of a search heuristic?

6. What is the enclosing Gen’s height parameter?

7. What is the simplest way to add a new type of MST to a search?

8. What is the effect of memoization on performance?

9. What is the simplest way to generate a monad?

10. What is the simplest way to solve the social golfer problem?

11. What is the simplest way to solve the second performance problem?

12. What is the simplest way to implement this?

13. What is the AST that is used to build a heuristic?

14. How do the authors preserve the sharing obtained through memoization in the backend?

15. What is the AST for a search heuristic?

Figures

Citations

A Compiler and Virtual Machine for Constraint-logic Object-oriented Programming with Muli

Painless programming combining reduction and search

Applications of Muli: Solving Practical Problems with Constraint-Logic Object-Oriented Programming

Lifting Term Rewriting Derivations in Constructor Systems by Using Generators.

Painless programming combining reduction and search: design principles for embedding decision procedures in high-level languages

References

MiniZinc: towards a standard CP modelling language

Mixin-based inheritance

Observable Sharing for Functional Circuit Description

Monadic constraint programming

Type-safe observable sharing in Haskell

Related Papers (5)

Constraint Logic Programming.

On Constraint Logic Programming

Cooperative Constraint Functional Logic Programming.

Applications of Constraint Logic Programming.

Higher-Order Logic Programming as Constraint Logic Programming.