2
2

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 5 years have passed since last update.

Data.Vector型をリストのようにパターンマッチさせる

Last updated at Posted at 2014-09-22

Data.Vector型をリストのようにパターンマッチできないかなと調べていた時に、2ちゃんねるのプログラム技術板に以下のようなコメントを目にした。

スクリーンショット 2014-09-22 11.49.21.png

このヒントに基づいて、以下のように、データ列をData.Vector.UnboxedとしたIIRフィルタを実装してみた。が、Cで書いたものよりも、2000倍!ほど遅い。ただならぬ遅さである。データ列をData.Array.Unboxedにして実装してみたら、Cの約3倍ほど遅いが、まあ使える。
しかしVector型のものはどうしてこんなに遅いのだろう??要検討

{-# LANGUAGE PatternSynonyms #-}
{-# LANGUAGE ViewPatterns #-}

iir :: UV.Vector Double -> UV.Vector Double -> UV.Vector Double -> UV.Vector Double
iir b a x = y
  where
    y = iir' b a w x
    w = UV.replicate mn 0
    m = UV.length b
    n = UV.length a
    mn = max m n

iir' :: UV.Vector Double -> UV.Vector Double -> UV.Vector Double -> UV.Vector Double -> UV.Vector Double
iir' b a w v = do
  case v of
    Nils
      -> UV.fromList []
    Cons x xs
      -> UV.cons y (iir' b a w' xs)
      where y = UV.sum $ UV.zipWith (*) b w'
            w0 = x - (UV.sum $ UV.zipWith (*) (UV.tail a) (UV.init w))
            w' = UV.cons w0 (UV.init w)


uncons :: UV.Unbox a => UV.Vector a -> Maybe (a, UV.Vector a)
uncons v = do
  case UV.null v of
    False -> Just (UV.head v, UV.tail v)
    True  -> Nothing

pattern Cons x xs <- (uncons -> Just (x, xs))
pattern Nils <- (uncons -> Nothing)


2
2
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
2
2

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?