typerep-map step by step

Date: July 11, 2018
Author: Veronika Romashkina

Recently, I have been working on a very interesting and sophisticated project called typerep-map. A lot of advanced features and tricks were used during the development process and I have discovered many amusing and new sides of Haskell. So, I decided to share the ideas, steps, issues, etc. in this blog post.

If you want to skip all the funny parts, here is the link to the code itself:

What it’s all about🔗

The basic idea behind typerep-map is to have a data structure like Map, but where types serve as keys, and values stored in the map are of the type specified in the corresponding key.

An example image of this structure:

Int: 42
Bool: True
String: "Haskell"

There can be only one key-value pair for each type.

And here is an example written in pseudo-code for better understanding:

insert (42 :: Int) empty = [(Int, 42)]
                             ^     ^
                            key   value

insert (42 :: Int) (insert True empty) = [(Int, 42), (Bool, True)]
                                           ^     ^      ^     ^
                                          key   value  key  value

We also want our values to be indexed by a polymorphic type, but that will be explained later.

Existing Solutions🔗

There already exist some libraries that implement ideas similar to typerep-map:

  • type-map appears to resemble our project, however the interface is different. They track the elements in the types and don’t provide the desired parametrization.
  • dependent-map is closer to our goal in terms of the interface but the package has a complete reimplementation of Data.Map.Lazy, and the goal of the typerep-map project is to have an efficient implementation based on primitive unboxed arrays.

Motivation🔗

You might wonder what typerep-map brings to the table if there are other packages that aim to fulfil the same purpose. The primary goal is to use it in the caps library instead of the DMap type from dependent-map parametrized by TypeRep.

In caps the performance of lookups is extremely important so it makes sense to prioritize its performance above that of other functions.

Implementing TypeRepMap🔗

Sections below describe the details of the implementation phases and the general concepts.

NOTE: in this blog post I am talking about ghc-8.0.2 or higher.

Extensions used🔗

The code snippets in this blog post assume that the following extensions are enabled:

  • -XTypeApplications
  • -XScopedTypeVariables
  • -XGADTs
  • -XTypeInType
  • -XAllowAmbiguousTypes

Map-based implementation🔗

The reference implementation is more or less straightforward. It uses a lazy Map from containers as an internal representation of the desired data type.

Keys🔗

Normally, types in Haskell are only present at compile-time: after type-checking, they are completely erased from the program. And yet we want to use types as keys in a map. This requires a runtime representation for types. Luckily, the standard library provides a runtime representation for types in the form of TypeRep. But there are actually two different definitions of TypeRep in base:

The one in Type.Reflection was introduced in GHC 8.2 and the old one was rewritten to be based on it. Type.Reflection.TypeRep has kind TypeRep :: k -> * while the old one has kind TypeRep :: *.

To have the basic idea of what actually TypeRep is, you can think of the old TypeRep as an infinite ADT with all types enumerated as tag constructors:

data TypeRep = Int | Bool | Char | String | ...

and the new TypeRep is an equivalent to the infinite GADT:

 data TypeRep (a :: k) where
     Int    :: TypeRep Int
     Bool   :: TypeRep Bool
     Char   :: TypeRep Char
     String :: TypeRep String
     ...

If you are interested in the actual difference between old and new versions of the TypeRep and motivation for this change, here is a nice ICFP video by Simon Peyton Jones:

I use the old TypeRep that comes from Data.Typeable. And I have an explanation for that: there is a limitation in regular Map that all keys must be of the same type and this is not possible to achieve with parameterized TypeRep. Also, the old TypeRep will never be deprecated (from 8.2 it is just a different interface to the new TypeRep, so it’s not obsolete), and it is sufficient for our goal to support older GHC versions.

Here is a usage example of basic TypeRep interface:

ghci> :t typeRep
typeRep :: Typeable a => proxy a -> TypeRep

ghci> typeRep (Proxy @Int)
Int

ghci> :t it
it :: TypeRep

Values🔗

For the first prototype, I decided to use Dynamic as values in our TypeRepMap.

ghci> :t toDyn
toDyn :: Typeable a => a -> Dynamic

ghci> toDyn True
<<Bool>>

ghci> fromDynamic (toDyn "Haskell") :: Maybe String
Just "Haskell"

So we’ve got:

newtype TypeRepMap = TypeRepMap { unMap :: Map TypeRep Dynamic }

and the initial interface looks like this:

insert :: forall a . Typeable a => a -> TypeRepMap -> TypeRepMap
insert val = TypeRepMap . LMap.insert (typeRep (Proxy @a)) (toDyn val) . unMap

lookup :: forall a . Typeable a => TypeRepMap -> Maybe a
lookup = fromDynamic <=< LMap.lookup (typeRep $ Proxy @a) . unMap

When looking at the Dynamic data type implementation

data Dynamic = Dynamic TypeRep Obj
type Obj = Any

you can notice that it already stores TypeRep inside, so it seems like it’s a bit suboptimal decision due to redundancy. And we can safely use Any as our value type.

According to the Dynamic implementation, we can use unsafeCoerce function for the conversion to Any and from Any.

So we get:

newtype TypeRepMap = TypeRepMap { unMap :: LMap.Map TypeRep Any }

insert val =
      TypeRepMap
    . LMap.insert (typeRep $ Proxy @a) (unsafeCoerce val)
    . unMap

lookup = fmap unsafeCoerce . LMap.lookup (typeRep $ Proxy @a) . unMap

Let’s check how it’s all working:

ghci> let x = lookup $ insert (11 :: Int) empty

ghci> x :: Maybe Int
Just 11

ghci> x :: Maybe ()
Nothing

All right, we have a simple working version. But there are ways to improve it.

Parameterization🔗

The next step is to parametrize our data type by type variable f with kind f :: k -> *. This f will be the interpretation of our keys. Such parameterization allows us to encode additional structure common between all elements, making it possible to use TypeRepMap to model a variety of things from extensible records to monadic effects. This sort of parametrization may be familiar to users of vinyl records.

Note that the input kind is k — we want to support arbitrary kinds as well. Since TypeRep is poly-kinded, the interpretation can use any kind for the keys (see some examples in documentation).

newtype TypeRepMap (f :: k -> *) = TypeRepMap
    { unMap :: LMap.Map TypeRep Any
    }

The implementation of the functions stays the same, but the types are different:

insert :: forall a f . Typeable a => f a -> TypeRepMap f -> TypeRepMap f
lookup :: forall a f . Typeable a => TypeRepMap f -> Maybe (f a)

Our previous implementation is just equivalent to TypeRepMap Identity in the context of the described design.

NOTE: Another reason to get rid of the Dynamic: if we keep it then we have to specify Typeable (f a) constraint instead of Typeable a in the type declarations. And having Typeable a constraint would let us implement the following function efficiently:

hoist :: (forall x. f x -> g x) -> TypeRepMap f -> TypeRepMap g

Vector-based implementation🔗

The next step is to write an alternative implementation based on unboxed vectors, which is expected to be faster.

We want to use Vector (TypeRep, Any) instead of our lazy map. This vector is going to be sorted by TypeRep. insert/lookup algorithms should be implemented manually in the following way:

  • insert: allocate a new vector of n + 1 element, copy everything from the initial vector adding the new element and don’t forget to keep the sorting.
  • lookup: the simple binary search.

The point of the unboxed vector is that it helps to get rid of the pointer indirection. If we take just Vector we will observe this picture (Ty stands for TypeRep and El stands for an element):

  [   Pair₁,      Pair₂,      Pair₃,      Pair₄   ]
      /   \       /   \       /   \       /   \
    TyElTyElTyElTyEl

Instead of this what we would like to see is:

    [ Ty₁, El₁, Ty₂, El₂, Ty₃, El₃, Ty₄, El₄ ]

In this way, as the result, the access to the Ty or El is shorter for exactly one pointer dereference.

However, turned out that it’s more efficient to store keys and values in separate vectors under corresponding indices:

    [ Ty₁,   Ty₂,    Ty₃,    Ty₄ ]
       |      |       |       |
    [ El₁,   El₂,    El₃,    El₄ ]

Unfortunately, TypeRep doesn’t have the Unbox instance and it looks like it’s not possible to write it. So instead of storing TypeRep we will be storing a Fingerprint. Basically, Fingerprint is the hash for TypeRep, so it makes sense to move in this direction.

ghci> :t typeRepFingerprint
typeRepFingerprint :: TypeRep -> Fingerprint

ghci> typeRepFingerprint $ typeRep $ Proxy @Int
b1460030427ac0fa458cbf347f168b53

ghci> typeRepFingerprint $ typeRep $ Proxy @Bool
ebf3a8541b05453b8bcac4a38e8b80a4

TypeRep from Data.Typeable module is defined as

type TypeRep = SomeTypeRep

If we take a look at the Ord instance of SomeTypeRep in base we’ll see that it’s confirmed that Fingerprint is unique for each TypeRep. That means it’s okay to use Fingerprint as a key instead of TypeRep.

Vector🔗

This is initial vector-based implementation:

data TypeRepMap (f :: k -> *) = TypeRepMap
    { fingerprints :: Vector Fingerprint
    , anys         :: Vector Any
    }

We want to use unboxed vector as a type for the fingerprints field of TypeRepMap.

Every unboxed vector is the newtype wrapper over some primitive vector. In order to use an unboxed vector of Fingerprint we need to implement an instance of the Prim typeclass from the primitive package for Fingerprint. It was proposed to add this instance under this issue in primitive library (having this instance inside library would simplify implementation a lot):

As the reference for Prim instance implementation, we can use the Storable type class which contains similar functions. There is already the instance Storable for Fingerprint. An assumption is that there is no significant difference between Storable and Prim for our lookup checks and we can use storable vector instead of unboxed one. For more information about the difference between those typeclasses see this SO answer.

Though our initial assumptions were false and turned out that Storable doesn’t give the desired performance boost as shown with benchmarks.

Optimal Vector🔗

According to the source code, Fingerprint is a pair of (Word64, Word64). So instead of having a single vector of Fingerprints we can have a vector of Word64 where Fingerprint with number i stored on 2 * i and 2 * i + 1 indices.

But actually, it’s better to split it into two separate vectors of Word64 where one vector stores the first halves of Fingerprint and the other one stores the second halves correspondingly. It makes the implementation easier and also faster (checked with benchmarks) because of the assumption that it should be almost always enough to compare only the first part and it makes key comparison faster.

After all described optimizations were done our structure took the following form:

data TypeRepMap (f :: k -> *) = TypeRepMap
    { fingerprintAs :: Unboxed.Vector Word64
    , fingerprintBs :: Unboxed.Vector Word64
    , anys          :: Boxed.Vector Any
    }

And the lookup function was implemented like this:

lookup :: forall a f . Typeable a => TypeRepVector f -> Maybe (f a)
lookup tVect =
        fromAny . (anys tVect V.!)
    <$> binarySearch (typeRepFingerprint $ typeRep $ Proxy @a)
                     (fingerprintAs tVect)
                     (fingerprintBs tVect)

It uses a manually implemented version of the binary search algorithm optimized for unboxed vectors. The algorithm initially performs a binary search using the fingerprintAs vector only. And then, after finding the first half, walks through the fingerprintBs vector.

At first, a simple naive binary search was implemented but later it was rewritten into a cache-optimized binary search (see the description here) which boosted the performance significantly.

Array-based implementation🔗

But that’s not all. Later we noticed that every vector has the following definition:

data Vector a = Vector {-# UNPACK #-} !Int
                       {-# UNPACK #-} !Int
                       {-# UNPACK #-} !(Array a)

As you can see it contains two Int fields. So we can make our representation more optimal by using Array instead of boxed vector and PrimArray instead of unboxed vector directly in the TypeRepMap data type.

After all optimizations the final shape of the TypeRepMap is following:

data TypeRepMap (f :: k -> Type) = TypeRepMap
    { fingerprintAs :: {-# UNPACK #-} !(PrimArray Word64)
    , fingerprintBs :: {-# UNPACK #-} !(PrimArray Word64)
    , anys          :: {-# UNPACK #-} !(Array Any)
    }

Benchmarking🔗

Initially, I was frustrated about this part because I had no idea how to create the Map of 1000 elements as that means I needed to somehow generate 1000 types. But there was actually a pretty elegant solution for this puzzle — polymorphic recursion.

Let’s introduce the following data types as type-level natural numbers:

data Z
data S a

Using these data types we can now implement the function which builds TypeRepMap of the desired size.

buildBigMap
    :: forall a . Typeable a
    => Int
    -> Proxy a
    -> TypeRepMap Proxy
    -> TypeRepMap Proxy

so when I run buildBigMap with size n and Proxy a, it calls itself recursively with n - 1 and Proxy (S a) at each step, so the types are growing on each step.

But this wasn’t the only challenge in benchmarking TypeRepMap. There were also a few interesting things with benchmarks to keep in mind:

  • We should force maps to normal form before benchmarking.
  • We can’t use rnf function. Deriving NFData instance for TypeRepMap is not possible because there can be no NFData for Any. We won’t be able to use rnf because it would try to force both the keys and the values, as our values are Any (can’t force them), but since evaluating the values is not important at all for the benchmark, we could try to define a function like rnf but without touching the values.

For Map-based implementation we need to benchmark the lookup function on different depths of our tree (as Map is internally a tree). But the key can be very close to the root so our benchmarks won’t be honest enough. Thus we need to test on different Proxys with different types.

Here is the diagram of how the tree’s been constructed. You can notice that the Char element is the direct child of the root:

size: 16
tree:
   +--Proxy * (S (S (S (S (S (S (S (S (S (S (S (S (S (S Z))))))))))))))
   |
+--Char
|  |
|  |  +--Proxy * (S (S (S (S (S (S Z))))))
|  |  |
|  +--Proxy * (S (S (S (S (S (S (S (S (S (S (S Z)))))))))))
|     |
|     +--|
|
Proxy * (S (S (S (S (S (S (S (S (S (S (S (S (S Z)))))))))))))
|
|     +--Proxy * (S (S (S (S (S (S (S (S (S (S Z))))))))))
|     |
|  +--Proxy * (S (S (S (S (S (S (S (S (S Z)))))))))
|  |  |
|  |  +--Proxy * (S (S (S (S (S (S (S (S Z))))))))
|  |
+--Proxy * (S (S (S (S (S (S (S Z)))))))
   |
   |     +--Proxy * (S Z)
   |     |
   |  +--Proxy * (S (S (S Z)))
   |  |  |
   |  |  +--Proxy * (S (S (S (S (S (S (S (S (S (S (S (S Z))))))))))))
   |  |
   +--Proxy * (S (S (S (S Z))))
      |
      |  +--Proxy * (S (S Z))
      |  |
      +--Proxy * (S (S (S (S (S Z)))))
         |
         +--Proxy * Z

Since we can’t predict how Ord on TypeRep will behave we need to select a Proxy from our range randomly, however, because our types were huge we introduced the following type family to solve that issue:

type family BigProxy (n :: Nat) :: * where
    BigProxy 0 = Z
    BigProxy n = S (BigProxy (n - 1))

While running this version of benchmarks it turned out that rnf function was taking a lot of time mostly on normalisation of the enormous TypeRep keys which consisted of tall nested types like S (S (S ...)).

So, eventually, I end up using the ghc plugin ghc-typelits-knownnat and the type of the buildBigMap became:

buildBigMap :: forall (a :: Nat) . KnownNat a
            => Int
            -> Proxy a
            -> TypeRepMap (Proxy :: Nat -> *)
            -> TypeRepMap (Proxy :: Nat -> *)

In order to benchmark lookup function we implemented a special function fromList to use in place of the bunch of inserts, so we will be able to see the real time measures of lookup operation itself.

data TF f where
    TF :: Typeable a => f a -> TF f

fromList :: [TF f] -> TypeRepMap f

Now the buildBigMap function will have type

buildBigMap :: forall (a :: Nat) . KnownNat a
            => Int
            -> Proxy a
            -> [TF (Proxy :: Nat -> *)]
            -> [TF (Proxy :: Nat -> *)]

Benchmarks make 10 lookups to collect average performance statistics:

tenLookups :: TypeRepMap (Proxy :: Nat -> *)
           -> ( Proxy 10, Proxy 20, Proxy 30, Proxy 40
              , Proxy 50, Proxy 60, Proxy 70, Proxy 80
              )
tenLookups tmap = (lp, lp, lp, lp, lp, lp, lp, lp)
  where
    lp :: forall (a :: Nat) . Typeable a => Proxy a
    lp = fromJust $ lookup tmap

and compare the work of map-based implementation with optimal array-based implementation. Here are the achieved results:

NOTE: time in the report is for 10 lookups. To get the average time of single lookup you need to divide time by 10.

  • Benches GHC-8.4.3
benchmarking map-based/lookup
time                 2.198 μs   (2.195 μs .. 2.202 μs)
                     1.000 R²   (1.000 R² .. 1.000 R²)
mean                 2.196 μs   (2.193 μs .. 2.199 μs)
std dev              10.46 ns   (8.436 ns .. 12.67 ns)

benchmarking dependent map/lookup
time                 819.0 ns   (810.7 ns .. 829.1 ns)
                     0.999 R²   (0.999 R² .. 1.000 R²)
mean                 815.8 ns   (812.1 ns .. 822.5 ns)
std dev              16.11 ns   (9.371 ns .. 23.09 ns)

benchmarking vector-binary-search/lookup
time                 370.7 ns   (368.9 ns .. 372.2 ns)
                     1.000 R²   (1.000 R² .. 1.000 R²)
mean                 368.9 ns   (368.2 ns .. 369.7 ns)
std dev              2.512 ns   (1.938 ns .. 3.474 ns)

benchmarking array-cache-optimized-binary-search/lookup
time                 183.5 ns   (183.2 ns .. 183.8 ns)
                     1.000 R²   (1.000 R² .. 1.000 R²)
mean                 183.6 ns   (183.3 ns .. 184.4 ns)
std dev              1.535 ns   (958.3 ps .. 2.631 ns)

Conclusion🔗

In this blog post, I wanted to show the difficulties, tricks, and useful information which I personally learned during the implementation of an optimized version of TypeRepMap. Also, I needed to somehow structure the knowledge I’ve gained while working on this project. You can say that some parts of the post can be skipped or might be irrelevant but I wrote it in such a way on purpose to highlight the topics that I find very hard to find and understand quickly. So I hope you too will find this knowledge useful!

Acknowledgments🔗

Many thanks to Vladislav Zavialov (@int-index) for mentoring this project! It was the great experience for me.

Bonus🔗

A few more challenges on the way to the release typerep-map:

KindOf🔗

During interface enhancement I ran into some weird issue described below.

It’s nice to have the member function and it makes sense to implement it using already written lookup function:

member :: forall a f . Typeable a => TypeRepMap f -> Bool
member trMap = case lookup @a trMap of
    Nothing -> False
    Just _  -> True

Type of the lookup function is the following:

lookup :: forall a f . Typeable a => TypeRepMap f -> Maybe (f a)

Unfortunately, this implementation of member doesn’t compile! The problem is in the fact that the compiler can’t infer that type variable a and the argument to f have the same kind. These two functions have the following type with implicitly inferred kinds:

lookup :: forall {k} (a :: k) (f :: k -> *) . Typeable a
       => TypeRepMap f -> Maybe (f a)
member :: forall {k1} {k2} (a :: k1) (f :: k2 -> *) . Typeable a
       => TypeRepMap f -> Bool

After this ghc proposal is implemented, it should be possible to write such type signatures directly in code. The current workaround is to use this trick with KindOf type:

type KindOf (a :: k) = k

member :: forall a (f :: KindOf a -> Type) . Typeable a => TypeRepMap f -> Bool

New TypeRep performance🔗

During benchmarking the Map-based implementation of TypeRepMap, very perceptible performance degradation was noticed. Here is the comparison table with the results we have with our Map-based implementation.

ghc version containers performance
8.0.2 0.5.7.1 556.9 ns
8.2.2 0.5.10.2 2.076 μs
8.4.3 0.5.11.0 2.464 μs

We didn’t observe this performance degradation when we used Fingerprint as keys, so it’s probably an issue with the new TypeRep.

KnownNat and Typeable🔗

Initial version of buildBigMap function had this type signature:

buildBigMap :: forall (a :: Nat) . (Typeable a, KnownNat a) => ...

But, unfortunately, it became broken on GHC-8.4.3! Turned out that Typeable and KnownNat constraints don’t play well together. This observation resulted in the following ghc ticket with quite an interesting discussion: