[GHC] #15001: Add strict product field demands to lambda binders

4 Apr 2018

      #15001: Add strict product field demands to lambda binders
-------------------------------------+-------------------------------------
           Reporter:  sgraf          |             Owner:  (none)
               Type:  task           |            Status:  new
           Priority:  normal         |         Milestone:  ⊥
          Component:  Compiler       |           Version:  8.5
           Keywords:  DmdAnal        |  Operating System:  Unknown/Multiple
       Architecture:                 |   Type of failure:  None/Unknown
  Unknown/Multiple                   |
          Test Case:                 |        Blocked By:
           Blocking:                 |   Related Tickets:
Differential Rev(s):  D4244          |         Wiki Page:
-------------------------------------+-------------------------------------
 This is a follow-up on the first improvement in [Phab:D4244#126311].

 Consider the following two functions with identical semantics:

 {{{
 module Foo where

 {-# LANGUAGE BangPatterns #-}
 data Complex a = !a :+ !a

 foo :: Complex Double -> Int -> Complex Double
 foo !x 0 = x
 foo (a :+ b) _ = b :+ a

 bar :: Complex Double -> Int -> Complex Double
 bar x 0 = x
 bar (a :+ b) _ = b :+ a
 }}}

 In the corresponding simplified Core, `x` gets strictness `S(SS)` for
 `foo`, but only `S` for `bar`. We could do better by looking at `x`s type
 and see that its a product type with strict fields.

 This is because currently only case binders get
 [https://github.com/ghc/ghc/blob/72b5f649ede82ab3bb429aa72ee1c572f415b0eb/com...
 special treatment] for strict product fields through
 `addDataConStrictness`.

 This ticket tracks if the same is worthwhile for lambda binders, as in
 `bar` above. Note that intuitively, this shouldn't make any difference
 because when we eventually unleash an `S` demand on a `Complex`
 constructor, we add `seqDmd` on strict fields anyway, amounting to
 `S(SS)`. So this is only a matter of recording things early because some
 parts of the analysis code might not be smart enough to add strict field
 demands by themselves.

 Also note that this doesn't touch impliciations on usage analysis at all,
 where unleashing the same `seqDmd` twice might mean we accidentally make
 some product field lose its single-entrieness.

-- 
Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/15001
GHC http://www.haskell.org/ghc/
The Glasgow Haskell Compiler

GHC

GHC

GHC

GHC

GHC

GHC

tags

participants (1)