Better occurrence analysis with castswip/T20988

This patch addresses #20988 by refactoring the way the
occurrence analyser deals with lambdas.

Previously it used collectBinders to split off a group of binders,
and deal with them together.  Now I deal with them one at a time
in occAnalLam, which allows me to skip casts easily.  See
    Note [Occurrence analysis for lambda binders]
about "lambda-groups"

This avoidance of splitting out a list of binders has some good
consequences.  Less code, more efficient, and I think, more clear.

The Simplifier needed a similar change, now that lambda-groups
can inlude casts.  It turned out that I could simplify the code
here too, in particular elminating the sm_bndrs field of StrictBind.
Simpler, more efficient.

Compile-time metrics improve slightly; here are the ones that are
+/- 0.5% or greater:

                                  Baseline
              Test    Metric         value     New value Change
--------------------------------------------------------------------
   T11303b(normal) ghc/alloc    40,736,702    40,543,992  -0.5%
    T12425(optasm) ghc/alloc    90,443,459    90,034,104  -0.5%
    T14683(normal) ghc/alloc 2,991,496,696 2,956,277,288  -1.2%
    T16875(normal) ghc/alloc    34,937,866    34,739,328  -0.6%
   T17977b(normal) ghc/alloc    37,908,550    37,709,096  -0.5%
    T20261(normal) ghc/alloc   621,154,237   618,312,480  -0.5%
     T3064(normal) ghc/alloc   190,832,320   189,952,312  -0.5%
     T3294(normal) ghc/alloc 1,604,674,178 1,604,608,264  -0.0%
   T5321FD(normal) ghc/alloc   270,540,489   251,888,480  -6.9% GOOD
  T5321Fun(normal) ghc/alloc   300,707,814   281,856,200  -6.3% GOOD
     WWRec(normal) ghc/alloc   588,460,916   585,536,400  -0.5%

         geo. mean                                        -0.3%

Metric Decrease:
    T5321FD
    T5321Fun

1 files changed, 8 insertions, 7 deletions

diff --git a/compiler/GHC/Core/SimpleOpt.hs b/compiler/GHC/Core/SimpleOpt.hs
index 720bc895c8..92f1c7987e 100644
--- a/compiler/GHC/Core/SimpleOpt.hs
+++ b/compiler/GHC/Core/SimpleOpt.hs
@@ -28,7 +28,7 @@ import GHC.Core.FVs
 import GHC.Core.Unfold
 import GHC.Core.Unfold.Make
 import GHC.Core.Make ( FloatBind(..) )
-import GHC.Core.Opt.OccurAnal( occurAnalyseExpr, occurAnalysePgm )
+import GHC.Core.Opt.OccurAnal( occurAnalyseExpr, occurAnalysePgm, zapLambdaBndrs )
 import GHC.Types.Literal
 import GHC.Types.Id
 import GHC.Types.Id.Info  ( realUnfoldingInfo, setUnfoldingInfo, setRuleInfo, IdInfo (..) )
@@ -332,20 +332,21 @@ simple_app env (App e1 e2) as
   = simple_app env e1 ((env, e2) : as)
 
 simple_app env e@(Lam {}) as@(_:_)
-  | (bndrs, body) <- collectBinders e
-  , let zapped_bndrs = zapLamBndrs (length as) bndrs
+  = do_beta env (zapLambdaBndrs e n_args) as
     -- Be careful to zap the lambda binders if necessary
     -- c.f. the Lam case of simplExprF1 in GHC.Core.Opt.Simplify
     -- Lacking this zap caused #19347, when we had a redex
     --   (\ a b. K a b) e1 e2
     -- where (as it happens) the eta-expanded K is produced by
     -- Note [Typechecking data constructors] in GHC.Tc.Gen.Head
-  = do_beta env zapped_bndrs body as
   where
-    do_beta env (b:bs) body (a:as)
+    n_args = length as
+
+    do_beta env (Lam b body) (a:as)
       | (env', mb_pr) <- simple_bind_pair env b Nothing a NotTopLevel
-      = wrapLet mb_pr $ do_beta env' bs body as
-    do_beta env bs body as = simple_app env (mkLams bs body) as
+      = wrapLet mb_pr $ do_beta env' body as
+    do_beta env body as
+      = simple_app env body as
 
 simple_app env (Tick t e) as
   -- Okay to do "(Tick t e) x ==> Tick t (e x)"?