[Git][ghc/ghc][wip/T13253] Revert the multi-case inline fix

Simon Peyton Jones gitlab at gitlab.haskell.org
Fri Jul 24 09:10:00 UTC 2020



Simon Peyton Jones pushed to branch wip/T13253 at Glasgow Haskell Compiler / GHC


Commits:
eaa84461 by Simon Peyton Jones at 2020-07-24T10:08:56+01:00
Revert the multi-case inline fix

It makes GHC run a bit slower in some cases.
Let's try again without.

- - - - -


1 changed file:

- compiler/GHC/Core/Opt/Simplify/Utils.hs


Changes:

=====================================
compiler/GHC/Core/Opt/Simplify/Utils.hs
=====================================
@@ -1350,13 +1350,14 @@ postInlineUnconditionally env top_lvl bndr occ_info rhs
         -- This is very important in practice; e.g. wheel-seive1 doubles
         -- in allocation if you miss this out
 
-      OneOcc { occ_in_lam = in_lam, occ_int_cxt = int_cxt, occ_n_br = n_br }
+      OneOcc { occ_in_lam = in_lam, occ_int_cxt = int_cxt }
+             -- occ_n_br = n_br
         ->  -- See Note [Suppress exponential blowup]
-            n_br < (case int_cxt of
+            {- n_br < (case int_cxt of
                         IsInteresting  -> 16
                         NotInteresting -> 4)
 
-           && smallEnoughToInline dflags unfolding     -- Small enough to dup
+           && -} smallEnoughToInline dflags unfolding     -- Small enough to dup
                         -- ToDo: consider discount on smallEnoughToInline if int_cxt is true
                         --
                         -- NB: Do NOT inline arbitrarily big things, even if occ_n_br=1



View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/eaa8446170ae6685978c21a44f0ceb2c275aea3c

-- 
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/eaa8446170ae6685978c21a44f0ceb2c275aea3c
You're receiving this email because of your account on gitlab.haskell.org.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-commits/attachments/20200724/6f9a0796/attachment.html>


More information about the ghc-commits mailing list