Pointer Tagging

  
This patch implements pointer tagging as per our ICFP'07 paper "Faster
laziness using dynamic pointer tagging".  It improves performance by
10-15% for most workloads, including GHC itself.

The original patches were by Alexey Rodriguez Yakushev
<mrchebas@gmail.com>, with additions and improvements by me.  I've
re-recorded the development as a single patch.

The basic idea is this: we use the low 2 bits of a pointer to a heap
object (3 bits on a 64-bit architecture) to encode some information
about the object pointed to.  For a constructor, we encode the "tag"
of the constructor (e.g. True vs. False), for a function closure its
arity.  This enables some decisions to be made without dereferencing
the pointer, which speeds up some common operations.  In particular it
enables us to avoid costly indirect jumps in many cases.

More information in the commentary:

http://hackage.haskell.org/trac/ghc/wiki/Commentary/Rts/HaskellExecution/PointerTagging

1 files changed, 12 insertions, 3 deletions

diff --git a/compiler/codeGen/CgProf.hs b/compiler/codeGen/CgProf.hs
index 27ee54c50d..651f0eaa82 100644
--- a/compiler/codeGen/CgProf.hs
+++ b/compiler/codeGen/CgProf.hs
@@ -20,7 +20,7 @@ module CgProf (
 	emitSetCCC, emitCCS,
 
 	-- Lag/drag/void stuff
-	ldvEnter, ldvRecordCreate
+	ldvEnter, ldvEnterClosure, ldvRecordCreate
   ) where
 
 #include "HsVersions.h"
@@ -242,9 +242,12 @@ enter_cost_centre closure_info ccs body
   where
     enc_ccs    = CmmLit (mkCCostCentreStack ccs)
     re_entrant = closureReEntrant closure_info
-    node_ccs   = costCentreFrom (CmmReg nodeReg)
+    node_ccs   = costCentreFrom (cmmOffsetB (CmmReg nodeReg) (-node_tag))
     is_box     = isBox body
 
+    -- if this is a function, then node will be tagged; we must subract the tag
+    node_tag = funTag closure_info
+
 -- set the current CCS when entering a PAP
 enterCostCentrePAP :: CmmExpr -> Code
 enterCostCentrePAP closure = 
@@ -448,9 +451,14 @@ ldvRecordCreate closure = stmtC $ CmmStore (ldvWord closure) dynLdvInit
 -- The closure is not IND or IND_OLDGEN because neither is considered for LDV
 -- profiling.
 --
+ldvEnterClosure :: ClosureInfo -> Code
+ldvEnterClosure closure_info = ldvEnter (cmmOffsetB (CmmReg nodeReg) (-tag))
+  where tag = funTag closure_info
+        -- don't forget to substract node's tag
+  
 ldvEnter :: CmmExpr -> Code
 -- Argument is a closure pointer
-ldvEnter cl_ptr 
+ldvEnter cl_ptr
   =  ifProfiling $
      -- if (era > 0) {
      --    LDVW((c)) = (LDVW((c)) & LDV_CREATE_MASK) |
@@ -458,6 +466,7 @@ ldvEnter cl_ptr
     emitIf (CmmMachOp mo_wordUGt [loadEra, CmmLit zeroCLit])
 	   (stmtC (CmmStore ldv_wd new_ldv_wd))
   where
+        -- don't forget to substract node's tag
     ldv_wd = ldvWord cl_ptr
     new_ldv_wd = cmmOrWord (cmmAndWord (CmmLoad ldv_wd wordRep)
 				       (CmmLit (mkWordCLit lDV_CREATE_MASK)))