[Sbcl-commits] CVS: sbcl NEWS,1.939,1.940 version.lisp-expr,1.2629,1.2630

[Juho S. <js...@us...>]

Update of /cvsroot/sbcl/sbcl
In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv26832

Modified Files:
	NEWS version.lisp-expr 
Log Message:
0.9.8.19:
        Changes to GENCGC memory zeroing behaviour that give a big 
        performance boost, especially on modern processors.
        
        * Instead of zeroing memory by remapping memory with
          munmap/mmap at GC time, pages are just marked as needing
          zeroing and zeroed with memset when they're added to a new
          allocation region. This reduces GC latency both for the
          common and worst cases.

        * To keep the memory footprint down, clear the pages by
          remapping after major GCs (arbitrarily defined as a
          collection of generation 2 or older).  The memory freed from
          a minor GC is just going to get used again immediately , so
          releasing them back to the OS would make little sense.

        * Add a GENCGC mode (#define READ_PROTECT_FREE_PAGES) for
          catching attempts to read unallocated pages

        * See sbcl-devel "Changes to GENCGC memory zeroing" in 2005-12
          for more details and performance measurements. (Note
          that many parts of this patch have already been committed
          piecemeal over the last month, this is just the most
          significant chunk).

        * Performance effect on BSDs (which used a different zeroing
          strategy than Linux before this) is unknown.


Index: NEWS
===================================================================
RCS file: /cvsroot/sbcl/sbcl/NEWS,v
retrieving revision 1.939
retrieving revision 1.940
diff -u -d -r1.939 -r1.940
--- NEWS	6 Jan 2006 03:31:26 -0000	1.939
+++ NEWS	7 Jan 2006 18:53:39 -0000	1.940
@@ -9,6 +9,7 @@
   * bug fix: interrupt handling on NetBSD (thanks to Richard M
     Kreuter)
   * bug fix: saving a core corrupted callbacks on x86/x86-64
+  * optimization: major improvements to GC efficiency on GENCGC platforms
   * optimization: faster implementation of EQUAL
   * optimization: emit more efficient opcodes for some common 
     immediate->register MOV instructions on x86-64.  (thanks to Lutz Euler)

Index: version.lisp-expr
===================================================================
RCS file: /cvsroot/sbcl/sbcl/version.lisp-expr,v
retrieving revision 1.2629
retrieving revision 1.2630
diff -u -d -r1.2629 -r1.2630
--- version.lisp-expr	7 Jan 2006 18:22:58 -0000	1.2629
+++ version.lisp-expr	7 Jan 2006 18:53:39 -0000	1.2630
@@ -17,4 +17,4 @@
 ;;; checkins which aren't released. (And occasionally for internal
 ;;; versions, especially for internal versions off the main CVS
 ;;; branch, it gets hairier, e.g. "0.pre7.14.flaky4.13".)
-"0.9.8.18"
+"0.9.8.19"