[Zlib-devel] [PATCH] Vectorized adler32 calculation

[Stefan Fuhrmann]

Hi there,

please find attached the finished patch for speeding up 
adler32() by using SSE. The patch has been created 
against the released 1.2.4 version.

Changes to the "patch-in-progress" code:

* uses intrinsics now
* uses SSE2 only (no MMX nor SSE3) but with higher
  throughput on that
* automatically disables the feature for older MS compilers
  as well as for GCC, if the target processor does not 
  support SSE2
* configure script now detects SSE2 & intrinsics support
* Win32 builds with MSC assume at least an i486 processor

I tested the code on Win32/x86 (VS2005) as well as
linux/x64 (GCC 4.3).

-- Stefan^2.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: zlib-1.2.4.adler32.patch
Type: text/x-patch
Size: 14300 bytes
Desc: not available
URL: <http://madler.net/pipermail/zlib-devel_madler.net/attachments/20100417/3cc72d8b/attachment.bin>