Hi,
Also, add a missing include of opensslconf.h so that we properly
get the OPENSSL_SYSNAME_ULTRASPARC define even in the 32-bit case.
These changes give a pretty reasonable speed boost.
On a SPARC T4-2, without these changes:
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
des cbc 39148.28k 40071.83k 40669.53k 40685.80k 40864.43k
des ede3 14358.54k 14528.02k 14611.88k 14632.96k 14641.83k
and with them:
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
des cbc 54334.03k 56280.92k 57101.01k 57549.14k 57619.80k
des ede3 20010.98k 20682.62k 20834.90k 20873.56k 20886.87k
I can't confirm your "before" numbers. I get
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
des cbc 51501.71k 53611.82k 54637.06k 54721.74k 54976.51k
des ede3 19572.77k 20307.14k 20472.83k 20513.45k 20526.42k
Which kind of makes sense, because the difference between "before" and
"after" is difference between 8 byte loads + 6 shifts-n-ors and pair of
little-endian 32-bit loads, and there is no way former would be that
much slower. Therefore I reckon that
http://git.openssl.org/gitweb/?p=openssl.git;a=commitdiff;h=d8f3ed23062539c4a6f7b4153633f921ce16a15f
is sufficient.
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List openssl-dev@openssl.org
Automated List Manager majord...@openssl.org
