arm: Add optimized add_n and sub_n
authorRichard Henderson <rth@twiddle.net>
Wed, 27 Feb 2013 00:17:51 +0000 (16:17 -0800)
committerRichard Henderson <rth@twiddle.net>
Wed, 6 Mar 2013 15:47:26 +0000 (07:47 -0800)
commita6f2f4353886ba4868c3f41001d2eac2fcebf4b8
tree3bbc6e07f695ed7e5bfe3b6c81692dd65df87a6d
parentfe62ba37a30b83be7e2a63b667fd519ca7e1b28d
arm: Add optimized add_n and sub_n

Written from scratch rather than copied from GMP, due to LGPL 2.1 vs
GPL 3, but tested with the GMP testsuite.

This is 250% faster than the generic code as measured on Cortex-A15,
and the same speed as GMP on the same core, and probably everywhere.
ports/ChangeLog.arm
ports/sysdeps/arm/add_n.S [new file with mode: 0644]
ports/sysdeps/arm/sub_n.S [new file with mode: 0644]