[PATCH] T3/T4 sparc shifts, plus more timings

David Miller davem at davemloft.net
Mon Apr 1 19:43:54 CEST 2013

From: Torbjorn Granlund <tg at gmplib.org>
Date: Mon, 01 Apr 2013 19:34:11 +0200

> Clearly, the new lshiftc code is not optimal for US1 through US4.  It
> runs 0.5 c/l slower on them all, compared to what one would hope for
> 2-way unrolled code.

I think with some tinkering one of us could get rid of that extra
cycle somehow, it's a matter of pipeline details rather than a
fundamental issue.

More information about the gmp-devel mailing list