[PATCH 2/2] Optimize 64-bit mpn_add_N and mpn_sub_N for sparc T3 and later.

Torbjorn Granlund tg at gmplib.org
Thu Mar 7 07:00:00 CET 2013

David Miller <davem at davemloft.net> writes:

  > I optimised submul_1.asm, and then edited both addmul_1 and submul_1 to
  > use as similar operand order as possible.  Please test these using
  > tests/devel/try, and please time this new submul_1.
  The testsuite starts failing very early with these changes.
Sorry about that.

As always with assembly code, tests/devel/try is our friend.
It points out the error, and tests thoroughly.

I believe it would have complained about the return value and nothing
else.  There was a typo where I use %o0 instead of %i0 for the
destination register in the last arithmetic instruction of both

Why doesn't your functions use the 'return' insns, btw?

PS. I have created a poor man's T3 with some m4 macros.  This will allow
me to test T3 code on an old v9 machine.


More information about the gmp-devel mailing list