diff --git a/x86_64/sha3-permute.asm b/x86_64/sha3-permute.asm index f58d787c33bce318a5374616e662c94d67a189c6..64c4cfc4907a473d8598c557c8a193c109b9ea1b 100644 --- a/x86_64/sha3-permute.asm +++ b/x86_64/sha3-permute.asm @@ -17,12 +17,6 @@ C along with the nettle library; see the file COPYING.LIB. If not, write to C the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, C MA 02111-1301, USA. -C NOTE: This code is currently unused. It gives a modest speedup on -C Intel processors, running at roughly two instructions per cycle, and -C a large slowdown on AMD processors, running at only a single -C instruction per cycle. Maybe it can be rewritten to not move data -C back and forth between xmm registers and regular 64-bit registers. - define(<CTX>, <%rdi>) C 25 64-bit values, 200 bytes. define(<COUNT>, <%r8>) C Avoid clobbering %rsi, for W64.