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.