sha1-alpha.pl: commentary update.

crypto/sha/asm/sha1-alpha.pl

@@ -10,8 +10,9 @@
 # SHA1 block procedure for Alpha.
 
 # On 21264 performance is 33% better than code generated by vendor
-# compiler, and 75% better than GCC [3.4]. Implementation features
-# vectorized byte swap, but not Xupdate.
+# compiler, and 75% better than GCC [3.4], and in absolute terms is
+# 8.7 cycles per processed byte. Implementation features vectorized
+# byte swap, but not Xupdate.
 
 @X=(	"\$0",	"\$1",	"\$2",	"\$3",	"\$4",	"\$5",	"\$6",	"\$7",
 	"\$8",	"\$9",	"\$10",	"\$11",	"\$12",	"\$13",	"\$14",	"\$15");