# Atom 570 0.26
#
# The ratio values mean that 128-byte blocks will be processed
-# 21-27% slower, 256-byte blocks - 12-16%, 382-byte blocks - 8-11%,
+# 21-27% slower, 256-byte blocks - 12-16%, 384-byte blocks - 8-11%,
# etc. Then keep in mind that input sizes not divisible by 128 are
# *effectively* slower, especially shortest ones, e.g. consecutive
# 144-byte blocks are processed 44% slower than one would expect,
#
# Performance summary. aes-586.pl column lists large-block CBC
# encrypt/decrypt/with-hyper-threading-off(*) results in cycles per
-# byte processed with 128-bit key, and vpaes-x86.pl column -
-# encrypt/decrypt.
+# byte processed with 128-bit key, and vpaes-x86.pl column - [also
+# large-block CBC] encrypt/decrypt.
#
# aes-586.pl vpaes-x86.pl
#
$k_dsbo=0x2c0; # decryption sbox final output
&data_word(0x7EF94000,0x1387EA53,0xD4943E2D,0xC7AA6DB9);
&data_word(0x93441D00,0x12D7560F,0xD8C58E9C,0xCA4B8159);
-&asciz ("Vector Permutation AES for x86, Mike Hamburg (Stanford University)");
+&asciz ("Vector Permutation AES for x86/SSSE3, Mike Hamburg (Stanford University)");
&align (64);
&function_begin_B("_vpaes_preheat");
# Performance summary. aes-x86_64.pl column lists large-block CBC
# encrypt/decrypt/with-hyper-threading-off(*) results in cycles per
# byte processed with 128-bit key, and vpaes-x86_64.pl column -
-# encrypt/decrypt.
+# [also large-block CBC] encrypt/decrypt.
#
# aes-x86_64.pl vpaes-x86_64.pl
#
.Lk_dsbo: # decryption sbox final output
.quad 0x1387EA537EF94000, 0xC7AA6DB9D4943E2D
.quad 0x12D7560F93441D00, 0xCA4B8159D8C58E9C
-.asciz "Vector Permutaion AES for x86_64, Mike Hamburg (Stanford University)"
+.asciz "Vector Permutaion AES for x86_64/SSSE3, Mike Hamburg (Stanford University)"
.align 64
.size _vpaes_consts,.-_vpaes_consts
___