summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorAndy Polyakov <appro@openssl.org>2009-03-17 20:04:11 +0000
committerAndy Polyakov <appro@openssl.org>2009-03-17 20:04:11 +0000
commit4e52b9845e0b0aeff2c93958fde59a0cccd70ab2 (patch)
tree42c07c3f4d7a4f2516e7b2fe38cd2f9c6314ed7f
parente22b864846d0ba7be13bf791843001f53ee54f70 (diff)
aes-390x.pl: commentary update.
-rw-r--r--crypto/aes/asm/aes-s390x.pl10
1 files changed, 5 insertions, 5 deletions
diff --git a/crypto/aes/asm/aes-s390x.pl b/crypto/aes/asm/aes-s390x.pl
index 6ee2097a15..4b27afd92f 100644
--- a/crypto/aes/asm/aes-s390x.pl
+++ b/crypto/aes/asm/aes-s390x.pl
@@ -35,11 +35,11 @@
#
# Add support for hardware AES192/256 and reschedule instructions to
# minimize/avoid Address Generation Interlock hazard and to favour
-# dual-issue z10 pipeline. This gave ~25% improvement on z10. The gain
-# should be larger on earlier CPUs, because being dual-issue z10 makes
-# it improssible to eliminate the interlock condition, critial path is
-# not long enough. Yet z10 spends ~24 cycles per byte processed with
-# 128-bit key.
+# dual-issue z10 pipeline. This gave ~25% improvement on z10 and
+# almost 50% on z9. The gain is smaller on z10, because being dual-
+# issue z10 makes it improssible to eliminate the interlock condition:
+# critial path is not long enough. Yet it spends ~24 cycles per byte
+# processed with 128-bit key.
#
# Unlike previous version hardware support detection takes place only
# at the moment of key schedule setup, which is denoted in key->rounds.