# instructions and achieve "64-bit" performance even in 31-bit legacy
# application context. The feature is not specific to any particular
# processor, as long as it's "z-CPU". Latter implies that the code
-# remains z/Architecture specific. On z900 SHA256 was measured to
+# remains z/Architecture specific. On z990 SHA256 was measured to
# perform 2.4x and SHA512 - 13x better than code generated by gcc 4.3.
$flavour = shift;