Eliminate platform specific isX86 check in codegen agnostic Recognized Call Transformer

[fjeremic]

As part of [1] there seems to be a x86 specific check introduced in a codegen agnostic transformation. The reason the check exists is because the Z codegen currently pattern matches the direct calls at direct call evaluator time and inlines these toUpper and toLower methods. We should switch to using the common Recognized Call Transformer infrastructure for Z and eliminate these platform agnostic checks.

There also seems to be another isX86() check in the same file for Math.min acceleration. We should try to get rid of this one as well. No reason why this shouldn't be done on all platforms.

[1] #1366 (review)

[fjeremic]

This was completed as part of #2320.