Add cross-architecture drcachesim trace analysis and postprocessing support

It is useful to be able to analyze and post-process drcachesim offline traces gathered on one architecture on a different platform: e.g., to trace an ARM application and pull the trace onto an x86 machine for analysis.

Most of the analysis code is already cross-platform, but the decoder for tools like opcode_mix is the key missing piece. We would need something like #1684 to continue using DR's decoder there, but we'd need it in full DR and not just drdecodelib (because raw2trace uses full DR and it is complex to downgrade it).