Hi JC;

   To answer your question definitely, I wish that the Application Profiler feature of PB 2017 also worked for the 64bit Apps as it does for the 32bit Apps. Then we could have an absolute way of measuring the true run-time performance differences.

   I can say though through observation and testing of my own framework which is now 64bit compliant - that the applications I have developed & deployed from PB 2017 seem to run just as well from the 64bit EXE vs their 32bit EXE counter-part. My framework does measure the DBMS interaction times of each DW object and records these timings in its framework log. From what I have seen comparing the DB part of the logs - the DBMS interaction time seems comparable in the 32bit vs 64bit App comparison. Of course, DBMS response times might be better for you depending on how well your particular DBMS processes 64bit client DML requests vs 32bit ones. My DBMS tests thus far have focused mainly on SQL Anywhere v17 vs ASE, SS or Oracle.

  If you look at the 64bit world in general, 64bit Apps tend to load a little slower but once instantiated process faster. While 32bit Apps load faster, they need to run in the emulation mode known as Windows On Windows or WOW for short. WOW does require a little extra over-head to use at run time - but, that is marginal. So I would think that the trade-off of WOW vs native 64bit operations should place the 32bit vs 64bit Apps on a somewhat even playing field performance wise overall.

  I am just working with a PB App this week that I wrote that walks through MS-Windows control blocks. So its very memory & CPU intensive. The 64bit version seems   faster than the 32bit P-Code version for sure and "on par" with the 32bit M-Code compiled version - speed wise. Just an observation I had noticed yesterday. However, this is not your typical PB business App processing profile.

HTH

Regards ... Chris