Path: utzoo!utgpu!watmath!att!tut.cis.ohio-state.edu!brutus.cs.uiuc.edu!wuarchive!swbatl!texbell!nuchat!moray!urchin!f506.n106.z1.fidonet.org!Scott.Ladd From: Scott.Ladd@f506.n106.z1.fidonet.org (Scott Ladd) Newsgroups: comp.lang.c++ Subject: C++ Translators/Compilers Message-ID: <18728.24DED452@urchin.fidonet.org> Date: 7 Aug 89 13:06:54 GMT Sender: ufgate@urchin.fidonet.org (newsout1.26) Organization: FidoNet node 1:106/506 - Fulcrum's Edge, Spring TX Lines: 20 Just for general educational purposes, I thought I'd pass on this quick comparison of various C++ translators. I use the Guidelines and Intek translators in addition to the Zortech compiler. I keep the translators just so I can be sure that the programs I write are compilable with translators close to the original CFRONT developed at AT&T. Compiling a test program revealed the following statistics: Intek Guidelines Zortech ----- ---------- ------- Size (bytes) 22064 21840 10166 Time (secs) 55.40 54.12 29.46 Full optimization was used for the compiles. The translators used Microsoft C 5.1 as their C compiler (it was not possible to use Zortech C). The two translators produced DOZENS of warnings, usually involving the truncation of long identifiers. Even with its slow global optimization pass, Zortech came out so far ahead as to make the other products laughable. An interesting point: When using the translators, do not invoke Microsoft C with the /Oi or /Ox (which includes /Oi) switches on; this generates "duplicate definition" errors for all intrinsic functions such as memset() and strcpy(). Yuck. Verdict: I'm sticking with Zortech! ------------------------------------------------------------------------------ The above is a cross posting from the FidoNet C_Echo. Scott Ladd is a writer thus tries to be a neutral party in the "compiler wars". Roy Browning