| Title: | C++ |
| Notice: | Read 1.* and use keywords (e.g. SHOW KEY/FULL KIT_CXX_VAX_VMS) |
| Moderator: | DECCXX::AMARTIN |
| Created: | Fri Nov 06 1987 |
| Last Modified: | Thu Jun 05 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 3604 |
| Total number of notes: | 18242 |
Is there any comparative bench on C++ compiler which gives the gain of
performances:
- WITH optimized code either in volume or speed
- and WHITHOUT optimization.
The customer context is DIGITAL UNIX 3.2C and TeMIP application
Regards,
G�rard
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 3423.1 | Some figures for TeMIP | TAEC::FLAUW | Marc Flauw, TeMIP Technical Office, VBO | Mon Feb 03 1997 10:59 | 21 |
Gerard, I don't have any comparative benchmark, but I can give you some figures in the context of Visual TeMIP. In average, our libraries without debug are 3 times smaller than the libraries with debug. (5.17 Mb compared to 15.08 Mb). This is mainly due to the symbol table in the debug code. In average on our application, the optimised code (-O2) is between 30% to 50% faster than debug code. The impact of using -O4 versus -O2 seems small (< 5%) in our application (TeMIP). I am not qualified to say if it is specific to our application or not. Best regards, Marc. PS: These numbers are coming from informal measures and I don't have any written document I can provide you with those figures in it. | |||||
| 3423.2 | BEGIN::ROTITHOR | Mon Feb 03 1997 13:13 | 4 | ||
There is no standard C++ bench to intercompare performance. There are however, some public domain benches commonly used/cited for evaluating performance of C++ compiler. Some of these can be obtained from: http://www.kai.com/index.html | |||||