The same curious can compare performances (expressed in second) between gcc and tcc (both 32/64 bits) on the same Windows 7 x64 machine. While LAP compiled code (interpreted by a virtual machine) is generally twice as fast as the interpreter, tcc is two times slower. GNU GCC 4.5 (Mgw) XP (Compile) | 1.047 | GNU GCC 4.5 (Mgw) XP (32 bits) | 2.237 | GNU GCC 4.5 (Mgw) XP64 (Compile) | 1.208 | GNU GCC 4.5 (Mgw) XP64 (64 bits) | 1.985 | Tiny C Compiler XP (32 bits) | 5.717 | Tiny C Compiler XP (Compile) | 10.104 | Tiny C Compiler XP64 (64 bits) | 5.642 | Tiny C Compiler XP64 (Compile) | 10.108 |
Here is the complet table on the same Windows machine with all compilers I use for my non regression tests: Name (mode) | 9.7.0 | ---------------------------------+-------+- GNU GCC 4.5 (Mgw) XP (Compile) | 1.047 | Microsoft C 16.0 XP (Compile) | 1.139 | GNU GCC 4.5 (Mgw) XP64 (Compile) | 1.208 | Watcom C 1.90 XP (Compile) | 1.263 | Microsoft C 16.0 XP64 (Compile) | 1.311 | Borland C 5.60 XP (Compile) | 1.341 | IBM Visual Age C XP (Compile) | 1.358 | Digital Mars 8.52 XP (Compile) | 1.364 | LCC C Compiler XP (Compile) | 1.373 | Microsoft C 16.0 XP (Unicode) | 1.871 | Microsoft C 16.0 XP (32 bits) | 1.902 | GNU GCC 4.5 (Mgw) XP64 (64 bits) | 1.985 | Microsoft C 16.0 XP64 (64 bits) | 1.985 | Microsoft C 16.0 XP64 (Unicode) | 1.996 | Microsoft C 16.0 XP (64float) | 1.997 | GNU GCC 4.5 (Mgw) XP (32 bits) | 2.237 | Borland C 5.60 XP (32 bits) | 2.608 | Watcom C 1.90 XP (32 bits) | 2.649 | Pelles C Compiler XP64 (64 bits) | 2.834 | Pelles C Compiler XP (32 bits) | 2.862 | IBM Visual Age C XP (32 bits) | 3.027 | Digital Mars 8.52 XP (32 bits) | 3.313 | LCC C Compiler XP (32 bits) | 3.624 | Tiny C Compiler XP64 (64 bits) | 5.642 | Tiny C Compiler XP (32 bits) | 5.717 | Tiny C Compiler XP (Compile) | 10.104 | Tiny C Compiler XP64 (Compile) | 10.108 | ________________________________ From: tinycc-devel-bounces+eligis=orange...@nongnu.org [mailto:tinycc-devel-bounces+eligis=orange...@nongnu.org] On Behalf Of Stanley Steel Sent: mercredi 2 février 2011 16:59 To: tinycc-devel@nongnu.org Subject: Re: [Tinycc-devel] i just can't get over how farging fast tcc is... For the curious: I have been compiling several of my projects with both tcc and gcc for awhile now. Specifically, I've been compiling an ANTLR C project that consist of about 60,000 lines of ANLTR generated code that translates another language to C. As you would expect, the compilation with tcc is noticeably faster. The runtime performance of scanning, parsing, and translating a ~100 line file is: gcc= ~.000850 sec tcc= ~.001050 sec I've done other test as well (microbenchmarks), but in those test there was no noticeable difference. The state of my machine seemed to be the biggest variable even for long running loops. On Wed, Feb 2, 2011 at 3:53 AM, Gary Birkett <liq...@gmail.com> wrote: On Wed, Feb 2, 2011 at 3:20 AM, Stephan Beal <sgb...@googlemail.com> wrote: On Wed, Feb 2, 2011 at 4:09 AM, Gary Birkett <liq...@gmail.com> wrote: :) Why not test it? :) The milliseconds won't matter for what i'm doing - the database access times swamp any overhead introduced by a non-optimizing compiler. :) (You're not going to convince me to not be impressed with tcc's speed ;) I am already impressed with its reported build speed, though I am curious at just how much of a difference the optimisation steps in gcc actually make. I would test myself, but most of the things I build are for ARM, hence my asking you to try some tests since you are using it :) Regards, Gary Birkett -- ----- stephan beal http://wanderinghorse.net/home/stephan/ _______________________________________________ Tinycc-devel mailing list Tinycc-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/tinycc-devel _______________________________________________ Tinycc-devel mailing list Tinycc-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/tinycc-devel _______________________________________________ Tinycc-devel mailing list Tinycc-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/tinycc-devel
