Date: Wed, 27 May 1998 22:52:23 +0200 From: Eivind Eklund <eivind@yes.no> To: Amancio Hasty <hasty@rah.star-gate.com> Cc: "John S. Dyson" <toor@dyson.iquest.net>, current@FreeBSD.ORG Subject: Re: Fix for undefined "__error" and discussion of shared object versioning Message-ID: <19980527225223.43868@follo.net> In-Reply-To: <199805270007.RAA06110@rah.star-gate.com>; from Amancio Hasty on Tue, May 26, 1998 at 05:07:56PM -0700 References: <19980526225254.45152@follo.net> <199805270007.RAA06110@rah.star-gate.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, May 26, 1998 at 05:07:56PM -0700, Amancio Hasty wrote: > Since, most of the time over here all I do is compile this or compile that > makes me interested on the speed of the compiler . for instance, if it > can speed up make world that would be great and it the compiler generates > decent code all the much better. For selected applications I am very > interested on the code generation . I'm interested in code generation more than compile speed everywhere. Anyway; here are the results: gcc -O2 -pipe: 11.41s (average over 3 runs) tcc -Ysystem: 8.35s (average over 3 runs) tcc -Ysystem, -Wt,-X: 8.21s (average over 3 runs) gcc -pipe: 6.94s (average over 3 runs) tcc defaults to all optimizations on; I'd guess that's about equal to -O2. -Wt,-X disables "all" optimizations - this seems to only disable high-level optimizations, and not peepholes, which probably take the largest amount of time. Before each set of test runs I did a 'get-everything-into-the-cache' run with the same options, to avoid that penalty. The runs were done from a DTP RAID-1 array with a 16MB cache in front, so disk access times should not be a significant part. Eivind. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980527225223.43868>