[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Build 2.1.64 on OS 10.3. error

From: Michael Martin
Subject: Re: Build 2.1.64 on OS 10.3. error
Date: Tue, 21 Dec 2004 12:53:26 -0600

On Dec 21, 2004, at 12:22 PM, Per Persson wrote:

Sorry if I'm sounding negative, but it is my experience that applying something as agressive[1] as the -fast option will not work well on something as complex as octave.

Caveat simulator!

I second Per Persson's warning. I have had the -fast optimization optimize some of my loops right out of existence and thereby produce incorrect results.

We have a number of spacecraft simulations, which have varying degrees of complexity. I ran an extensive set of test cases on our basic simulator, slowly increasing the optimization levels. The no-optimization compile was considered the baseline, and was shown to agree with the results from our solaris and sgi installations of the simulator. When I compiled on OSX, -O1 was fine, so was -O2, as was -O3, but -fast, while it was the fastest, it could not reproduce the numbers seen in the baseline. Since I would rather not have my spacecraft fall apart on me (bad for business, you know). Our make files are set to -O3 for OSX.

OSX is not unique in this by any means. Our simulators run on several Unix platforms and tolerate different degrees of optimization on each platform. We run test cases to find what does and does not work for each simulator/platform/OS/compiler combination. It is different for each.

Michael W. Martin                   Phone: (281) 333-2177
Draper Laboratory                   FAX:   (281) 333-5276
2200 Space Park Dr.                 EMail: address@hidden
Houston, TX 77058                   WWW:
USA                                 Mail Code: EG/Draper

Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:
How to fund new projects:
Subscription information:

reply via email to

[Prev in Thread] Current Thread [Next in Thread]