discuss-gnuradio
[Top][All Lists]
Advanced

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

Re: [Discuss-gnuradio] accuracy problem?


From: David P. Reed
Subject: Re: [Discuss-gnuradio] accuracy problem?
Date: Thu, 02 Jun 2005 07:41:01 -0400
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.6) Gecko/20050317 Thunderbird/1.0.2 Mnenhy/0.6.0.104

Eric Blossom wrote:

On Wed, Jun 01, 2005 at 04:12:16PM -0400, David P. Reed wrote:
I'm trying to install the gnuradio software from CVS on Fedora Core 3 on an AMD64 (x86_64) laptop (and python 2.3). Aside from annoying pathname bugs and quirks with FFTW linking, which I will report separately, I get the following error when testing gnuradio-core, where a unittest fails. Anyone who has a clue why this imprecision might be happening? If I change this one test to check for 5 digits after the decimal point, it works...

Also, has anyone gotten gnuradio to run on x86_64?

Hi David,

Some folks have GNU Radio running on x86_64.  I just took delivery on
one the other day, but haven't had a chance to spend any time with it.

I think the difference you're seeing is that GCC for the x86_64 uses
the SSE scalar mode instructions at 32-bits instead of the x87
instructions with 80-bits of precision.

Eric


Yeah, but should the test process stop the build process, or should the check be for 5 digits rather than 6?

I'm still trying to figure out why I can't build the gsm codec, apparently because of linker errors with fftw (multiply defined symbols).




reply via email to

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