Anyone running Julia on a Big-Endian machine?

[Kevin Squire]

Hi,

I'm wondering if anyone has attempted to run Julia on a Big-Endian machine?  In theory, there should be support, as long as LLVM supports the architecture.  But I haven't heard anything one way or another.

The main reason I'm asking is that I'm coding support for a file format (bam format, for encoding genetic sequencing data) which is encoded in little endian.  On a little endian architecture, I can just dump (most of) the data into the type and be done with it.  On a big-endian machine, I have to do extra work (e.g., with StrPack.jl).  I'm just wondering how worth my while it is to do so, especially since the files I'm using are large (20-100GB compressed), and working through them fast is useful, so I'd like to avoid as much overhead as possible.  (Of course, I don't know how much overhead StrPack really adds.)

Thanks,

   Kevin

[Stefan Karpinski]

Unless someone pipes up here and is willing to test, I would not bother trying to support big-endian systems. Without people actively using the code on big-endian machines, the chances that it would actually work are negligible.