[Biodevelopers] Re: [Bioclusters] working on NCBI "port" to Opteron

[Chris Dagdigian]

I'll likely have the chance to get long term access to Opteron systems 
for bioclustering stuff but this will absolutly be a few weeks away as 
I've got to head to Brisbane for the ISMB/BOSC meetings.

Anyone else have them yet? Anyone have them who is not under NDA? heh.

Joe- I'm particularly interested how different compilers do when 
building for Opteron. Any feedback on that would be appreciated.

-Chris



Joe Landman wrote:
> Hi folks:
> 
>   I have had the joy of playing with an Opteron for a few days, and have
> done some preliminary testing on it using some BLAST runs.  Hopefully I
> will be able to summarize in a few days.
> 
>   One of the interesting side effects of working on this platform is the
> need to update one of the NCBI header files.  Here is my quick diff:
> 
> diff ./ncbi/include/ncbilcl.h.original ./ncbi/include/ncbilcl.h
> 90a91,92
> 
>>#elif defined(__x86_64__) || #cpu(x86_64) || #machine(x86_64)
>>#define PROC_X86_64
> 
> 
>   Without this, it does compile, but it complains about not knowing
> which processor it was being compiled for.  Hopefully I can get longer
> term access to an Opteron along with better compilers, so I can complete
> this work.
>  
>   Just an FYI.
> 
> Joe
>