[Bioclusters] blat and hmmer on some architectures?

[Tim Cutts]

On 25 Oct 2005, at 1:05 pm, Jason Stajich wrote:

> re: blat
>
> Now that is not really fair about changing all the makefiles, you  
> only need to edit inc/common.mk to set the CFLAGS/COPT and CC  
> options.  I built blast with icc on i686 and gcc on x86_64 with no  
> problems.
>
> I don't understand why you think i686 is only supported.  just set  
> the $MACHTYPE variable.
>
> echo "MACHTYPE is: '$$MACHTYPE'" ; \
>                 echo "it should be something simple: i386 i686  
> sparc alpha x86_64 ppc, etc..." ;
>
> So just edit the inc/common.mk to specify your compiler flags and  
> set the MACHTYPE environment variable to uname -m

Hmm - I'm sure I had to edit a lot more than that; I had to remove  
all of the -lm's for one thing.  I don't think you have to do that  
any more, but for older versions of icc you did, if you wanted it to  
use Intel's math routines rather than the GLIBC ones.

Didn't mean to be unduly critical.  At least BLAT *has* a set of  
makefiles.  There are plenty of bits of bioinf software around that  
don't, and building some of them on new architectures can be a real  
chore.

Tim

-- 
Dr Tim Cutts
Informatics Systems Group, Wellcome Trust Sanger Institute
GPG: 1024D/E3134233 FE3D 6C73 BBD6 726A A3F5  860B 3CDD 3F56 E313 4233