On 3/5/2010, at 10:19 a.m., David Watson wrote:
> On May 3, 2010, at 8:53 AM, Ben Roberts wrote:
>> [snip]
>> Like Alan, I upgraded to 11.1.088, and I'm running OS X 10.6.3. I see the same bugs in AmberTools that he did.
>
> One thing to keep in mind is that you can "downgrade" by downloading an older compiler from your Intel Support account.
> If I remember correctly, when my account was current, I had a choice of downgrading to the 10.x compilers.
Which means I could almost certainly go all the way back to 11.1.067 if I wish (I kept my previous versions on the computer). But I wanted to find out the extent of the problem as it stands before resorting to drastic measures. :)
To put the cat among the pigeons some more, Amber serial is presenting me with problems too:
cd gact_ips && ./Run.ipsnve
forrtl: severe (174): SIGSEGV, segmentation fault occurred
cd gact_ips && ./Run.ips_dfft
forrtl: severe (174): SIGSEGV, segmentation fault occurred
cd gact_ips && ./Run.ips_vacuum
forrtl: severe (174): SIGSEGV, segmentation fault occurred
cd newton_raph && ./Run.newton_raph
diffing nmd_min.out.save with nmd_min.out
possible FAILURE: check nmd_min.out.dif
(on inspection, many numerical differences)
export TESTsander='../../exe/pmemd'; cd tip5p && ./Run.tip5p
diffing mdout.tip5p.save with mdout.tip5p
possible FAILURE: check mdout.tip5p.dif
(which we already knew about, so I'm merely mentioning it "for the record")
I've attached my diff file so people can see what the "failures" are. I'm also about to test the parallel bits and pieces, especially since I see Alan asked after pmemd.MPI. (I'm using OpenMPI 1.4.1.)
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon May 03 2010 - 08:00:03 PDT
