On Mon, Aug 06, 2018, Thomas Pochapsky wrote:
> Just a general shout-out to the community...
>
> We have used residual dipolar coupling analysis (RDCs) as implemented in
> sander for ~8 years, and this has proven to be a really nice way to look
> at conformational changes that occur in cytochrome P450 enzymes upon
> substrate binding. Is there any plan to implement RDC restraints in
> pmemd? As most (all?) of the recent improvements in AMBER have been
> made there, I would volunteer to do the implementation myself, but I am
> but a poor organic chemist, and would get quickly lost in the maze of
> programming. OTOH, I would be happy to act as a beta tester if there
> are any volunteers.
Hi Tom: your request is noted, but I wonder if you could say a bit more:
are you eager to use pmemd.cuda for RDC calculations? For NMR-size
proteins, and parallel CPU calculations are not that much slower on
sander than on pmemd. For example, I just ran the benchmark at
$AMBERHOME/benchmarks/dhfr on 4 processors for 1000 steps:
sander time: 41 sec.; pmemd time: 22 sec.
That's for explicit water; if I run myoglobin with gb:
sander time: 34 sec.; pmemd time: 32 sec.
A factor of 2 (for explicit water) is certainly not negligible, but
it's not a deal-breaker either. The real "win" would be go get a GPU
implementation of RDC's. That's not impossible, but I wouldn't hold my
breath, either.
....regards...dave
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Aug 14 2018 - 08:30:03 PDT
