search

radical-cybertools / radical.repex.at

This is the github location for RepEx developed by the RADICAL team in conjunction with the York Lab.
Other
4 stars 3 forks source link

defining exchange groups in 2D #17

Closed haoyuanchen closed 9 years ago

haoyuanchen commented 9 years ago

Hi Antons,

As we've discussed this morning, if we're running a 2D replica exchange simulation, for each exchange stage it should only exchange within one dimension and all the exchange partners should have the same value of the non-exchange variable. For example, if we have 4 replicas:

0 (T1, S1)

1 (T1, S2)

2 (T2, S1)

3 (T2, S2)

in a temperature exchange stage it should try to exchange (#0 with #2), and (#1 with #3) (no inter-group exchanges!). In a salt concentration exchange stage, it should try to exchange (#0 with #1), and (#2 with #3) (no inter-group exchanges!). If two replicas are different in both T and S (say, #0 and #3), then we shouldn't try to exchange them at all.

Thanks! Haoyuan

antonst commented 9 years ago

Hi Haoyuan,

Will be implemented.

Thanks, Antons

On Wed, Dec 17, 2014 at 3:57 PM, haoyuanchen notifications@github.com wrote:

Hi Antons,

As we've discussed this morning, if we're running a 2D replica exchange simulation, for each exchange stage it should only exchange within one dimension and all the exchange partners should have the same value of the non-exchange variable. For example, if we have 4 replicas:

0 (T1, S1)

1 https://github.com/radical-cybertools/RepEx/issues/1 (T1, S2)

2 https://github.com/radical-cybertools/RepEx/issues/2 (T2, S1)

3 https://github.com/radical-cybertools/RepEx/issues/3 (T2, S2)

in a temperature exchange stage it should try to exchange (#0 with #2 https://github.com/radical-cybertools/RepEx/issues/2), and (#1 https://github.com/radical-cybertools/RepEx/issues/1 with #3 https://github.com/radical-cybertools/RepEx/issues/3) (no inter-group exchanges!). In a salt concentration exchange stage, it should try to exchange (#0 with #1 https://github.com/radical-cybertools/RepEx/issues/1), and (#2 https://github.com/radical-cybertools/RepEx/issues/2 with #3 https://github.com/radical-cybertools/RepEx/issues/3) (no inter-group exchanges!). If two replicas are different in both T and S (say, #0 and #3 https://github.com/radical-cybertools/RepEx/issues/3), then we shouldn't try to exchange them at all.

Thanks! Haoyuan

— Reply to this email directly or view it on GitHub https://github.com/radical-cybertools/RepEx/issues/17.