[Wien] About the machines file for mpi

[liyh]

Dear Blaha,

       I am still not very clear about your suggest.
For example, for a large system, we use 2 k-points, and 4 nodes. then the .machines file should like this?
granularity:1
1:comp1 comp2
1:comp3 comp4
residue:comp1
lapw0:comp1 comp2 comp3 comp4

Thank you!

Yonghua

>Yes, your .machines file means you are running lapw0 on those two machines 
>in paralllel and then distribute the k-points into two sets, and each 
>k-point is calculated on two machnies in mpi-parallel mode.
>Of course your .machines file does NOT make sense.
>You should first try to distribute the k-points (because this is very 
>efficient) and you should not use the same machines for k-point AND 
>mpi-parallel mode (your example runs 2 lapw1_mpi jobs on each of those 
>machines).
>
>PS: I think the description of the .machines file is pretty exhausting in 
>the UG and shows almost all possible combinations of cluster and shared 
>memory machines.
>
>>        Do you means if I have two k-points the machines file should like this?
>> #For lapw0_mpi
>> lapw0: comp124 comp123
>> #For lapw1_mpi/lapw2_mpi
>> 1:comp124 comp123
>> 2:comp124 comp123
>> Does it correct?
>> I have looked the UG, it is very simple and only give the example for k-points parrel, but not for mpi.
>> 
>> Yonghua
>> 
>> >Please have a look at page 61 in the UG. 
>> >
>> >The lapw0 line must be a single line like
>> >
>> >lapw0:comp124 comp123 comp80 ...
>> >
>> >For lapw1 AND lapw2 again for each k-point you must use a single line only 
>> >like:
>> >
>> >1:comp124 comp123 comp80 ...
>> >
>> >
>> >
>> >>       I want to run wien2k in parallel using mpi method. But I dont know how to creat the .machines file. This is part of our script file for pbs submitting.
>> >> ...........
>> >> this is the created .machines file by this script.
>> >> 
>> >> lapw0:comp124
>> >> comp123
>> >> comp80
>> >> comp120
>> >> comp69
>> >> comp127
>> >> lapw1:comp124
>> >> comp123
>> >> comp80
>> >> comp120
>> >> comp69
>> >> comp127
>> >> lapw2:comp124
>> >> comp123
>> >> comp80
>> >> comp120
>> >> comp69
>> >> comp127
>> >> 
>> >> Does it correct? Could you give me some sugest?
>> >> 
>> >> Best Wish to You!
>> >> 
>> >> Yonghua
>> >> 
>> >> _______________________________________________
>> >> Wien mailing list
>> >> Wien at zeus.theochem.tuwien.ac.at
>> >> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>> >> 
>> >
>> >
>> >                                      P.Blaha
>> >--------------------------------------------------------------------------
>> >Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
>> >Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
>> >Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
>> >--------------------------------------------------------------------------
>> >
>> >_______________________________________________
>> >Wien mailing list
>> >Wien at zeus.theochem.tuwien.ac.at
>> >http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>> 
>> = = = = = = = = = = = = = = = = = = = =
>> 			
>> 
>>         ?
>> ?!
>>  
>> 				 
>>         liyh
>>         lyhua at fudan.edu.cn
>>           2005-01-22
>> 
>> 
>> _______________________________________________
>> Wien mailing list
>> Wien at zeus.theochem.tuwien.ac.at
>> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>> 
>
>
>                                      P.Blaha
>--------------------------------------------------------------------------
>Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
>Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
>Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
>--------------------------------------------------------------------------
>
>_______________________________________________
>Wien mailing list
>Wien at zeus.theochem.tuwien.ac.at
>http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

= = = = = = = = = = = = = = = = = = = =
			

　　　　　　　　致
礼！
 
				 
　　　　　　　　liyh
　　　　　　　　lyhua at fudan.edu.cn
　　　　　　　　　　2005-01-22