[CP2K-user] dbcsr_distribution_get() subroutine problem

linfeng gan ganli... at gmail.com
Fri May 17 10:35:27 UTC 2019


Thanks you Alfio. I will try DBCSR with version 1.1.0.

Linfeng

在 2019年5月17日星期五 UTC+8下午5:27:37,Alfio Lazzaro写道:
>
> Dear Linfeng,
> The current DBCSR is in pre-release for v2, please use the official tag 
> included in the CP2K (https://github.com/cp2k/dbcsr/releases).
>
> The DBCSR v2 will be released soon and it requires changes in CP2K (as you 
> noticed).
>
> Alfio
>
>
>
>
> Il giorno venerdì 17 maggio 2019 05:03:14 UTC+2, linfeng gan ha scritto:
>>
>> Dear CP2K developers,
>> When I compild the newest version CP2K(2019-05-16) with newest dbcsr 
>> 2019-04-16 version, I runned across following error,
>>
>> error #6627: This is an actual argument keyword name, and not a dummy 
>> argument name.   [ROW_CLUSTER]
>>             CALL dbcsr_distribution_get(dist_qs, row_dist=blk_distr, 
>> row_cluster=cluster_distr)
>>
>> It seems there are some arguments problem in the source code. And I found 
>> that the error local in the src/mscfg_types.F
>>
>>          IF (dimen == 1) THEN !rows
>>             CALL dbcsr_distribution_get(dist_qs, row_dist=blk_distr, 
>> row_cluster=cluster_distr)
>>             CALL dbcsr_get_info(submatrix_in, row_blk_size=blk_sizes)
>>          ELSE !columns
>>             CALL dbcsr_distribution_get(dist_qs, col_dist=blk_distr, 
>> col_cluster=cluster_distr)
>>             CALL dbcsr_get_info(submatrix_in, col_blk_size=blk_sizes)
>>          ENDIF
>>
>> and the dbcsr_distribution_get() subroutine is local in the 
>> exts/dbcsr/src/dbcsr_api.F, which is 
>>    SUBROUTINE dbcsr_distribution_get(dist, row_dist, col_dist, &
>>                                      nrows, ncols, has_threads, &
>>                                      group, mynode, numnodes, nprows, 
>> npcols, myprow, mypcol, pgrid, &
>>                                      subgroups_defined, prow_group, 
>> pcol_group)
>>
>> It seems there not row_cluster argument lists in the 
>> dbcsr_distribution_get() subroutine. Is this a bug or something else? 
>>
>> Yours,
>> Linfeng
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20190517/3e3d2fec/attachment.htm>


More information about the CP2K-user mailing list