[molpro-user] Query related to high precision vector results.

[Andy May]

Raghu,

I agree the error is slightly less for the energies, but the dipoles do
not appear acceptable, hence why the job failed.

Please could you try with the most recent version of Molpro?

Thanks,

Andy

On 13/09/10 11:21, Swamy, Raghavendra wrote:
> Hello Molpro developers,
> 
>  
> 
> I have a specific query,
> 
> 1. The following verification tests failures are reported for ‘make
> quickest’, This failures happens when the libacml_mv.so’s high precision
> vector math routines are linked to the malpro.
> 
> -          ar2_dk_dummy.test
> 
> -          bccd_save.test
> 
> -          c2h4_prop.test
> 
> -          cn_tranh.test
> 
> -          cn_uno.test
> 
> -          dftacg_h2o_vdz.test
> 
> -          form_optlmp2.test
> 
> -          h2o_cfit.test
> 
> -          h2o_cfit_opt.test
> 
> -          h2o_cosmo_opt.test
> 
> -          h2o_dft.test
> 
> -          h2o_direct.test
> 
> -          h2o_extrapolate.test
> 
> -          h2o_extrapol_proc.test
> 
> -          h2o_freq.test
> 
> -          h2o_local.test
> 
> -          h2o_numfreq.test
> 
> -          h2o_opt_dflmp2.test
> 
> -          h2o_opt.test
> 
> -          h2o_optrs2.test
> 
> -          h2o_select.test
> 
> -          h2o_userop.test
> 
> -          h2o_vdztest
> 
> -          h2o_vdz_qcisd.test
> 
> -          o2_rs2.test
> 
> -          o2_vdz.test
> 
> -          sf6_dk.test
> 
> -          So the query is  “The reference h2o_dft.out has the
> deviations in the precision values when compared to the vector routine’s
> h2o_dft.errout for the same test case h2o_dft.test. Is it ok to go ahead
> by ignoring these failures since they seems to be more precise ?”.
> Please also let us know the permissible precision levels.
> 
> -          The sample reference and vector routines outputs for the test
> case “h2o_dft.test” has been attached.
> 
>  
> 
>  
> 
> Other details,
> 
> -          Molpro version - molpro2008.2
> 
> -          ACML-libm version   -  2.1
> 
> -          SUSE Linux Enterprise Server 11 (x86_64)
> 
> -          Open64 compiler
> 
>  
> 
> Do let me know if any other details are required.
> 
>  
> 
> 
> Thanks and regards,
> 
> Raghu
> 
> 
> 
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user