Hi Morten,
That's a good idea.   In the meantime...I have checked and I think it's not the Rosetta version that gives the changed output you found, it is probably the compiler. (At least on my machine the current working Rosetta from svn and version 2.6 give the same output).  I put in an "ignore" on this change of output so starting today any phenix versions won't give an error message for that test if that output is changed.
All the best,
Tom T


From: [email protected] [[email protected]] on behalf of Morten Groftehauge [[email protected]]
Sent: Tuesday, January 08, 2013 4:04 AM
To: PHENIX user mailing list
Subject: Re: [phenixbb] dev-1261 wizard.test fail

Or you could have the test specify what version of Rosetta the results were compiled with. It's a bit opaque right now.

Cheers,
Morten


On 7 January 2013 17:43, Terwilliger, Thomas C <[email protected]> wrote:
Hi Morten,

Yes, you've guessed correctly that the regression tests do check the numbers in the output compared to a standard run. In some cases (as the two that you show) some minor change in the algorithms can give "failures" that are just due to small changes in algorithms.  I think you can probably ignore these two errors. I'll be checking these here and I'll just have it ignore these differences if they are uninteresting.

All the best,
Tom T

On Jan 7, 2013, at 6:35 AM, Morten Groftehauge wrote:

> Hi guys,
>
> I just updated Rosetta to 3.4 and it seems to have compiled correctly (at least no error interruptions). I've done this enough to not be satisfied by that and ran
> > phenix_regression.wizards.test_command_line_rosetta_quick
> which goes past the Rosetta tests correctly but fails at
> =================================================================
> Running test_prerefine
> ------------------------------------------------------------
> Numbers do not match: 0.47825 0.53601
> r {{ 0.87799   0.06946  -0.47362}
>    {-0.02049   0.99396   0.10779}
>    { 0.47825  -0.08494   0.87411}}
> ---
> r {{ 0.84375   0.06393  -0.53291}
>    {-0.02789   0.99676   0.07541}
>    { 0.53601  -0.04877   0.84280}}
>
> ------------------------------------------------------------
>
> FAILED
> See log files test_prerefine/test_prerefine.log  test_prerefine/test_prerefine_current.log
> =================================================================
> Running test_prerefine_double
> OK
> =================================================================
> Running test_prerefine_no_data
> ------------------------------------------------------------
> Numbers do not match: 0.47825 0.53601
> r {{ 0.87799   0.06946  -0.47362}
>    {-0.02049   0.99396   0.10779}
>    { 0.47825  -0.08494   0.87411}}
> ---
> r {{ 0.84375   0.06393  -0.53291}
>    {-0.02789   0.99676   0.07541}
>    { 0.53601  -0.04877   0.84280}}
>
> ------------------------------------------------------------
>
> FAILED
> See log files test_prerefine_no_data/test_prerefine_no_data.log  test_prerefine_no_data/test_prerefine_no_data_current.log
> =================================================================
> When I look in the log there are no error messages. I don't need to prerefine so I'm good but I still find it weird.
> I tried again with Rosetta 3.3 and got no error flags.
>
> Is it because the quick tests are checking the result against the results that you got earlier and the algorithm has been changed between 3.3 and 3.4?
>
> Cheers,
> Morten
>
> PS Really enjoyed the talks at the CCP4 study weekend.
>
> --
> Morten K Grøftehauge, PhD
> Pohl Group
> Durham University
> _______________________________________________
> phenixbb mailing list
> [email protected]
> http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________
phenixbb mailing list
[email protected]
http://phenix-online.org/mailman/listinfo/phenixbb



--
Morten K Grøftehauge, PhD 
Pohl Group
Durham University