>
> While restructuring the tests I have also changed the tmod tool to become a
> general command line interface to lensfun. So that might already include
> everything you need.
Will it work natively with different pixel types?
If yes, then *i think*, that is pretty much all what i need.
At least a test where we compare the output of the SSE and non SSE
> functions and
> assert that they deliver equal results for some test points. But we can
> further
> look at this when the test framework is ready.
I checked, and *at this moment*, results from old normal code and my new
SSE code *has different MD5*.
So most likely the comparison will need to be a little bit more clever than
memcmp(3) <http://en.cppreference.com/w/cpp/string/byte/memcmp>
I will merge your code and my changes into master during the weekend so we
> can
> work on the same code base.
Sounds good! :)
On Fri, Oct 10, 2014 at 1:12 PM, Sebastian Kraft <[email protected]>
wrote:
> > Well, as far as i am aware, i will need to modify lensfun/tests
> ><https://github.com/LebedevRI/lensfun/blob/master/tests/tmod/tmod.cpp>
> >/tmod/tmod.cc
> ><https://github.com/LebedevRI/lensfun/blob/master/tests/tmod/tmod.cpp>
> anyway,
> > when i'm done with lf_f32 pixel type and ready to move on to others
> (to get
> >a tool that can be run from "console" to measure performance)
> >
>
> While restructuring the tests I have also changed the tmod tool to become a
> general command line interface to lensfun. So that might already include
> everything you need.
>
> > >And we should add some tests to verify that SSE code and normal
> > >code both give accurate results. Maybe you can help to write tests for
> > >the vignetting part?
> > Not sure i'm following, how should results be verified?
> > By per-pixel comparison of the same input image processed through
> "normal"
> >code and hand-optimized one?
> > Such kind of tests would be indeed a great tool to have, but i will
> need a
> >sample (e.g. such a test for TCA) to do something about it.
> >
>
> At least a test where we compare the output of the SSE and non SSE
> functions and
> assert that they deliver equal results for some test points. But we can
> further
> look at this when the test framework is ready.
>
> I will merge your code and my changes into master during the weekend so we
> can
> work on the same code base.
>
> Sebastian
>
------------------------------------------------------------------------------
Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer
Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports
Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper
Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer
http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk
_______________________________________________
Lensfun-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/lensfun-users
