[vsipl++] maxval_test

[Jules Bergmann]

Assem Salama wrote:
> Everyone,
>  This test tests the maxval operator. It creates a vector on a subset of 
> the processors to test the processor mapping.

Assem,


This looks good.  Please address the feedback below and then check it in.

				thanks,
				-- Jules

> ------------------------------------------------------------------------

Assem,

[1] This file should be called tests/maxval.cpp or tests/par_maxval.cpp. 
  The _test suffix isn't necessary for files in the tests directory.

[2] Please add a header.  All source files in the library need a header. 
  In the future, please do this before posting a patch.


> 
> #include <vsip/initfin.hpp>
> #include <vsip/support.hpp>
> #include <vsip/map.hpp>
> #include <vsip/vector.hpp>
> #include <vsip/selgen.hpp>
> #include <vsip_csl/output.hpp>

> #include <vsip/opt/general_dispatch.hpp>
> #include <vsip/core/reductions/reductions_idx.hpp>

[3] Why is it necessary to include these directly?



> 
>   if(max == float(size-1) && max_idx[0] == size-1)
>   {
> #if DEBUG == 1
>     std::cout << "Test passed\n";
> #endif
>     return 0;
>   } else
>   {
> #if DEBUG == 1
>     std::cout << "Test failed\n";
> #endif
>     return -1;
>   }

[4] To make this test easier to extend and easier to debug (if it 
fails), it would better to use test_assert than returning the error code 
from main, i.e. something like

	#if DEBUG == 1
	  if(max == float(size-1) && max_idx[0] == size-1)
	    std::cout << "Test passed\n";
	  else
	    std::cout << "Test passed\n";
	#endif
	test_assert(max == float(size-1) && max_idx[0] == size-1);

-- 
Jules Bergmann
CodeSourcery
jules at codesourcery.com
(650) 331-3385 x705