Resolves #49

Either vector_base/matrix_base always create shallow copies in their copy-CTOR,
or they create deep copies in their copy-CTOR. The side-effects of shallow
copies can be corrupted data, leading to wrong results. On the contrary,
deep copies are likely to have only poor performance as a side-effect,
which is not as bad. The only use for shallow copies is within
proxy objects, which now have suitable overloads for their copy-CTOR.

Resolves #60

Now avoiding unnecessary temporary buffers.

FFT: Rewrote to fix VS2013 compilation errors

Assuming the complex numbers are in Cartesian form this should be identical in function to the prior function.

Use of 'uint' is rejected by some compilers, e.g. Visual Studio.

Never include this without a separate switch for Apple systems.

Resolves #68. Resolves #76.

v1 -= pow(v1, v2); might be poorly conditioned if v1 is close to 1.0.
This fix lifts the values of v1 to be at least 1.1, hopefully fixing the
repeated issues seen in the nightly tests in the past.

Execution times on CPU are otherwise excessive.
Provide runtime flag for switching sizes later.

About a factor of 20 faster than previous implementation.
I estimate that more microtuning can get another factor of 2.
Higher performance gains will most likely require intrinsics.

Apparently incomplete refactoring from output of tuner.

Older devices such as a GeForce 8600 GT have been mapped incorrectly.
The new detection fixes this for the devices currently available.
Needs to be revisited once the post-Maxwell generation is out.

Some macros have not been properly prefixed. Now every proprocessor macro carries
a prefix 'VIENNACL_'.

Issue introduced with a89695e5

The default local workgroup size is now 128, providing better compatibility with
weaker and older hardware. Also, preprocessor defines are properly guarded with a
VIENNACL-prefix.

The following is now legitimate:
  std::vector<T> x(N);
  viennacl::vector<T> vcl_x;
  viennacl::copy(x, vcl_x);
and resizes the ViennaCL vector accordingly (in the default context).
A similar support for the iterator interface is not possible without substantial refactoring.

Supplements 80697cf1 and addresses #71.

Resulted in a conversion warning and in a nightly test failure.

Resolves #71.

This way a user can conveniently free memory. Resolves #98.

If a user has provided her own memory buffers, the use of copy() to update
the matrix should reside in the same user-provided buffer and not create a new one.
Also, memory_write() is usually faster than memory_create().
Fixes #77.

Requires Doxygen 1.8.x or above.

Could cause some linkage issue when matrix_proxy.hpp was included but not matrix.hpp