Testing so far has proven this to be quite safe as texture memory read 
added a 2-5ms load to the current cache.

This also does some fixes on compressed textures reinterpret and on the
Fermi2D engine in general.

Also corrects some asserts.

This fixes clouds in SMO Cap Kingdom and lens on Cloud Kingdom.
Also moved accurate_gpu setting check to Pick Strategy

Fixed ASTC mipmaps loading
Fixed alignment on openGL upload/download
Fixed Block Height Calculation
Removed unalign_height

Instead of using Common::AlignUp, use Common::AlignBits to align the
texture compression factor.

Instead of storing all block width, height and depths in their shifted
form:

block_width = 1U << block_shift;

Store them like they are provided by the emulated hardware (their
block_shift form). This way we can avoid doing the costly
Common::AlignUp operation to align texture sizes and drop CPU integer
divisions with bitwise logic (defined in Common::AlignBits).

Due to our current infrastructure, it is possible for a mipmap to be set 
on as a render target before a texception of that mipmap's superset be 
set afterwards. This is problematic as we rely on texture views to set 
up texceptions and protecting render targets targets for 3D texture 
rendering.

One simple solution is to configure framebuffers after texture setup but 
this brings other problems. This solution, forces a reconfiguration of 
the framebuffers after such event happens.