[compat] on ImageCore, Primes

[timholy]

Closes #12

@ChantalJuntao, I can't run the tests, can you do so and report whether this works?

[ChantalJuntao]

I'm not sure why I'm getting these warnings

Precompiling project...
  56 dependencies successfully precompiled in 171 seconds. 35 already precompiled.        
  2 dependencies had warnings during precompilation:
┌ BFloat16s [ab4f0b2a-ad5b-11e8-123f-65d77653426b]
│  WARNING: could not import Printf.ini_hex into BFloat16s
│  WARNING: could not import Printf.ini_HEX into BFloat16s
└
┌ CUDA [052768ef-5323-5732-b1bb-66c8b64840ba]
│  WARNING: Wrapping `Vararg` directly in UnionAll is deprecated (wrap the tuple instead).
│  WARNING: Wrapping `Vararg` directly in UnionAll is deprecated (wrap the tuple instead).
└
     Testing Running tests...
Test Summary:          | Pass  Total   Time
kernel_conv_components |    6      6  23.8s
Test Summary:       | Pass  Total  Time
kernel_calcNumDenom |    4      4  2.8s
Test Summary:    | Pass  Total   Time
Mismatch 2D test |    3      3  17.7s
Test Summary:    | Pass  Total  Time
Mismatch 3D test |    2      2  9.5s
Test Summary:              | Pass  Total  Time
Mismatch apertures 2D test |   60     60  4.1s
Test Summary:              | Pass  Total  Time
Mismatch apertures 3D test |    2      2  1.5s
     Testing RegisterMismatchCuda tests passed

But the tests seem to pass fine.

[timholy]

They are essentially "code quality" warnings:

• the first says the package BFloat16.jl is trying to import some symbols from the Printf stdlib that are no longer in Printf. Fixed in Remove ini_hex and ini_HEX on 1.6+ JuliaMath/BFloat16s.jl#37 and there's a new (breaking) release. If you're curious, you can use pkg> status --outdated -m to see what's holding it back. But this is not anything to worry about.
• the second is saying that CUDA.jl (the version you're using) is doing something like this. We're waaay behind on CUDA versions: this package requires CUDA 3.x and the latest is CUDA 5.x. But I probably can't change this without sitting down at a machine that has an NVIDIA card.