Merge pull request #14 from ssfrr/patch-1

updates README with `init` clarification and other use-facing functions

Author: SimonDanisch

README.md

@@ -7,12 +7,20 @@ Implementation of the [abstract GPU Array Interface](https://github.com/JuliaGPU
 
 CLArray uses [Transpiler.jl](https://github.com/SimonDanisch/Transpiler.jl) to compile Julia functions for the GPU using OpenCL.
 
-It implements the full abstract gpu interface from GPUArrays.
-To learn how to use it, please refer to the GPUArray documentations:
+It implements the full abstract gpu interface from GPUArrays, and most interactions will be through those functions.
+To learn how to use it, please refer to the GPUArray documentation:
 
 [![](https://img.shields.io/badge/docs-latest-blue.svg)](https://JuliaGPU.github.io/GPUArrays.jl/latest)
 
-The only noteworthy functionality that isn't part of the GPUArray interface is how to select and initialize devices:
+CLArrays includes several other OpenCL-specific functions:
+
+* `CLArrays.devices()` returns a list of the OpenCL compute devices (CPU and GPU) available on the system.
+* `CLArrays.init(dev::OpenCL.cl.Device)` will set the given device to be the active device. If you do not initialize a device explicitly, a default device will be chosen automatically, prioritizing GPU devices over CPU devices.
+* `is_gpu(dev::OpenCL.cl.Device)` returns `true` if the given device is a GPU.
+* `is_cpu(dev::OpenCL.cl.Device)` returns `true` if the given device is a CPU.
+* `gpu_call(kernel::Function, A::GPUArray, args::Tuple, configuration = length(A))` calls the given function on the GPU. See the function documentation for more details.
+
+### Example
 
 ```Julia
 using CLArrays
@@ -21,6 +29,7 @@ for dev in CLArrays.devices()
     CLArrays.init(dev)
     x = zeros(CLArray{Float32}, 5, 5) # create a CLArray on device `dev`
 end
+
 # you can also filter with is_gpu, is_cpu
 gpu_devices = CLArrays.devices(is_gpu)
 ```