We should consider renaming this to float4 or vec4 (and ditto for uint4, etc.)

We also may want to use the <> syntax for something else in the future.

I have discovered this package and being a new user went straight to the docs to see how to use each of these functions. I tried as.gpuMatrix and naturally would think as..... is converting whatever I give it to a gpuMatrix so I tried:

gpuR::as.gpuMatrix(c(1:3),"vector")

This results in:

Error in (function (classes, fdef, mtable) :
unable to find an inherited method for

In the array example program, we're copying data into arrays and from arrays. While we're checking the CUDA runtime API return values, we're not checking that the copying results in the expected data. We may be printing information from which it's possible to deduce that, but not actually checking.

Let's do that... @codecircuit ?

The build instructions mention that the JAVA home variable must be set, but doesn't have any examples

Make sure the environment variable JAVA_HOME is defined and pointing a to a valid JDK installation. If JAVA_HOME is not defined, the build will fail.

This should be updated to clarify how this variable should be configured.