Incidentally this second method doesn't work.
On Monday, October 26, 2015 at 1:46:03 PM UTC, Kristoffer Carlsson wrote:
>
> Or if you want to manually handle the devices (device 0 in this case):
>
> CUDArt.init([0])
> C = CUDArt.CudaArray(Float64, (10,10))
> fill!(C, 2.0)
> println(to_host(C))
> CUDArt.close([0])
>
>
julia> using CUDArt
julia> CUDArt.init([0])
ERROR: UndefVarError: init not defined
julia> C = CUDArt.CudaArray(Float64, (10,10))
CUDArt.CudaArray{Float64,2}(CUDArt.CudaPtr{Float64}(Ptr{Float64} @
0x0000000b034a0c00),(10,10),0)
julia> fill!(C, 2.0)
ERROR: KeyError: (0,"fill_contiguous",Float64) not found
[inlined code] from essentials.jl:58
in getindex at dict.jl:719
in fill! at /home/mcp50/.julia/v0.5/CUDArt/src/arrays.jl:158
julia> println(to_host(C))
[-131072.0011474827 8.290725244651441e-76 -131072.66519424706 -
131072.79735252648 7.114269581235735e81 6.805647992284872e38
6.805647850367426e38 -2.376364460420424e-212 7.124662174923278e38
5.370009195891395e-152
[...]
-131072.26610348016 9.67251204874471e-76 8.290722196472935e-76
3.022659974162853e-77 1.2624824795916418e-190 6.1835059447865614e-229
5.075888440444962e-116 7.114269581235735e81 2.190889012369281e-154 -
6.805647751035016e38]
julia> CUDArt.close([0])
ERROR: MethodError: `close` has no method matching close(::Array{Int64,1})