diff --git a/lib/cudadrv/devices.jl b/lib/cudadrv/devices.jl
index 543524cd19..89f1d061e6 100644
--- a/lib/cudadrv/devices.jl
+++ b/lib/cudadrv/devices.jl
@@ -9,7 +9,7 @@ export
 
 Get a handle to a compute device.
 """
-struct CuDevice
+struct CuDevice <: AbstractGPUDevice
     handle::CUdevice
 
     function CuDevice(ordinal::Integer)
diff --git a/lib/cudadrv/events.jl b/lib/cudadrv/events.jl
index 8c4192d9c8..2d7e838e2a 100644
--- a/lib/cudadrv/events.jl
+++ b/lib/cudadrv/events.jl
@@ -133,3 +133,6 @@ macro elapsed(ex)
         elapsed(t0, t1)
     end
 end
+
+
+Adapt.get_compute_unit_impl(@nospecialize(TypeHistory::Type), e::CuEvent) = device(e.ctx)
diff --git a/lib/cusparse/CUSPARSE.jl b/lib/cusparse/CUSPARSE.jl
index 3fddf946f3..8b84303c25 100644
--- a/lib/cusparse/CUSPARSE.jl
+++ b/lib/cusparse/CUSPARSE.jl
@@ -11,7 +11,8 @@ using CEnum: @cenum
 using LinearAlgebra
 using LinearAlgebra: HermOrSym
 
-using Adapt: Adapt, adapt
+import Adapt
+using Adapt: Adapt, adapt, AbstractGPUDevice
 
 using SparseArrays
 
diff --git a/lib/cusparse/array.jl b/lib/cusparse/array.jl
index 34d52c0a85..333ef13776 100644
--- a/lib/cusparse/array.jl
+++ b/lib/cusparse/array.jl
@@ -15,6 +15,9 @@ const AbstractCuSparseMatrix{Tv, Ti} = AbstractCuSparseArray{Tv, Ti, 2}
 
 Base.convert(T::Type{<:AbstractCuSparseArray}, m::AbstractArray) = m isa T ? m : T(m)
 
+Adapt.get_compute_unit_impl(@nospecialize(TypeHistory::Type), A::AbstractCuSparseArray) = device(A.nzVal)
+
+
 mutable struct CuSparseVector{Tv, Ti} <: AbstractCuSparseVector{Tv, Ti}
     iPtr::CuVector{Ti}
     nzVal::CuVector{Tv}
diff --git a/src/CUDA.jl b/src/CUDA.jl
index 395f062d6c..7ad5508c6a 100644
--- a/src/CUDA.jl
+++ b/src/CUDA.jl
@@ -8,7 +8,8 @@ using LLVM
 using LLVM.Interop
 using Core: LLVMPtr
 
-using Adapt: Adapt, adapt, WrappedArray
+import Adapt
+using Adapt: Adapt, adapt, WrappedArray, AbstractGPUDevice
 
 using Requires: @require
 
diff --git a/src/array.jl b/src/array.jl
index c72cca5ce7..e4e0f8b1b3 100644
--- a/src/array.jl
+++ b/src/array.jl
@@ -244,6 +244,13 @@ function device(A::CuArray)
   return device(A.storage.buffer.ctx)
 end
 
+Adapt.get_compute_unit_impl(@nospecialize(TypeHistory::Type), A::CuArray) = device(A)
+
+Adapt.adapt_storage(dev::CuDevice, x) = device!(() -> Adapt.adapt_storage(CuArray, x), dev)
+
+Sys.total_memory(dev::CuDevice) = CUDA.totalmem(dev)
+Sys.free_memory(dev::CuDevice) = unsigned(CUDA.device!(CUDA.available_memory, dev))
+
 
 ## derived types