Support MadNLPGPU.jl v0.7.13

Author: Alexis Montoison

Project.toml

@@ -21,7 +21,7 @@ FFTW = "1.8.1"
 Krylov = "0.10.1"
 LinearAlgebra = "1.10"
 MadNLP = "0.8.9"
-MadNLPGPU = "0.7.12"
+MadNLPGPU = "0.7.13"
 NLPModels = "0.21.5"
 Random = "1.10"
 SparseArrays = "1.10"

src/gpu_utils.jl

@@ -36,7 +36,7 @@ end
     ROCSparseMatrixCSC to ROCMatrix
 =#
 
-function MadNLPGPU.transfer!(y::ROCMatrix{T}, x::ROCSparseMatrixCSC{T}) where {T}
+function MadNLPGPU.gpu_transfer!(y::ROCMatrix{T}, x::ROCSparseMatrixCSC{T}) where {T}
     n = size(y, 2)
     fill!(y, zero(T))
     backend = ROCBackend()