Skip to content

Store ROCArray offset in bytes instead of elements #898

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
maleadt merged 1 commit into from
Apr 10, 2026
+13 −4
Merged

Store ROCArray offset in bytes instead of elements #898

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 4 additions & 4 deletions src/array.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
mutable struct ROCArray{T, N, B} <: AbstractGPUArray{T, N}
buf::DataRef{Managed{B}}
dims::Dims{N}
offset::Int # Offset is in number of elements (not bytes).
offset::Int # offset of the data in memory, in bytes

function ROCArray{T, N, B}(::UndefInitializer, dims::Dims{N}) where {T, N, B <: Mem.AbstractAMDBuffer}
check_eltype("ROCArray", T)
Expand Down Expand Up @@ -92,7 +92,7 @@ GPUArrays.storage(a::ROCArray) = a.buf

function GPUArrays.derive(::Type{T}, x::ROCArray, dims::Dims{N}, offset::Int) where {N, T}
ref = copy(x.buf)
offset += (x.offset * Base.elsize(x)) ÷ aligned_sizeof(T)
offset = x.offset + offset * aligned_sizeof(T)
ROCArray{T, N}(ref, dims; offset)
end

Expand Down Expand Up @@ -298,7 +298,7 @@ Adapt.adapt_storage(::Float32Adaptor, xs::AbstractArray{Float16}) =
roc(xs) = adapt(Float32Adaptor(), xs)

Base.unsafe_convert(typ::Type{Ptr{T}}, x::ROCArray{T}) where T =
convert(typ, x.buf[]) + x.offset * aligned_sizeof(T)
convert(typ, x.buf[]) + x.offset

# some nice utilities

Expand Down Expand Up @@ -351,7 +351,7 @@ function Base.convert(
buf = convert(Mem.AbstractAMDBuffer, a.buf[])
ptr = convert(Ptr{T}, typeof(buf) <: Mem.HIPBuffer ?
buf : buf.dev_ptr)
llvm_ptr = AMDGPU.LLVMPtr{T,AS.Global}(ptr + a.offset * aligned_sizeof(T))
llvm_ptr = AMDGPU.LLVMPtr{T,AS.Global}(ptr + a.offset)
ROCDeviceArray{T, N, AS.Global}(a.dims, llvm_ptr)
end

Expand Down
9 changes: 9 additions & 0 deletions test/core/rocarray_base.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,6 +52,15 @@ end
@test collect(c) == 3:4
end

@testset "reinterpret of view with non-aligned offset" begin
# reinterpreting a view to a larger element type where the byte offset
# is not a multiple of the new element size
a = ROCArray(Int32[1,2,3,4,5,6,7,8,9])
v = view(a, 2:7) # offset of 1 Int32 = 4 bytes
r = reinterpret(Int64, v) # Int64 = 8 bytes; 4 is not a multiple of 8
@test Array(r) == reinterpret(Int64, @view Array(a)[2:7])
end

@testset "resize!" begin
a_h = Array(range(1, 10))
a_d = a_h |> roc
Expand Down