MemoryViews.jl

Simple, low-level views into memory in Julia
Author BioJulia
Popularity
17 Stars
Updated Last
5 Months Ago
Started In
May 2024

MemoryViews

Latest Release MIT license Documentation

This package implements MemoryView, a simple, low-level view into a chunk of Memory, as well as the MemoryKind trait to guide dispatch of generic methods to memory views. It is intended to be used as a foundational base for other packages.

To learn how to use the package, read the documentation

Example

Basic usage

# Create complex nested types backed by memory
v1 = view(codeunits("abc"), :)
v2 = view([0x61, 0x62, 0x63], Base.OneTo(3))

# This call hits the slow, generic fallback implementation in Base,
# because it's very difficult to correctly cover all possible
# combinations of types in the method
copyto!(v2, v1)

# These are ordinary `AbstractVector`s, in this case with
# element type UInt8.
mem1 = MemoryView(v1)
mem2 = MemoryView(v2)

# Both views are `MemoryView{UInt8}`, which has the fast path
# implemented. Precisely because we represent "memory" as a simple,
# concrete types, it's easier to provide these kinds of guarantees.
copyto!(mem2, mem1)

# Use the memory views as ordinary vectors
fst = mem1[1]
reverse!(mem1) # ... etc

Dispatching to MemoryView

function foo(x::ImmutableMemoryView)
    # low-level implementation
end

function foo(::NotMemory, x::AbstractArray)
    # slow, generic fallback
end

# Dispatch with the `MemoryKind` trait
foo(::IsMemory, x) = foo(ImmutableMemoryView(x))
foo(x) = foo(MemoryKind(typeof(x)), x)

# Optionally: Also support strings
foo(x::AbstractString) = foo(codeunits(x))

Good to know

  • MemoryViews.jl currently only works on the master branch of Julia.
  • Slicing a memory view produces a memory view - it does not copy.

Limitations

  • Many optimised fast methods for more established types like Vector are missing for MemoryView. These are added over time. Please make an issue or a PR as you encounter missing methods.

  • Currently, MemoryView does not make use of Core.GenericMemory's additional parameters, such as atomicity or address space. This may easily be added with a GenericMemoryView type, similar to Memory / GenericMemory.

  • I can't figure out how to support reinterpreted arrays. Any way I can think of doing so will sigificantly complicate MemoryView, which takes away some of the appeal of this type's simplicity. It's possible that reinterpreted arrays are so outside Julia's ordinary memory management that this simply can't be done.

  • Currently, Strings are not backed by Memory in Julia. Therefore, creating a MemoryView of a string requires heap-allocating a new Memory pointing to the existing memory of the string. This can be fixed if String is re-implemented to be backed by Memory, but I don't know enough details about the implementation of String to know if this is practical.

Contributing

Make an issue or PR on this repository, or get in touch with the BioJulia community over at the Julia Slack or Zulip servers.