Expand description

Implementation agnostic memory allocator for Vulkan like APIs.

This crate is intended to be used as part of safe API implementations.
Use with caution. There are unsafe functions all over the place.


Start with fetching DeviceProperties from gpu-alloc-<backend> crate for the backend of choice.
Then create GpuAllocator instance and use it for all device memory allocations.
GpuAllocator will take care for all necessary bookkeeping like memory object count limit, heap budget and memory mapping.

Backends implementations

Backend supporting crates should not depend on this crate.
Instead they should depend on gpu-alloc-types which is much more stable, allowing to upgrade gpu-alloc version without gpu-alloc-<backend> upgrade.


Allocation flags

Configuration for GpuAllocator

Properties of the device that will be used for allocating memory objects.

Memory allocator for Vulkan-like APIs.

Specifies range of the mapped memory region.

Memory block allocated by GpuAllocator.

Defines memory heap.

Memory properties type.

Defines memory type.

Memory request for allocator.

Memory usage type. Bits set define intended usage for requested memory.


Enumeration of possible errors that may occur during memory allocation.

Hints for allocator to decide on allocation strategy.

Memory mapped error.

Enumeration of possible errors that may occur during memory mapping.

Memory exhausted error.


Abstract device that can be used to allocate memory objects.