tess::gpu::FieldMirrorDesc Struct Reference

#include <tess/gpu/descriptors.h>

Public Member Functions
constexpr auto	total_bytes_fits () const noexcept -> bool
constexpr auto	total_bytes () const noexcept -> std::uint64_t

Public Attributes
std::uint32_t	field_index = 0
GpuFieldFormat	format = GpuFieldFormat::U8
std::uint32_t	value_bytes = 0
std::uint64_t	tiles_per_chunk = 0
std::uint64_t	bytes_per_chunk = 0
std::uint64_t	chunk_count = 0

Friends
constexpr auto	operator== (const FieldMirrorDesc &, const FieldMirrorDesc &) noexcept -> bool=default

Detailed Description

Byte layout of a dense, chunk-key-major field mirror.

A backend can size one buffer containing chunk_count contiguous slices of bytes_per_chunk. Selective sparse mirrors may reuse the layout metadata while choosing a different chunk set.

Member Function Documentation

total_bytes()

auto tess::gpu::FieldMirrorDesc::total_bytes ( ) const -> std::uint64_t

inlinenodiscardconstexprnoexcept

Returns the dense byte count, saturating when a manual descriptor wraps.

Generated descriptors are compile-time checked and never saturate. The saturation keeps malformed consumer-created descriptors from appearing to request a small, valid buffer after unsigned multiplication wraps.

total_bytes_fits()

auto tess::gpu::FieldMirrorDesc::total_bytes_fits ( ) const -> bool

inlinenodiscardconstexprnoexcept

Returns whether the dense mirror byte count is representable.

---

The documentation for this struct was generated from the following file:

• include/tess/gpu/descriptors.h