pub struct AdapterInfo {
pub vendor: String,
pub architecture: String,
pub device: String,
pub description: String,
pub is_fallback: bool,
pub shader_f16: bool,
pub subgroups: bool,
pub max_buffer_size: u64,
pub max_storage_buffer_binding_size: u64,
pub max_compute_workgroup_size_x: u32,
pub max_compute_invocations_per_workgroup: u32,
}Expand description
What a host learned about its adapter, in WebGPU’s own vocabulary.
The fields mirror GPUAdapterInfo, GPUSupportedLimits, and the two
optional GPUFeatureNames that matter for compute. WebGPU deliberately
does not expose PCI vendor/device IDs — vendor and architecture are
lowercase strings such as "nvidia" / "ampere", and may be empty when
the browser chooses to withhold them for fingerprinting reasons.
Fields§
§vendor: StringGPUAdapterInfo.vendor, e.g. "nvidia", "amd", "intel", "apple".
architecture: StringGPUAdapterInfo.architecture, e.g. "ampere", "rdna-3".
device: StringGPUAdapterInfo.device — often empty in browsers.
description: StringGPUAdapterInfo.description — often empty in browsers.
is_fallback: boolGPUAdapter.isFallbackAdapter — a software rasteriser.
shader_f16: boolThe "shader-f16" feature was granted on the device.
subgroups: boolThe "subgroups" feature was granted on the device.
max_buffer_size: u64GPUSupportedLimits.maxBufferSize.
max_storage_buffer_binding_size: u64GPUSupportedLimits.maxStorageBufferBindingSize.
max_compute_workgroup_size_x: u32GPUSupportedLimits.maxComputeWorkgroupSizeX.
max_compute_invocations_per_workgroup: u32GPUSupportedLimits.maxComputeInvocationsPerWorkgroup.
Trait Implementations§
Source§impl Clone for AdapterInfo
impl Clone for AdapterInfo
Source§fn clone(&self) -> AdapterInfo
fn clone(&self) -> AdapterInfo
1.0.0 (const: unstable) · Source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
source. Read more