List of all items
Structs
- cache::KvBlockPool
- cache::KvCache
- cache::KvPoolExhausted
- cache::PagedKvCache
- cache::PagedKvStore
- cache::PagedStoreExhausted
- expert_store::ExpertKey
- expert_store::ExpertLease
- expert_store::ExpertStore
- expert_store::ExpertStoreStats
- expert_store::FileRangeSource
- instance::InstanceConflict
- instance::InstanceGuard
- instance::InstanceInfo
- kernel_registry::Entry
- kernel_registry::Key
- kernel_registry::Lookup
- kernel_registry::Registry
- kernel_registry::SealReport
- kernel_registry::StrictKernelError
- kv_block::BlockHash
- kv_block::BlockHasher
- kv_disk::DecodedBlock
- kv_disk::DiskConfig
- kv_disk::DiskKvStore
- kv_disk::DiskStats
- kv_disk::ReadHandle
- kv_signature::CacheSignature
- kv_signature::KvBlock
- kv_signature::UnverifiedBlock
- kv_swa::BlockLayout
- tensor::Tensor
Enums
- instance::InstancePolicy
- kernel_registry::Backend
- kernel_registry::Outcome
- kernel_registry::Phase
- kernel_registry::Severity
- kv_disk::BlockFormatError
- kv_disk::StoreError
- kv_signature::KvDtype
- kv_signature::SignatureError
- kv_swa::BlockLayoutError
- weight_matrix::BatchActs
- weight_matrix::QuantKind
- weight_matrix::WeightBytes
- weight_matrix::WeightMatrix
Traits
Functions
- attention::apply_rope
- attention::apply_rope_back
- attention::apply_rope_interleaved
- attention::apply_rope_interleaved_back
- attention::apply_rope_interleaved_with_freq_factors
- attention::apply_rope_with_freq_factors
- attention::causal_gqa_attention
- attention::causal_gqa_attention_paged
- attention::causal_gqa_attention_prefill
- attention::causal_gqa_attention_prefill_shared_kv
- attention::causal_gqa_attention_prefill_shared_kv_windowed
- attention::causal_gqa_attention_sinks
- attention::causal_gqa_attention_softcap
- attention::causal_gqa_attention_windowed
- attention::causal_gqa_attention_windowed_softcap
- attention::causal_mla_attention
- attention::causal_mla_attention_sparse
- attention::lightning_indexer_topk
- csa_hca_compress::channel_gated_pool
- csa_hca_compress::compress_block
- deepseek_v4_attention::csa_attention
- deepseek_v4_attention::hca_attention
- instance::current_backend
- instance::live_instances
- instance::register
- instance::registry_dir
- kernel_registry::enabled
- kernel_registry::global
- kernel_registry::hit
- kernel_registry::miss
- kernel_registry::miss_by_design
- kernel_registry::record_build
- kernel_registry::seal
- kernel_registry::seal_or_error
- kernel_registry::strict
- kernel_registry::verbose
- kv_block::full_blocks
- kv_disk::decode_block
- kv_disk::encode_block
- kv_disk::encoded_len
- kv_swa::aligned_block_size
- matmul::geglu
- matmul::gelu
- matmul::layer_norm
- matmul::matmul_f32
- matmul::rms_norm
- matmul::rms_norm_per_head
- matmul::silu
- matmul::situ_and_mul
- matmul::softcap_inplace
- matmul::swiglu
- threads::current_qos_name
- threads::for_each_chunk_init
- threads::for_each_row
- threads::gemv_num_threads
- threads::init_cpu_pool
- threads::init_gemv_pool
- threads::perf_core_count
- threads::resolve_cpu_threads
- threads::resolve_gemv_threads
- threads::should_parallelize
- turboquant::dequantize_turbo4_groups
- turboquant::fwht_inplace
- turboquant::fwht_orthonormal_inplace
- turboquant::quantize_turbo4_groups
- weight_matrix::active_backend
- weight_matrix::cpu_int_dot_enabled
- weight_matrix::cpu_int_dot_kind_supported
- weight_matrix::cuda_matvec_kind_supported
- weight_matrix::default_cpu_int_dot_on
- weight_matrix::metal_matvec_kind_name
- weight_matrix::metal_mul_mm_kind_supported