List of all items

Structs

• arena::Arena
• autotune::SearchSpace
• autotune::TuneResult
• calibrate::Calibrator
• config::RuntimeConfig
• cost::Cost
• cost::HwModel
• executor::ExternalBuffers
• kernel_config::KernelConfig
• llada2_gate::GateAttrs
• moe_residency::LayerHostBind
• moe_residency::MoeHostBind
• moe_residency::MoeResidencyStats
• moe_topk_capture::MoeTopkCapture
• op_registry::CpuKernelRegistry
• splat::ArenaPrepareArgs
• splat::ArenaRasterizeArgs
• splat::ArenaRenderArgs
• splat::ArenaRenderBwdArgs
• splat::HostBackwardArgs
• splat::HostRenderArgs
• thunk::ThunkSchedule
• tile::Coord2
• tile::RowMajorTile
• tile::StridedTile
• tile::Strides2
• tile::Tile2

Enums

• asm_check::AsmCheckError
• dispatch::Override
• kernel_config::CpuArch
• kernel_config::OpClass
• op_registry::CpuTensorMut
• op_registry::CpuTensorRef
• thunk::Thunk

Traits

• op_registry::CpuKernel
• tile::TileIO

Functions

• asm_check::assert_function_contains
• asm_check::disassemble_self
• asm_check::function_section
• attention_bwd::attention_backward
• autotune::active_config
• autotune::autotune
• blas::bias_add
• blas::dgemm
• blas::dgeqrf_full
• blas::dgesv
• blas::dgesvd_thin
• blas::dpotrf
• blas::dsyevd
• blas::dtrsm_lower_or_upper
• blas::par_sgemm
• blas::par_sgemm_bias
• blas::sgemm
• blas::sgemm_accumulate
• blas::sgemm_auto
• blas::sgemm_bias
• blas::sgemm_bias_epilogue
• blas::sgemm_bt
• blas::sgemm_epilogue
• blas::sgemm_general
• blas::sgemm_strided
• blas::sgemv_at
• blas::sger
• blas::sgesv
• blas::sscal
• cost::hw_model
• dequant_cache::clear_dequant_cache
• dequant_cache::gguf_weight_f32
• dispatch::resolve
• dispatch::resolve_current
• dispatch::set_override
• executor::execute
• gdn::gdn_step_blas
• gguf_matmul::dequant_moe_weights_to_grouped_f32
• gguf_matmul::gguf_grouped_matmul_bt
• gguf_matmul::gguf_matmul_bt
• gguf_matmul::gguf_matmul_bt_parallel
• gguf_matmul::grouped_moe_sort_plan
• gguf_matmul::grouped_moe_unpermute_out
• kernel_config::current_config
• kernel_config::kernel_config_for
• kernels::bias_gelu
• kernels::conv_transpose2d_nchw
• kernels::gelu_approx_inplace
• kernels::gelu_inplace
• kernels::group_norm_nchw
• kernels::layer_norm2d_nchw
• kernels::layer_norm_row
• kernels::neon_sgemm_bias_small
• kernels::neon_sgemm_small
• kernels::neon_softmax
• kernels::par_bias_gelu
• kernels::par_gelu_approx_inplace
• kernels::par_gelu_inplace
• kernels::par_residual_bias_ln
• kernels::par_silu_inplace
• kernels::residual_bias_layer_norm
• kernels::residual_bias_rms_norm
• kernels::resize_nearest_2x_nchw
• kernels::scalar_gelu_approx
• kernels::silu_inplace
• llada2_gate::execute_gate_f32
• llada2_gate::execute_gate_in_f32_arena
• llada2_gate::group_limited_topk
• lm_head::f32_tied_lm_argmax
• lm_head::f32_tied_lm_topk
• lm_head::gguf_tied_lm_argmax
• lm_head::gguf_tied_lm_topk
• moe_residency::bind_host_weights
• moe_residency::clear_mask
• moe_residency::expert_on_device
• moe_residency::expert_on_device_for_layer
• moe_residency::host_expert_weight_ptr
• moe_residency::next_gmm_ord
• moe_residency::peek_stats
• moe_residency::record_expert_tokens
• moe_residency::reset_gmm_counters
• moe_residency::set_mask
• moe_residency::set_per_layer_masks
• moe_residency::take_last_forward_stats
• moe_residency::take_stats
• naive::bias_add
• naive::erf
• naive::gelu
• naive::layer_norm
• naive::matmul
• naive::matmul_bias_gelu
• naive::sdpa
• naive::sdpa_no_mask
• naive::silu
• naive::softmax
• op_registry::global_cpu_kernels
• op_registry::lookup_cpu_kernel
• op_registry::register_cpu_kernel
• pool::num_threads
• pool::par_for
• splat::backward_host_slices
• splat::execute_gaussian_splat_prepare
• splat::execute_gaussian_splat_rasterize
• splat::execute_gaussian_splat_render
• splat::execute_gaussian_splat_render_backward
• splat::register_splat_executors
• splat::render_host_slices
• thunk::compile_thunks
• thunk::dequant_matmul_nvfp4
• thunk::execute_axial_rope2d_f32
• thunk::execute_compiled
• thunk::execute_conv_transpose2d_nchw_f32
• thunk::execute_cumsum_backward_f32
• thunk::execute_dequant_grouped_matmul_gguf_f32
• thunk::execute_dequant_matmul_fp8_f32
• thunk::execute_dequant_matmul_gguf_f32
• thunk::execute_dequant_matmul_int4_f32
• thunk::execute_dequant_matmul_nvfp4_f32
• thunk::execute_fft1d_f32
• thunk::execute_fft1d_f64
• thunk::execute_gated_delta_net_f16
• thunk::execute_gated_delta_net_f32
• thunk::execute_gather_backward_f32
• thunk::execute_group_norm_nchw_f32
• thunk::execute_layer_norm2d_nchw_f32
• thunk::execute_resize_nearest_2x_f32
• thunk::execute_rms_norm_backward_beta_f32
• thunk::execute_rms_norm_backward_gamma_f32
• thunk::execute_rms_norm_backward_input_f32
• thunk::execute_rope_backward_f32
• thunk::execute_thunks
• thunk::execute_thunks_active
• tile::copy_tile
• tile::for_each_coord
• training_bwd::cumsum_backward_row
• training_bwd::gather_axis_backward
• training_bwd::group_norm_backward_beta_nchw
• training_bwd::group_norm_backward_gamma_nchw
• training_bwd::group_norm_backward_input_nchw
• training_bwd::rms_norm_backward_row
• training_bwd::rope_backward_row

Constants

• gdn::GDN_MAX_STATE