Index

A

Abs (class in vortex_torch.cache.elementwise) (class in vortex_torch.indexer.elementwise) (ElementwiseOpType attribute) Add (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) (ElementwiseBinaryOpType attribute) add_aux_flops() (ContextBase method)

add_aux_memory() (ContextBase method) Add_Mul (class in vortex_torch.cache.elementwise) (class in vortex_torch.indexer.elementwise) (ElementwiseOpType attribute) approxTopK (class in vortex_torch.indexer.output_func) as_vtensor() (in module vortex_torch.abs.tensor) assert_created() (ContextBase method) auxilary_func_def_lines (Context attribute), [1]

B

batch_size (Context property) (MetaData attribute) BATCHED (FORMAT attribute), [1] BF16 (QuantizationType attribute)

block_reserved_bos (Context attribute) block_reserved_eos (Context attribute) block_size (Context attribute), [1] BlockSparseAttention (class in vortex_torch.flow.algorithms) build_vflow() (in module vortex_torch.flow.loader)

C

CenteredBlockSparseAttention (class in vortex_torch.flow.algorithms) clear_aux_flops() (ContextBase method) clear_aux_memory() (ContextBase method) compilation_cache_dir (Context attribute), [1] compilation_header_lines (Context attribute), [1] Context (class in vortex_torch.cache.context) (class in vortex_torch.indexer.context)

ContextBase (class in vortex_torch.abs.context_base) Conv1d (class in vortex_torch.indexer.scan) create() (Context method), [1] (ContextBase method) create_cache() (vFlow method) created (ContextBase property)

D

dense_block_tables (MetaData attribute) dense_kv_indices (MetaData attribute) dense_kv_indptr (MetaData attribute) dense_seqlens (MetaData attribute)

device (vTensor attribute) dim() (vTensor method) dtype (vTensor attribute) DTYPE_STR_TO_TORCH (in module vortex_torch.utils)

E

Elementwise (class in vortex_torch.cache.elementwise) (class in vortex_torch.indexer.elementwise) Elementwise_Binary (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) ElementwiseBinaryOpType (class in vortex_torch.utils) ElementwiseOpType (class in vortex_torch.utils)

execute (Mode attribute) execute() (ContextBase method) (vOp method) Exp (class in vortex_torch.cache.elementwise) (class in vortex_torch.indexer.elementwise) (ElementwiseOpType attribute)

F

Fill (class in vortex_torch.cache.fill) FlowInitError FlowLoadError FORMAT (class in vortex_torch.abs.tensor)

forward_cache() (vFlow method) forward_indexer() (vFlow method) FP8_E4M3 (QuantizationType attribute) FP8_E5M2 (QuantizationType attribute)

G

GeMM (class in vortex_torch.cache.matmul) (class in vortex_torch.indexer.matmul) GeMV (class in vortex_torch.indexer.matmul) get() (in module vortex_torch.flow.registry) get_cache_meta_info() (vFlow method) get_chunkwise_hn2nh_transpose() (in module vortex_torch.indexer.utils_sglang) get_chunkwise_nh2hn_transpose() (in module vortex_torch.indexer.utils_sglang) get_ctx() (in module vortex_torch.cache.context) (in module vortex_torch.indexer.context)

get_decode_planner() (in module vortex_torch.indexer.utils_sglang) get_decode_planner_trtllm() (in module vortex_torch.indexer.utils_sglang) get_prefill_planner() (in module vortex_torch.indexer.utils_sglang) get_token_ratio() (vFlow method) GQABlockSparseAttention (class in vortex_torch.flow.algorithms) GQAQuestSparseAttention (class in vortex_torch.flow.algorithms) group_size (Context attribute)

H

has() (in module vortex_torch.flow.registry)

head_dim (Context attribute), [1] head_num (Context attribute)

I

impl_backend (Context attribute), [1] indent_block() (in module vortex_torch.utils)

initialize() (vFlow method) is_hopper() (in module vortex_torch.utils) is_hopper_or_newer() (in module vortex_torch.utils)

K

Kron (class in vortex_torch.indexer.kron)

kv_last_page_len (MetaData attribute)

L

L2Norm (class in vortex_torch.cache.reduce) (class in vortex_torch.indexer.reduce) (ReduceType attribute) L2NormInterleave (class in vortex_torch.cache.reduce_interleave) list_keys() (in module vortex_torch.flow.registry)

Load (class in vortex_torch.indexer.save_load) Log (class in vortex_torch.cache.elementwise) (class in vortex_torch.indexer.elementwise) (ElementwiseOpType attribute) LServeCentroidSparseAttention (class in vortex_torch.flow.algorithms) LServeSparseAttention (class in vortex_torch.flow.algorithms)

M

MaskedQuestSparseAttention (class in vortex_torch.flow.algorithms) MaskSlice (class in vortex_torch.cache.mask) (class in vortex_torch.indexer.mask) Max (class in vortex_torch.cache.reduce) (class in vortex_torch.indexer.reduce) (ReduceType attribute) max_bs (Context attribute) max_new_tokens_per_batch (Context attribute) max_num_blocks (Context attribute) max_num_blocks_per_request (Context attribute) max_num_pages (Context attribute) max_num_pages_per_request (Context attribute) max_num_workloads (Context attribute) max_topk_val (Context attribute) Maximum (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) (ElementwiseBinaryOpType attribute) MaxInterleave (class in vortex_torch.cache.reduce_interleave) Mean (class in vortex_torch.cache.reduce) (class in vortex_torch.indexer.reduce) (ReduceType attribute) MeanInterleave (class in vortex_torch.cache.reduce_interleave) MetaData (class in vortex_torch.indexer.context) metadata (Context attribute) Min (class in vortex_torch.cache.reduce) (class in vortex_torch.indexer.reduce) (ReduceType attribute) Minimum (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) (ElementwiseBinaryOpType attribute) MinInterleave (class in vortex_torch.cache.reduce_interleave) missing() (ContextBase method) Mode (class in vortex_torch.utils) mode (Context attribute), [1] (ContextBase attribute) module vortex_torch vortex_torch.abs vortex_torch.abs.context_base vortex_torch.abs.op vortex_torch.abs.tensor vortex_torch.cache vortex_torch.cache.context vortex_torch.cache.elementwise vortex_torch.cache.elementwise_binary vortex_torch.cache.fill vortex_torch.cache.mask vortex_torch.cache.matmul vortex_torch.cache.reduce vortex_torch.cache.reduce_interleave vortex_torch.cache.reshape vortex_torch.cache.triton_kernels vortex_torch.cache.triton_kernels.elementwise_binary_impl vortex_torch.cache.triton_kernels.elementwise_impl vortex_torch.cache.triton_kernels.fill_impl vortex_torch.cache.triton_kernels.matmul_impl vortex_torch.cache.triton_kernels.reduce_impl vortex_torch.cache.triton_kernels.set_kv vortex_torch.flow vortex_torch.flow.algorithms vortex_torch.flow.flow vortex_torch.flow.loader vortex_torch.flow.registry vortex_torch.indexer vortex_torch.indexer.context vortex_torch.indexer.elementwise vortex_torch.indexer.elementwise_binary vortex_torch.indexer.kron vortex_torch.indexer.mask vortex_torch.indexer.matmul vortex_torch.indexer.output_func vortex_torch.indexer.reduce vortex_torch.indexer.reshape vortex_torch.indexer.save_load vortex_torch.indexer.scan vortex_torch.indexer.select vortex_torch.indexer.transpose vortex_torch.indexer.triton_kernels vortex_torch.indexer.triton_kernels.elementwise_binary_impl vortex_torch.indexer.triton_kernels.elementwise_impl vortex_torch.indexer.triton_kernels.matmul_impl vortex_torch.indexer.triton_kernels.mv_impl vortex_torch.indexer.triton_kernels.normalize_impl vortex_torch.indexer.triton_kernels.reduce_impl vortex_torch.indexer.triton_kernels.save_load_impl vortex_torch.indexer.triton_kernels.softmax_impl vortex_torch.indexer.triton_kernels.transpose_impl vortex_torch.indexer.triton_kernels.utils_impl vortex_torch.indexer.utils_sglang vortex_torch.utils vortex_torch.version

Mul (ElementwiseBinaryOpType attribute) Multiply (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary)

N

name (Context attribute), [1] (ContextBase attribute) ndim (vTensor property) needs_padding() (vTensor method) Normalize (class in vortex_torch.indexer.scan)

num_blocks_per_page (Context attribute), [1] num_kv_heads (Context attribute) num_pages_per_workload (Context attribute) num_qo_heads (Context attribute) num_sms (Context attribute), [1] numel() (vTensor method)

O

op_list (Context attribute), [1] op_to_input_tensor_list (Context attribute), [1]

op_to_output_tensor_list (Context attribute), [1] output_tensor_to_op_list (Context attribute), [1]

P

padded_shape (vTensor attribute) page_size (Context attribute), [1]

PAGED (FORMAT attribute), [1] preallocate() (MetaData class method)

Q

QuantizationType (class in vortex_torch.utils)

query_arg_names (Context attribute)

R

RAGGED (FORMAT attribute), [1] Reduce (class in vortex_torch.cache.reduce) (class in vortex_torch.indexer.reduce) ReduceInterleave (class in vortex_torch.cache.reduce_interleave) ReduceType (class in vortex_torch.utils) register() (in module vortex_torch.flow.registry) RegistryError

Relu (class in vortex_torch.cache.elementwise) (class in vortex_torch.indexer.elementwise) (ElementwiseOpType attribute) Reshape (class in vortex_torch.cache.reshape) (class in vortex_torch.indexer.reshape) resolve_dtype() (in module vortex_torch.utils) RunningAvgBlockSparse (class in vortex_torch.flow.algorithms)

S

S (Schedule attribute) Save (class in vortex_torch.indexer.save_load) Schedule (class in vortex_torch.utils) schedule (vOp attribute) set_batch_size() (Context method) (MetaData method) shape (vTensor attribute) side_effect_op_ids (Context attribute) Sigmoid (class in vortex_torch.cache.elementwise) (class in vortex_torch.indexer.elementwise) (ElementwiseOpType attribute) Silu (class in vortex_torch.cache.elementwise) (class in vortex_torch.indexer.elementwise) (ElementwiseOpType attribute)

size() (vTensor method) Softmax (class in vortex_torch.indexer.scan) sparse_attention_name (Context attribute), [1] sparse_block_tables (MetaData attribute) sparse_kv_indices (MetaData attribute) sparse_kv_indptr (MetaData attribute) sparse_seqlens (MetaData attribute) Sum (class in vortex_torch.indexer.reduce) (ReduceType attribute) summary() (ContextBase method)

T

tensor_id (vTensor attribute) tensor_id_to_tensor_name_map (Context attribute), [1] tensor_list (Context attribute), [1] topK (class in vortex_torch.indexer.output_func) TopK (class in vortex_torch.indexer.select)

topk_ratio (Context attribute) topk_val (Context attribute) total_num_blocks (Context attribute) total_num_pages (Context attribute) Transpose (class in vortex_torch.indexer.transpose)

U

Union (class in vortex_torch.indexer.output_func)

use_tensor_core (Context attribute)

V

VEnergyGatedCentroid (class in vortex_torch.flow.algorithms) vFlow (class in vortex_torch.flow.flow) vOp (class in vortex_torch.abs.op) vortex_attention_backend (Context attribute) vortex_dtype (Context attribute), [1] (ContextBase attribute) vortex_torch module vortex_torch.abs module vortex_torch.abs.context_base module vortex_torch.abs.op module vortex_torch.abs.tensor module vortex_torch.cache module vortex_torch.cache.context module vortex_torch.cache.elementwise module vortex_torch.cache.elementwise_binary module vortex_torch.cache.fill module vortex_torch.cache.mask module vortex_torch.cache.matmul module vortex_torch.cache.reduce module vortex_torch.cache.reduce_interleave module vortex_torch.cache.reshape module vortex_torch.cache.triton_kernels module vortex_torch.cache.triton_kernels.elementwise_binary_impl module vortex_torch.cache.triton_kernels.elementwise_impl module vortex_torch.cache.triton_kernels.fill_impl module vortex_torch.cache.triton_kernels.matmul_impl module vortex_torch.cache.triton_kernels.reduce_impl module vortex_torch.cache.triton_kernels.set_kv module vortex_torch.flow module vortex_torch.flow.algorithms module vortex_torch.flow.flow module vortex_torch.flow.loader module

vortex_torch.flow.registry module vortex_torch.indexer module vortex_torch.indexer.context module vortex_torch.indexer.elementwise module vortex_torch.indexer.elementwise_binary module vortex_torch.indexer.kron module vortex_torch.indexer.mask module vortex_torch.indexer.matmul module vortex_torch.indexer.output_func module vortex_torch.indexer.reduce module vortex_torch.indexer.reshape module vortex_torch.indexer.save_load module vortex_torch.indexer.scan module vortex_torch.indexer.select module vortex_torch.indexer.transpose module vortex_torch.indexer.triton_kernels module vortex_torch.indexer.triton_kernels.elementwise_binary_impl module vortex_torch.indexer.triton_kernels.elementwise_impl module vortex_torch.indexer.triton_kernels.matmul_impl module vortex_torch.indexer.triton_kernels.mv_impl module vortex_torch.indexer.triton_kernels.normalize_impl module vortex_torch.indexer.triton_kernels.reduce_impl module vortex_torch.indexer.triton_kernels.save_load_impl module vortex_torch.indexer.triton_kernels.softmax_impl module vortex_torch.indexer.triton_kernels.transpose_impl module vortex_torch.indexer.triton_kernels.utils_impl module vortex_torch.indexer.utils_sglang module vortex_torch.utils module vortex_torch.version module vTensor (class in vortex_torch.abs.tensor)

W

W (Schedule attribute) WhereEqual (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) (ElementwiseBinaryOpType attribute) WhereGreater (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) (ElementwiseBinaryOpType attribute) WhereGreaterEqual (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) (ElementwiseBinaryOpType attribute) WhereLess (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) (ElementwiseBinaryOpType attribute)

WhereLessEqual (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) (ElementwiseBinaryOpType attribute) WhereNotEqual (class in vortex_torch.cache.elementwise_binary) (class in vortex_torch.indexer.elementwise_binary) (ElementwiseBinaryOpType attribute) winfo_chunk_size (MetaData attribute) winfo_is_first_workload_per_batch (MetaData attribute) winfo_kv_lens (MetaData attribute) winfo_kv_offsets (MetaData attribute) winfo_num_workloads (MetaData attribute) winfo_q_indices (MetaData attribute) workload_chunk_size (Context attribute)