Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/LMCache/LMCache
/ functions
Functions
13,371 in github.com/LMCache/LMCache
⨍
Functions
13,371
◇
Types & classes
2,044
↳
Endpoints
181
↓ 1 callers
Function
create_metadata
Create metadata for the LMCacheEngine.
tests/disagg/test_nixl_cache_engine.py:121
↓ 1 callers
Function
create_pending_request
Create an asyncio.Event for a request and store it.
examples/disagg_prefill_mp/disagg_proxy_server.py:215
↓ 1 callers
Function
create_prefetch_policy
Create a prefetch policy instance by name. Args: name: Registered policy name. Returns: A new PrefetchPolicy instance.
lmcache/v1/distributed/storage_controllers/prefetch_policy.py:110
↓ 1 callers
Function
create_record_strategy
(config)
lmcache/v1/lookup_client/record_strategies/__init__.py:41
↓ 1 callers
Function
create_scheduler_adapter
( server_url: str, zmq_context: zmq.Context, vllm_config: VllmConfig, mq_timeout: float, h
lmcache/integration/vllm/lmcache_mp_connector_0201.py:111
↓ 1 callers
Function
create_scheduler_adapter
( server_url: str, zmq_context: zmq.Context, vllm_config: VllmConfig, mq_timeout: float, h
lmcache/integration/vllm/lmcache_mp_connector_0180.py:104
↓ 1 callers
Function
create_singleton_config
Create thread-safe singleton configuration access pattern. Args: getter_func_name: Name for the singleton getter function config_
lmcache/v1/config_base.py:488
↓ 1 callers
Function
create_store_policy
Create a store policy instance by name. Args: name: Registered policy name. Returns: A new StorePolicy instance. R
lmcache/v1/distributed/storage_controllers/store_policy.py:125
↓ 1 callers
Function
create_test_config
Create test configuration for LMCacheEngine.
tests/v1/utils.py:790
↓ 1 callers
Function
create_test_config_with_dual_plugins
Create config with two fs_connector instances.
tests/v1/storage_backend/test_fs_connector.py:47
↓ 1 callers
Function
create_test_config_with_plugin
Create a test configuration for HF3fsConnector using remote_storage_plugins. Args: base_path: Base path for storage mount_poi
tests/v1/storage_backend/test_hf3fs_connector.py:86
↓ 1 callers
Function
create_test_config_with_plugin
Create a test configuration for FSConnector using remote_storage_plugins.
tests/v1/storage_backend/test_fs_connector.py:33
↓ 1 callers
Function
create_test_layer_key
Create a LayerCacheEngineKey for testing layer-wise GDS operations.
tests/v1/storage_backend/test_gds_backend.py:65
↓ 1 callers
Function
create_test_memory_obj
Create a test MemoryObj for testing.
lmcache/v1/check/check_mode_test_remote.py:51
↓ 1 callers
Function
create_test_metadata
Create test metadata for LMCacheEngine.
tests/v1/utils.py:769
↓ 1 callers
Function
create_test_metadata
Create test metadata with configurable KV shape.
tests/v1/storage_backend/test_pd_backend_buffer_alignment.py:18
↓ 1 callers
Function
create_transfer_strategy
Create the non-GPU transfer strategy for a registered context. Args: storage_manager: Storage manager used by the selected strategy.
lmcache/v1/multiprocess/modules/server_transfer.py:37
↓ 1 callers
Function
create_trtllm_metadata
Construct LMCacheMetadata from TRT-LLM args and the KV pool tensor. Args: llm_args: TRT-LLM ``TorchLlmArgs``. kv_cache_tensor: TR
lmcache/integration/tensorrt_llm/utils.py:50
↓ 1 callers
Function
create_worker_adapter
( server_url: str, zmq_context: zmq.Context, vllm_config: VllmConfig, mq_timeout: float, h
lmcache/integration/vllm/lmcache_mp_connector_0201.py:144
↓ 1 callers
Function
create_worker_adapter
( server_url: str, zmq_context: zmq.Context, vllm_config: VllmConfig, mq_timeout: float, h
lmcache/integration/vllm/lmcache_mp_connector_0180.py:137
↓ 1 callers
Function
cudagraphArgs
cudagraphArgs returns the cudagraph-mode flag set for the given mode. "eager" (and any unrecognized value) maps to --enforce-eager; "full_decode_only"
operator/internal/webhook/cacheblend_inject_builders.go:174
↓ 1 callers
Method
current_device
(self)
tests/v1/test_musa_support.py:73
↓ 1 callers
Function
decode_chunk
Write the decode output in target_buffer Expected shape: [nlayers (kv in total), ntokens, nchannels]
lmcache/storage_backend/serde/cachegen_decoder.py:59
↓ 1 callers
Function
decode_header
Parse a header from msgpack bytes.
lmcache/v1/mp_observability/trace/format.py:125
↓ 1 callers
Function
decode_hfbucket_object_name
Decode a reversible bucket object name back into an LMCache key string.
lmcache/v1/storage_backend/connector/hfbucket_connector.py:828
↓ 1 callers
Function
decode_record
Parse a record from msgpack bytes.
lmcache/v1/mp_observability/trace/format.py:130
↓ 1 callers
Method
delete
Delete a batch of keys from the remote backend. Submits a batch delete to the native connector and blocks until the demux thread sign
lmcache/v1/distributed/l2_adapters/native_connector_l2_adapter.py:291
↓ 1 callers
Method
deleteCoordinatorMetricsServiceIfExists
deleteCoordinatorMetricsServiceIfExists removes the headless metrics Service if it exists and is owned by this coordinator (e.g. when prometheus is di
operator/internal/controller/coordinator_reconcile_helpers.go:207
↓ 1 callers
Function
deleteNamespace
deleteNamespace tears down a test namespace. We tolerate NotFound so repeated cleanup calls don't fail; the namespace's contents are GC'd asynchronous
operator/test/e2e/smoke_helpers_test.go:98
↓ 1 callers
Function
deleteNode
(nodeName)
lmcache/lmcache_frontend/static/js/app.js:392
↓ 1 callers
Method
deleteRESPAuthSecretIfExists
deleteRESPAuthSecretIfExists removes the managed RESP auth secret if it exists (e.g. when authSecretRef is removed from the spec).
operator/internal/controller/reconcile_helpers.go:493
↓ 1 callers
Method
deleteRESPAuthSecretIfExists
deleteRESPAuthSecretIfExists removes the managed RESP auth secret if it exists (e.g. when authSecretRef is removed from the spec).
operator/internal/controller/cacheblend_reconcile_helpers.go:428
↓ 1 callers
Method
delete_many
Delete indexed keys and recycle their slots when allowed. Args: encoded_keys: Ordered encoded raw-block keys to delete.
lmcache/v1/storage_backend/raw_block/core.py:836
↓ 1 callers
Method
deregister_workers
Deregister all workers after benchmark
lmcache/tools/controller_benchmark/benchmark.py:416
↓ 1 callers
Method
describe
Run all section builders and emit.
lmcache/cli/commands/describe.py:110
↓ 1 callers
Method
describe
(self)
lmcache/v1/cache_controller/message.py:61
↓ 1 callers
Method
describe
(self)
lmcache/v1/cache_controller/message.py:546
↓ 1 callers
Method
describe
(self)
lmcache/v1/cache_controller/message.py:772
↓ 1 callers
Method
deserialize
Deserialize ``src`` byte buffer into ``dst`` group. Args: src: Source byte-buffer MemoryObj. dst: Destination group w
lmcache/v1/distributed/serde/multi.py:163
↓ 1 callers
Method
deserialize
Read fp8 bytes from src, cast to dst's dtype, copy into dst.
lmcache/v1/distributed/serde/fp8.py:74
↓ 1 callers
Method
deserialize
Deserialize src byte buffer into dst KV-shaped MemoryObj (in-place). Args: src: Source MemoryObj containing serialized bytes.
lmcache/v1/distributed/serde/base.py:93
↓ 1 callers
Function
deserialize_header
Parse an EncodedKV header from a contiguous byte buffer. The buffer must contain header + payload. Returns an EncodedKV with `header_bytes`
lmcache/v1/kv_codec/encoded_kv.py:286
↓ 1 callers
Function
determine_per_gpu_memory
()
benchmarks/long_doc_qa/long_doc_qa_recommender.py:15
↓ 1 callers
Method
discover
( self, kv_caches: DiscoverableKVCache, layout_hints: LayoutHints )
lmcache/v1/gpu_connector/kv_format/detectors/vllm.py:25
↓ 1 callers
Method
dispatched_count
(self)
lmcache/v1/multiprocess/native_completion.py:98
↓ 1 callers
Function
do_a2
()
tests/v1/storage_backend/test_pd_backend_async.py:740
↓ 1 callers
Function
do_alloc
()
tests/v1/storage_backend/test_pd_backend_async.py:547
↓ 1 callers
Function
do_b
()
tests/v1/storage_backend/test_pd_backend_async.py:728
↓ 1 callers
Function
draw_quadratic_interpolation
()
benchmarks/ttft-estimator/ttft-estimator.py:100
↓ 1 callers
Method
drop_instance_state
Drop blend state for a reaped instance (InstanceLivenessTarget hook). Only the CB rope state is held per instance; the GPU cache context is
lmcache/v1/multiprocess/modules/blend_v3.py:523
↓ 1 callers
Method
dtype
(self)
tests/v1/platform/test_cache_context_dispatch.py:75
↓ 1 callers
Method
dtype
Return the tensor dtype for ``layer_idx``.
lmcache/v1/gpu_connector/kv_format/specs/base.py:160
↓ 1 callers
Function
dumb_metadata_with_model_name
(model_name: str, kv_shape=(32, 2, 256, 8, 128))
tests/v1/utils.py:178
↓ 1 callers
Function
dumpControllerLogs
dumpControllerLogs writes the last <tail> lines of every controller-manager pod's log to GinkgoWriter. Used by AfterEach failure handlers to make CI d
operator/test/e2e/smoke_helpers_test.go:110
↓ 1 callers
Function
dumpNamespace
dumpNamespace writes events, pods, and the LMCacheEngine CR yaml from the given namespace to GinkgoWriter. Designed to be cheap to call from AfterEach
operator/test/e2e/smoke_helpers_test.go:128
↓ 1 callers
Function
dump_parsed_config
Dump the parsed aggregated config as compact JSON. The parent process captures stdout line-by-line and logs each line separately, so we emit
examples/mp_runtime_plugins/mp_plugin.py:68
↓ 1 callers
Method
dynamic_load_file
Open an existing file, DMA read into L1 memory, then clean up.
lmcache/v1/distributed/l2_adapters/nixl_store_dynamic_l2_adapter.py:228
↓ 1 callers
Method
dynamic_store_file
Write-to-temp-then-rename to publish the final file atomically. The DMA write goes to ``<file_path>.tmp.<uuid>`` in the same director
lmcache/v1/distributed/l2_adapters/nixl_store_dynamic_l2_adapter.py:180
↓ 1 callers
Method
elements_per_layer
Return the element count per layer (both K and V for non-MLA). Non-MP only (see the class docstring).
lmcache/v1/gpu_connector/kv_format/specs/base.py:153
↓ 1 callers
Function
encode_ntokens
Encode a batch of ntokens. :param cdf_int: int16 tensor on GPU with shape [nlayers, nchannels, Lp] :param encode_input: int8 tensor on GPU wi
lmcache/storage_backend/serde/cachegen_encoder.py:253
↓ 1 callers
Function
encode_po_string
Convert a Python string into one or more quoted ``.po`` lines.
tools/translate_docs_zh.py:105
↓ 1 callers
Function
encode_record
Serialize a record to msgpack bytes.
lmcache/v1/mp_observability/trace/format.py:120
↓ 1 callers
Function
endpoint_url
Build the chat-completions URL from ``TRANSLATION_API_BASE_URL``.
tools/translate_docs_zh.py:283
↓ 1 callers
Method
enqueue_evict
Queue a best-effort eviction of fingerprints by storage key. Args: object_keys: ``object_key`` values to evict.
lmcache/v1/mp_coordinator/blend_client.py:151
↓ 1 callers
Function
ensure_connection
(host, port)
tests/conftest.py:635
↓ 1 callers
Method
entry_offset
Return the raw-block slot offset for a legacy key.
lmcache/v1/storage_backend/plugins/rust_raw_block_backend.py:225
↓ 1 callers
Method
estimate_serialized_size
Upper bound on bytes produced by :meth:`serialize`. Args: layout_descs: Group of optional layout descriptors with
lmcache/v1/distributed/serde/multi.py:128
↓ 1 callers
Method
estimate_serialized_size
Return the byte size to allocate for the serialized temp buffer. Called BEFORE serialization to allocate the temp buffer. The returne
lmcache/v1/distributed/serde/base.py:69
↓ 1 callers
Method
estimate_serialized_size
Delegate to the sync serializer's estimate (includes margin).
lmcache/v1/distributed/serde/async_processor.py:140
↓ 1 callers
Method
evict
Tombstone one fingerprint, compacting when tombstones dominate. Args: poly: The chunk's full 64-bit poly hash. Returns:
lmcache/v1/mp_coordinator/blend_directory.py:166
↓ 1 callers
Method
execute
Execute the subcommand. Called by the CLI dispatcher in ``main.py`` via ``args.func(args)`` after argument parsing. The :met
lmcache/cli/commands/base.py:66
↓ 1 callers
Method
execute
Parse CLI arguments into config objects and launch the HTTP server. Args: args: Parsed CLI arguments. Raises:
lmcache/cli/commands/server.py:70
↓ 1 callers
Method
execute_eviction_action
(self, action: EvictionAction)
lmcache/v1/distributed/storage_controllers/eviction_controller.py:175
↓ 1 callers
Method
exists
(self, key: str)
tests/conftest.py:282
↓ 1 callers
Method
exists
(self, key: CacheEngineKey)
lmcache/v1/storage_backend/connector/redis_connector.py:269
↓ 1 callers
Method
exists
(self, key: CacheEngineKey)
lmcache/v1/storage_backend/connector/redis_connector.py:647
↓ 1 callers
Method
exists
Check if a key exists in Valkey.
lmcache/v1/storage_backend/connector/valkey_connector.py:363
↓ 1 callers
Method
exists
Check if the remote server contains the key Input: key: a CacheEngineKey Returns: True if the cache
lmcache/v1/storage_backend/connector/base_connector.py:149
↓ 1 callers
Method
exists
Return whether a full LMCache chunk exists for ``key``.
lmcache/v1/storage_backend/connector/hfbucket_connector.py:256
↓ 1 callers
Method
exists
(self, key: CacheEngineKey)
lmcache/v1/storage_backend/connector/bigtable_connector.py:248
↓ 1 callers
Method
exists
Check if a key exists in remote storage. Acquires a lease, checks existence, then releases immediately.
lmcache/v1/storage_backend/connector/sagemaker_hyperpod_connector.py:643
↓ 1 callers
Method
exists
(self, key: CacheEngineKey)
lmcache/v1/storage_backend/connector/lm_connector.py:83
↓ 1 callers
Method
exists
(self, key: CacheEngineKey)
lmcache/v1/storage_backend/connector/eic_connector.py:283
↓ 1 callers
Method
exists_in_put_tasks
Check whether key is in the ongoing submit_put_task tasks. :param key: The key to check :return: True if the key exists in p
lmcache/v1/storage_backend/nixl_storage_backend.py:1861
↓ 1 callers
Method
exists_in_put_tasks
(self, key: CacheEngineKey)
lmcache/v1/storage_backend/remote_backend.py:208
↓ 1 callers
Method
exists_in_put_tasks
(self, key: CacheEngineKey)
lmcache/v1/storage_backend/local_disk_backend.py:87
↓ 1 callers
Method
exists_in_put_tasks
(self, key: CacheEngineKey)
lmcache/v1/storage_backend/local_disk_backend.py:210
↓ 1 callers
Method
exists_in_put_tasks
Check whether ``key`` is tracked as an in-flight put task. Args: key: The cache key to check. Returns: ``Tru
lmcache/v1/storage_backend/plugins/dax_backend.py:267
↓ 1 callers
Method
exists_inflight
Return whether a key currently has an in-flight write. Args: encoded_key: Encoded raw-block key string. Returns:
lmcache/v1/storage_backend/raw_block/core.py:493
↓ 1 callers
Method
exists_sync
(self, key: str)
lmcache/v1/storage_backend/resp_client.py:168
↓ 1 callers
Method
exists_sync
(self, key: CacheEngineKey)
lmcache/v1/storage_backend/connector/s3_connector.py:280
↓ 1 callers
Method
expand_prompt
Expand placeholders and return the complete prompt text. Args: prompt: The prompt template string with {name} placeholders.
lmcache/cli/commands/query/_prompt.py:28
↓ 1 callers
Method
expected_payload_len
(self)
lmcache/v1/kv_codec/encoded_kv.py:210
↓ 1 callers
Function
export_results
Export analysis results to a JSON file. Args: analysis: Analysis results dictionary output_file: Output file path include
examples/chunk_statistics/analyze_chunk_hashes.py:239
↓ 1 callers
Method
extra_cxx_flags_for
All common extensions share the same ABI flag under CUDA.
setup_extensions/build_profiles/cuda.py:76
↓ 1 callers
Function
extract_content
Extract content from the response chunk. Args: chunk: The response chunk from OpenAI Chat Completions API. Returns: str:
benchmarks/long_doc_qa/long_doc_qa.py:184
↓ 1 callers
Function
extract_content
Extract content from the response chunk. Args: chunk: The response chunk from OpenAI API. Returns: str: The content extra
benchmarks/multi_doc_qa/multi_doc_qa.py:76
↓ 1 callers
Function
extract_content_completions
Extract content from a Completions stream chunk.
benchmarks/long_doc_qa/long_doc_qa.py:177
↓ 1 callers
Function
fetch_nodes_from_supplier
Fetch node information from node supplier. For each discovered node, attempt to retrieve its child nodes via ``/api/nodes`` (multiProcess lmc
lmcache/lmcache_frontend/app.py:204
↓ 1 callers
Function
fetch_remote_config
Fetch configuration from remote config service. The config server protocol: - Request: POST with JSON body containing 'current_config' and 'e
lmcache/v1/config_base.py:740
← previous
next →
4,001–4,100 of 13,371, ranked by callers