sglang

allocator

[Kernel] Introduce sglang.kernels namespace and migrate scattered triton_ops kernels (RFC #29630 , Phase 2) (#30044 )

2026-07-10 21:41:08 +08:00

cpp_radix_tree

Fix syntax errors in cpp_radix_tree (#13315 )

2025-11-15 11:35:43 +08:00

cpp_utils

[HiCache] Optimize HiCache hash generation with bulk token byte conversion (#28287 )

2026-07-01 15:44:23 +08:00

hybrid_cache

[mem_cache][7/N] refactor: move MLATokenToKVPoolHost to pool_host.mla (#30616 )

2026-07-13 14:23:38 +08:00

layout

feat(mem_cache): page-major (layer-major within a page) KV/state layout (#29533 )

2026-06-29 14:49:54 -07:00

pool_host

[mem_cache][7/N] refactor: move MLATokenToKVPoolHost to pool_host.mla (#30616 )

2026-07-13 14:23:38 +08:00

sparsity

hisparse: support NIXL DRAM KV destinations for HiSparse (#27563 )

2026-06-27 22:32:29 +08:00

storage

Let cache backend do not couple with owned committed kv details and avoid kv_committed_freed/kv_overallocated_freed fields (#29428 )

2026-07-15 14:43:10 +08:00

unified_cache_components

Let cache backend do not couple with owned committed kv details and avoid kv_committed_freed/kv_overallocated_freed fields (#29428 )

2026-07-15 14:43:10 +08:00

allocation_sizing.py

Lightweight extract allocation logic from mem_cache/common.py to more clearly show nearly parallel variants (#29431 )

2026-07-15 14:49:37 +08:00

allocation.py

Fix bookkeeping fields not encapsulated with real allocations in normal alloc, PD pre-alloc, DFlash and EAGLE (#29432 )

2026-07-15 14:52:21 +08:00

base_prefix_cache.py

[AMD] Enable unified-KV HiCache on DeepSeek-V4 (#29417 )

2026-07-09 12:58:47 -07:00

base_swa_memory_pool.py

Unify full→SWA index translation in init_forward_metadata; drop pool caches (#27091 )

2026-06-03 16:12:27 -07:00

cache_init_params.py

feat: first-class session identity in SGLang (#29436 )

2026-06-28 07:16:38 -07:00

chunk_cache.py

Let cache backend do not couple with owned committed kv details and avoid kv_committed_freed/kv_overallocated_freed fields (#29428 )

2026-07-15 14:43:10 +08:00

common.py

Fix bookkeeping fields not encapsulated with real allocations in normal alloc, PD pre-alloc, DFlash and EAGLE (#29432 )

2026-07-15 14:52:21 +08:00

deepseek_v4_compress_state.py

[AMD] Fix DeepSeek V4 MTP accuracy issue (#30333 )

2026-07-06 23:57:57 -07:00

deepseek_v4_memory_pool.py

[Kernel] Migrate DSA + DSV4 attention kernels to sglang.kernels (RFC #29630 , Phase 2.5, 5/7) (#30792 )

2026-07-15 11:11:22 +08:00

dsa_cache_layer_split.py

[Kernel] Migrate DSA + DSV4 attention kernels to sglang.kernels (RFC #29630 , Phase 2.5, 5/7) (#30792 )

2026-07-15 11:11:22 +08:00

events.py

fix: batch BlockRemoved events per radix node (#29265 )

2026-06-26 16:13:40 +08:00

evict_policy.py

[lint] Enable Ruff UP037 to drop redundant quoted annotations (#27984 )

2026-06-11 17:38:05 -07:00

flush_cache.py

…

hi_mamba_radix_cache.py

[refactor] Retire the legacy config accessor and the remaining process singletons (#30493 )

2026-07-09 02:10:47 -07:00

hicache_storage.py

feat(mem_cache): add client-side metadata cache for HiCacheFile storage (#29716 )

2026-07-07 18:21:46 -07:00

hiradix_cache.py

[mem_cache][7/N] refactor: move MLATokenToKVPoolHost to pool_host.mla (#30616 )

2026-07-13 14:23:38 +08:00

hisparse_memory_pool.py

[mem_cache][3/N] refactor: move HiSparse allocators to allocator/hisparse.py (#26678 )

2026-06-11 19:50:24 +08:00

kv_cache_builder.py

Extract hybrid-arch helpers into configs.hybrid_arch and ModelConfig (#31150 )

2026-07-14 15:55:32 +08:00

kv_cache_configurator.py

Lightweight extract allocation logic from mem_cache/common.py to more clearly show nearly parallel variants (#29431 )

2026-07-15 14:49:37 +08:00

kv_cache_dtype.py

Extract kv cache dtype configuration into mem_cache (#31147 )

2026-07-14 15:52:39 +08:00

kv_vmm_backing.py

Make KvVmmArena JIT stub unique per process (#30702 )

2026-07-09 15:56:26 -07:00

mamba_checkpoint_pool.py

[refactor] Retire the legacy config accessor and the remaining process singletons (#30493 )

2026-07-09 02:10:47 -07:00

mamba_radix_cache.py

Let cache backend do not couple with owned committed kv details and avoid kv_committed_freed/kv_overallocated_freed fields (#29428 )

2026-07-15 14:43:10 +08:00

memory_pool_host.py

[HiCache & HybridModel] nixl hicache backend support hybrid models (#29191 )

2026-07-13 09:09:37 -07:00

memory_pool.py

[Kernel] Migrate DSA + DSV4 attention kernels to sglang.kernels (RFC #29630 , Phase 2.5, 5/7) (#30792 )

2026-07-15 11:11:22 +08:00

mmap_allocator.py

Refactor NIXL hicache. Add O_DIRECT support (#25173 )

2026-06-01 17:28:53 +02:00

multi_ended_allocator.py

[Kernel] Introduce sglang.kernels namespace and migrate scattered triton_ops kernels (RFC #29630 , Phase 2) (#30044 )

2026-07-10 21:41:08 +08:00

multimodal_cache.py

[VLM] Chunk-aware ViT encoding with per-image cache and lazy device transfer (#22038 )

2026-04-04 16:55:17 +08:00

pure_swa_radix_cache.py

Let cache backend do not couple with owned committed kv details and avoid kv_committed_freed/kv_overallocated_freed fields (#29428 )

2026-07-15 14:43:10 +08:00

radix_cache_cpp.py

Let cache backend do not couple with owned committed kv details and avoid kv_committed_freed/kv_overallocated_freed fields (#29428 )

2026-07-15 14:43:10 +08:00

radix_cache.py

Let cache backend do not couple with owned committed kv details and avoid kv_committed_freed/kv_overallocated_freed fields (#29428 )

2026-07-15 14:43:10 +08:00

registry.py

Using UnifiedRadixTree by default for SWA, Mamba, and DSA models (#30468 )

2026-07-14 15:17:08 +08:00

session_radix_cache.py

feat: first-class session identity in SGLang (#29436 )

2026-06-28 07:16:38 -07:00

swa_memory_pool.py

Size KV pool after CUDA graph capture (opt-in) (#30157 )

2026-07-07 12:05:01 -07:00

swa_radix_cache.py

Let cache backend do not couple with owned committed kv details and avoid kv_committed_freed/kv_overallocated_freed fields (#29428 )

2026-07-15 14:43:10 +08:00

unified_memory_pool.py

[Kernel] Introduce sglang.kernels namespace and migrate scattered triton_ops kernels (RFC #29630 , Phase 2) (#30044 )

2026-07-10 21:41:08 +08:00

unified_radix_cache.py

Let cache backend do not couple with owned committed kv details and avoid kv_committed_freed/kv_overallocated_freed fields (#29428 )

2026-07-15 14:43:10 +08:00

utils.py

[Kernel] Introduce sglang.kernels namespace and migrate scattered triton_ops kernels (RFC #29630 , Phase 2) (#30044 )

2026-07-10 21:41:08 +08:00