[CPU] Refactor CPU affinity and memory management (#39781)

Signed-off-by: jiang1.li <jiang1.li@intel.com>
This commit is contained in:
Li, Jiang
2026-04-17 21:01:08 +08:00
committed by GitHub
parent b1dc87a098
commit d02421a7db
14 changed files with 721 additions and 423 deletions
-3
View File
@@ -101,8 +101,6 @@ class CacheConfig:
kv_cache_dtype_skip_layers: list[str] = field(default_factory=list)
"""Layer patterns to skip KV cache quantization. Accepts layer indices
(e.g., '0', '2', '4') or attention type names (e.g., 'sliding_window')."""
cpu_kvcache_space_bytes: int | None = None
"""(CPU backend only) CPU key-value cache space."""
mamba_page_size_padded: int | None = None
""" Optional override for mamba page size; used by hybrid mamba/attention
models to ensure exact alignment with attention page size."""
@@ -183,7 +181,6 @@ class CacheConfig:
"num_gpu_blocks_override",
"enable_prefix_caching",
"prefix_caching_hash_algo",
"cpu_kvcache_space_bytes",
"mamba_page_size_padded",
"user_specified_block_size",
"user_specified_mamba_block_size",