Crate spark_storage

Crate spark_storage 

Source

Re-exports§

pub use cuda_module::CudaEvent;
pub use cuda_module::CudaModule;
pub use cuda_module::launch_kernel;
pub use model_dims::ModelDims;
pub use backend::PosixBackend;
pub use backend::ReadRequest;
pub use backend::StorageBackend;
pub use backend::IoUringBackend;
pub use config::HighSpeedSwapConfig;
pub use eviction::EvictionPolicy;
pub use expert::ExpertKey;
pub use expert::ExpertLayout;
pub use expert::ExpertRecordHeader;
pub use expert::ExpertRecordId;
pub use expert::ExpertRecordSpec;
pub use expert::Proj;
pub use expert::ProjBytes;
pub use expert_arena::ExpertArena;
pub use expert_pack::ExpertFileReader;
pub use expert_pack::ExpertFileWriter;
pub use expert_pack::ExpertIndex;
pub use expert_pack::ProjData;
pub use expert_pack::ProjView;
pub use expert_pack::pack_record;
pub use expert_pack::unpack_record;
pub use expert_tier::ArenaSlot;
pub use expert_tier::ExpertResidency;
pub use expert_tier::ExpertTier;
pub use expert_tier::PosixTier;
pub use expert_tier::TierKind;
pub use expert_tier::UmaArenaTier;
pub use expert_tier::open_tier;
pub use expert_tier_rdma::RdmaTier;
pub use high_speed_swap::HighSpeedSwap;
pub use high_speed_swap::install_local;
pub use high_speed_swap::local_installed;
pub use high_speed_swap::with_local;
pub use ngram_cache::NgramRowCache;
pub use rdma_snapshot::RdmaSnapshotArena;
pub use predictor::Predictor;
pub use predictor::PredictorDims;
pub use probe::Backend;
pub use probe::ProbeConfig;
pub use probe::ProbeResult;
pub use probe::run_probe;
pub use projection::PredictorShape;
pub use projection::build_projection;
pub use tiled_attention::TiledAttention;
pub use tiled_attention::TiledAttentionDims;
pub use weight_lora_rdma::RdmaLoraLoader;
pub use weight_lora_rdma::LoraAbKind;
pub use weight_lora_rdma::LoraLandTarget;
pub use weight_peer::WeightManifest;
pub use weight_peer::WeightTensorRecord;
pub use weight_tier_rdma::RdmaWeightLoader;

Modules§

attention_ref
backend
bench
cache_peer
cascade_backend
cascade_policy
config
cuda_graph
cuda_min
cuda_module
eviction
expert
expert_arena
expert_pack
expert_peer
expert_tier
expert_tier_rdma
group
high_speed_swap
kv_paging
KV as a first-class paging kind (part of the tiered-cache consolidation, DEFAULT-OFF behind ATLAS_KV_PAGING).
layout
model_dims
ModelDims — model-shape descriptor threaded through every layer’s forward signature. Lives in spark-storage because the high-speed- swap orchestrator was the original consumer, but the type itself has no GPU state and must remain compilable on macOS / no-cuda builds where the swap orchestrator isn’t reachable.
ngram_cache
NVMe-backed row cache for the n-gram embedding tables.
predictor
predictor_ref
probe
projection
rdma_snapshot
Offset-addressed RDMA arena for the SSM-snapshot spill tier.
scratch_pool
snapshot_swap
tiled_attention
weight_lora_rdma
weight_peer
weight_tier_rdma

Functions§

rdma_verbs_enabled
true iff atlas_rdma_verbs was re-emitted for this crate by build.rs (the one-sided verbs shim lives in the CUDA-free atlas-rdma crate; rustc-cfg doesn’t cross crates, so build.rs re-emits it off atlas-rdma’s links metadata). rdma_verbs_probe_tests asserts it, so a silent cfg evaporation fails cargo test -p spark-storage --lib on verbs hosts instead of green-building with the gated modules compiled out.