class Types The class itself is the library's introspection: its version, and the hardware it can and does use. Everything else it holds is a nested type.
long entries = index.count(Key.of(42));
boolean present = index.contains(Key.of(42));
DenseIndexDescription description = index.describe();
DenseReachability reachability = index.health();
System.out.println(reachability.reachable() + " of " + reachability.audited() + " entries reachable");
System.out.println(Types.version() + " built for " + Types.capabilitiesCompiled() + ", here " + Types.capabilitiesAvailable());describe reports the frozen DenseMetadata with the size, capacity, memory usage and dispatched capabilities in one call. health audits how much of the graph a search reaches from its landmarks; entries it misses stay unfound until a refine re-links them. Types.capabilitiesCompiled lists the hardware capabilities the linked binary was compiled with, and Types.capabilitiesAvailable those the current machine offers.
TypesThe vocabulary every modality shares: entry identities, failure reporting, scalar kinds, the executor every fanned-out verb runs on, deferred calls, and the strided descriptors batches travel in.2class Types The class itself is the library's introspection: its version, and the hardware it can and does use. Everything else it holds is a nested type.
CPUClassA CPU family USearch ships tuned variants for, which a RAM or NVMe placement or executor may pin.enum CPUClass SERIAL | | The portable floor, with no SIMD assumptions. |
|---|---|---|
INTEL_HASWELL | | Haswell-class x86: AVX2 and FMA. |
INTEL_SIERRA | | Sierra Forest-class x86: AVX-VNNI without AVX-512. |
INTEL_SKYLAKE | | Skylake-class x86: the AVX-512 foundation. |
INTEL_ICELAKE | | Icelake-class x86: AVX-512 VNNI and VPOPCNT. |
AMD_GENOA | | Genoa-class x86: native bf16 dot products over AVX-512. |
INTEL_SAPPHIRE | | Sapphire Rapids-class x86: AVX-512 FP16 and the AMX tiles. |
INTEL_GRANITE | | Granite Rapids-class x86: the AMX-FP16 tiles. |
INTEL_DIAMOND | | Diamond Rapids-class x86: AVX10.2's FP8 conversions. |
AWS_G2 | | Graviton2-class Arm: the bare NEON floor. |
APPLE_M2 | | Armv8.6 NEON with dot products and bf16. |
APPLE_M4 | | Apple M4-class Arm: the SME outer-product tiles. |
AWS_G5 | | Graviton5-class Arm: Armv9.2. |
NVIDIA_VERA | | Vera-class Arm: FP8 dot products. |
ALIBABA_C930 | | Alibaba C930-class RISC-V: RVA23 vectors. |
WASM32_V128 | | WebAssembly with fixed-width simd128. |
WASM64_EXTENDED | | WebAssembly with relaxed SIMD, threads and 64-bit addressing. |
MONOVERSIONED | | The one class of a monoversioned build. |
GPUClassA GPU family USearch ships tuned variants for, which a VRAM placement or executor may pin.enum GPUClass NVIDIA_AMPERE | | Ampere-class NVIDIA, and every NVIDIA GPU without a class of its own. |
|---|---|---|
NVIDIA_ADA | | Ada-class NVIDIA: the fp8 conversions. |
NVIDIA_HOPPER | | Hopper-class NVIDIA. |
NVIDIA_BLACKWELL | | Blackwell-class NVIDIA. |
NVIDIA_BLACKWELLRTX | | Consumer Blackwell NVIDIA. |
NVIDIA_RUBIN | | Rubin-class NVIDIA. |
AMD_CDNA3 | | CDNA3 AMD: gfx942. |
AMD_CDNA4 | | CDNA4 AMD: gfx950. |
AMD_CDNA5 | | CDNA5 AMD: gfx1250. |
APPLE_METAL7 | | Apple7-family GPUs, and every Apple GPU without a class of its own. |
APPLE_METAL9 | | Apple9-family GPUs. |
APPLE_METAL10 | | Apple10-family GPUs. |
versionThe version of the linked library.capabilitiesCompiledThe hardware capabilities the linked binary was compiled with, comma separated.capabilitiesAvailableThe hardware capabilities the current machine offers at runtime, comma separated.