SIMD Aligned Memory Vectorization and Super-block K-Quants
Parent: Local Inference Micro-Architectures: Signaling, Quantization Superblocks, and Kernel Pipeline Scheduling · Published reference · snapshot 2026-09-30
↓ Facts as markdownall context files
Source-anchored research for SIMD Aligned Memory Vectorization and Super-block K-Quants. Run status: COMPLETED.
These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.
Structure and components
- Super-block K-quants (Q4_K) group 256 weights into chunks with a shared FP16 super-block scale and eight 6-bit sub-block scales, reducing quantization noise in attention projection layers. (high confidence) [source]
Facts and statements
- 32-byte memory alignment prevents unaligned memory access penalties and cacheline splits during vectorized 256-bit SIMD load instructions across host CPU and GPU threads. (high confidence) [source]
Children
- No children recorded.