Research

Research gives BitCal a context layer that is useful to engineers, not just decorative to readers. It collects the external references, related systems, and evolution notes that inform the redesign.

Research map

This section answers three kinds of questions:

• which papers, manuals, and references underpin the performance and algorithm narrative;
• which nearby libraries or systems are the most relevant comparison points;
• which trade-offs BitCal is making by choosing a C++23, x86-64-first, owner/view/algorithm model.

Reference set

1. Intel Intrinsics Guide

   Instruction reference

   Used when discussing x86 SIMD capabilities and instruction-level constraints.

2. Agner Fog optimization manuals

   Optimization manuals

   Relevant for dispatch costs, latency, throughput, and measurement interpretation on x86-64.

3. Sebastiano Vigna, Broadword Implementation of Rank/Select Queries

   Broadword paper

   Important background for word-parallel bit techniques and for understanding why algorithm shape matters.

Related systems

Use Evolution and Comparisons for the detailed comparison layer. The short list of meaningful reference points is:

• Boost dynamic_bitset for a C++-native dynamic bitset baseline;
• CRoaring for performance-conscious bitmap engineering with a different domain focus;
• libpopcnt for a focused example of how serious one low-level bit primitive can become.

How this section fits the reading chain

Research should usually be read after Reference, because external comparison only makes sense once BitCal's own contract is clear. End with Status to bring the scope back down to what the repository actually supports today.