Monthly Summary for 2025-10: OpenXiangShan-Nanhu project maintenance focused on dependency alignment and observability improvements across LinkNan and Nanhu-V5. Key features delivered include submodule bumps for Nanhu across multiple commits to align with latest state, and memory stress testing configuration options for targeted performance analysis. In Nanhu-V5, readiness update of the ready-to-run submodule ensured build consistency. An HWMon enhancement added deeper committed instruction telemetry, improving observability and debugging for performance and security contexts. No major bug fixes were reported this month; instead, the work emphasizes build hygiene, reproducibility, and enhanced runtime insight.

Monthly summary for 2025-09 focusing on feature delivery, stability improvements, and cross-project collaboration for OpenXiangShan-Nanhu/LinkNan. Key outcomes include CI release workflow enhancements with multiple ST single-core build variants, NEMU integration in the ST release, and dependency management with nanhu to ensure compatibility. Additionally, Device Tree ISA extension strings were added to the dtsi to define processor capabilities and improve software compatibility across platforms.

August 2025 monthly summary focusing on release engineering, platform coverage, and ecosystem alignment across Nanhu-V5 and LinkNan. Key changes include multi-target FPGA/ST release support, separation of EDA/FPGA release workflows, Nanhu submodule update, CSR mapping bug fix in Nanhu-V5, and CI/packaging hardening. These efforts improved release reliability, validated multiple configurations, and improved artifact management, enabling faster, safer releases and broader platform support.

Concise monthly summary for 2025-07 highlighting delivered features, reliability improvements, and cross-repo alignment that directly support faster, more reliable releases and lower maintenance costs.

2025-06 Monthly Summary — Delivered core cohesion and efficiency improvements across Nanhu-V5 and LinkNan, positioning the project for upcoming Zhujiang deployments. Key work focused on submodule synchronization, stability fixes, and memory footprint reductions that improve reliability in multi-core and resource-constrained scenarios.

May 2025 monthly summary for OpenXiangShan-Nanhu/Nanhu-V5 focused on dependency hygiene, environment consistency, and enhanced validation through PLDM-enabled Difftest. Delivered two key feature streams: (1) dependency updates for the ready-to-run submodule and nemu subproject to align versions and ensure a stable environment; included commits bumping ready-to-run nemu (63c737a304f73fcbd53bab0199a8a36fcc20bd4a) and bump nemu: turn off v extension (e0b44060b3f65f54104dde8e6327270a6cdd5e8b). (2) PLDM-enabled Difftest enhancements to improve simulation accuracy, including new registers (DTS, graphML, JSON, plusArgs), adjusted build configurations, refreshed difftest submodule references, and refined writeback data handling for vector loads; commit feat: add support on difftest pldm (1cfa6b65cc29c84c20ea5cbe5c5c96e75b102bda).

April 2025 performance summary for Nanhu-V5: Delivered observable improvements in hardware monitoring, configuration, and NoC tooling, while stabilizing the platform through targeted reliability fixes. Key work includes hardware monitoring integration with ROB/CSR, fast-simulation support via NactMiniConfig, and introduction of NoC configuration capabilities, complemented by focused refactors and robustness fixes that reduce risk in memory paths and improve build stability. These efforts enhance observability, accelerate validation cycles, and enable safer performance decisions across the stack.

February 2025 — Nanhu-V5 development highlights focused on verification tooling hardening, FP arithmetic correctness, and FP16 support through submodule integration. Key outcomes include enhancements to the difftest workflow, robust initialization fixes, and expanded FP capabilities to support AI workloads.

January 2025 (2025-01) performance summary for OpenXiangShan-Nanhu/Nanhu-V5. The focus this month was expanding 64-bit vector capabilities, improving decoding paths, and strengthening the modular, shareable vector infrastructure, while addressing critical masking and wiring issues to stabilize signal propagation. The efforts reduced development risk for future vector features by standardizing naming and refactoring for clearer maintenance and reuse across the VPU stack.

Month: 2024-11 — In November 2024, I focused on stabilizing the memory subsystem in Nanhu-V5 and improving repository health to enable faster future iterations. Key deliveries include a bug fix for SRAM read data handling in ICache and broad maintenance to update dependencies and MemBlock imports across submodules, preparing for upcoming enhancements. These efforts reduce runtime risk in the ICache path and improve build reliability and integration readiness.