[Under Review] Revisiting BaM's '× Cost Advantage' — The Denominator That Got Skipped
BaM claims 4.4×~21.8× cost advantage over DRAM-only systems. But this conclusion rests on $/GB — while the real bottleneck for these workloa…
Hand-written deep dives from first principles — on GPU dataflow, NVLink topology, CXL memory, KV-cache economics, and the architecture decisions that shape AI infrastructure.
BaM claims 4.4×~21.8× cost advantage over DRAM-only systems. But this conclusion rests on $/GB — while the real bottleneck for these workloa…
CXL/RDMA distributed memory pools upgrade memory from compute appendage to poolable network resource, but existing data analytics tooling (p…
Extend PyTorch Dataset/DataLoader API to build CXL memory pool-aware distributed data loading system — not a transparent solution, but expli…
The critical missing piece in the CXL-for-training stack is not hardware capability, but a fundamental rethinking of PyTorch's DataLoader ar…
GPU-Driven RDMA is not just 'GPU initiates RDMA' — it is the tight coupling of memory and communication in a multi-execution-unit system (GP…
OpenURMA is an open-source FPGA implementation of the UB (Ultra-Bus) connectionless RDMA transport, achieving ≈500ns remote load/store laten…
Start here for the most in-depth analysis — chip architecture, inference systems, and communication libraries.
NVIDIA GPU / CPU / 网络全产品线规格总览 · 含 8 篇 Agent 深度调研模块 (GPU 微架构 / 互联 / 内存 / 网络 / DGX / 边缘计算 / 学术论文 / 数据手册)
Read the analysis →AMD Advancing AI 2026 全栈深度调研 · EPYC Venice (Zen 6/6c, 256 核) / MI455X (CDNA 4, 2.61 PFLOPS FP8) / Helios 机架 / Ryzen AI Gorgon · 芯片 + 集群 + 路线图
Read the analysis →3 份主报告 + 16 篇 CAM Agent 深挖 + 10 篇 URPC 专题文档 · CAM 通信加速库 v2 (8章) / URPC 统一远程过程调用 (11篇) / CAM v1 · 21K+ 分析行数
Read the analysis →12 章统一分析 + 3 专题 (第一性原理 / 热路径 / KV-Cache 4 层 / 分布式 / Ascend Overlay / Perf Handbook) · 百万序列 Prefill 专题 12 章 · KV Cache 即 Buffer / 1M token = 1.15 GB / NIXL vs Mooncake
Read the analysis →vLLM / SGLang / LMCache / Mooncake / Dynamo 五大框架的 Prefill-Decode 分离 + KV Cache 路由内部实现源码级拆解
Read the analysis →Keynote + 19 篇论文逐篇深度解读后的跨论文战略综合 · 6 条主轴: 同步税 / 存储层级重定义 / P2P 转移 / Superchip 冲击 / 批判性转向 / 训练路线分叉
Read the analysis →AI-generated deep-dive analysis — paper-level research with full source tracing and cross-referencing.
© 2026 backyes · Human-driven, AI-amplified · Built with Claude Code