Mingxin Technology

Storage acceleration for LLM inference. KV-cache tiering on NVMe-oF: +29-40% throughput (signed benchmarks, reproducible). mingxinstorage.xyz

Joined on Jul 22, 2026

Mingxin Technology

Aug 4

KV Cache Reuse in Multi-Turn Dialogue: A Deployment Case Study

#kvcache #ai #storage #llm

5 min read

Mingxin Technology

Aug 4

How KV Cache Pooling and Sharing Improves Inference Resource Utilization

#kvcache #ai #storage #llm

5 min read

Mingxin Technology

Aug 4

Distributed KV Cache Load Balancing: Strategies and Measured Trade-offs

#kvcache #ai #storage #llm

6 min read

Mingxin Technology

Aug 4

Measured Evaluation of KV Cache Acceleration for Inference in Online Education Scenarios

#kvcache #ai #storage #llm

4 min read

Mingxin Technology

Aug 3

Deployment Path and Measured Performance of Domestic AI Inference Acceleration Cards in Real-Time Database Queries

#ai #storage #llm #performance

5 min read

Mingxin Technology

Aug 3

Case Analysis of Huawei OceanStor UCM Inference Acceleration Solution

#oceanstor #ucm #ai #storage

5 min read

Mingxin Technology

Jul 31

Why KV Cache Tiering Is the Key TCO Optimization Point in a Three-Tier Storage Architecture for Compute Centers

#tco #ai #storage #llm

6 min read

Mingxin Technology

Jul 31

The Hidden Cost of Model Switching: A Measured Path from 46.7% to 62.8% Effective Compute Utilization

#gpu #ai #storage #llm

5 min read

Mingxin Technology

Jul 28

Cross-Platform Methodology Porting: Validating KV Cache Memory Efficiency from Muxi N260 to MI308X

#rocm #gpu #ai #storage

3 min read

Mingxin Technology

Jul 28

Inference Deployment in Xinchuang Scenarios: Architecture Essentials for Data Not Leaving the Domain

#rocm #gpu #ai #storage

5 min read

Mingxin Technology

Jul 25

Clos Network Architecture: A Cost-Effectiveness and Selection Framework for Thousand-Card Inference Clusters

#tco #ai #storage #llm

7 min read

Mingxin Technology

Jul 24

How to Determine the Storage-to-Compute Ratio for Inference Clusters: Measured Basis for One Array Serving 8 Nodes

#kvcache #lmcache #vllm #ai

6 min read

Mingxin Technology

Jul 24

Engineering Practice Essentials of NVMe-oF + RoCEv2 in Inference Storage Scenarios: A Case Study with FX100 Measurements

#kvcache #lmcache #vllm #ai

5 min read

Mingxin Technology

Jul 22

What 90% Line-Rate Utilization on a Single 100GbE Port Means: Analyzing Network Bottlenecks in Inference Storage

#kvcache #lmcache #vllm #ai

5 min read

DEV Community

Mingxin Technology

Badges

Writing Debut

KV Cache Reuse in Multi-Turn Dialogue: A Deployment Case Study

How KV Cache Pooling and Sharing Improves Inference Resource Utilization

Distributed KV Cache Load Balancing: Strategies and Measured Trade-offs

Measured Evaluation of KV Cache Acceleration for Inference in Online Education Scenarios

Deployment Path and Measured Performance of Domestic AI Inference Acceleration Cards in Real-Time Database Queries

Case Analysis of Huawei OceanStor UCM Inference Acceleration Solution

Why KV Cache Tiering Is the Key TCO Optimization Point in a Three-Tier Storage Architecture for Compute Centers

The Hidden Cost of Model Switching: A Measured Path from 46.7% to 62.8% Effective Compute Utilization

Cross-Platform Methodology Porting: Validating KV Cache Memory Efficiency from Muxi N260 to MI308X

Inference Deployment in Xinchuang Scenarios: Architecture Essentials for Data Not Leaving the Domain

Clos Network Architecture: A Cost-Effectiveness and Selection Framework for Thousand-Card Inference Clusters

How to Determine the Storage-to-Compute Ratio for Inference Clusters: Measured Basis for One Array Serving 8 Nodes

Engineering Practice Essentials of NVMe-oF + RoCEv2 in Inference Storage Scenarios: A Case Study with FX100 Measurements

What 90% Line-Rate Utilization on a Single 100GbE Port Means: Analyzing Network Bottlenecks in Inference Storage