Mingxin Technology

Mingxin Technology

Aug 14

A Comparative Framework for KV Cache Performance on Domestic Accelerators: Measured Insights

#kvcache #ai #storage #llm

5 min read

Mingxin Technology

Aug 14

KV Cache Storage Product Comparison: Methodology and Measured Benchmarks

#kvcache #ai #storage #llm

5 min read

Mingxin Technology

Aug 14

KV Cache Memory vs. Hit Rate: Why Larger Caches Yield Diminishing Returns

#kvcache #ai #storage #llm

4 min read

Mingxin Technology

Aug 14

The Boundary Between External Memory and VRAM for KV Cache in Local LLM Deployment

#kvcache #ai #storage #llm

5 min read

Mingxin Technology

Aug 14

Optimizing Compute Rental Costs: Dynamic Scaling and On-Demand Allocation Strategies

#ai #storage #llm #performance

5 min read

Mingxin Technology

Aug 14

Measured Application Analysis of Domestic KV Cache Products in AI Inference

#kvcache #ai #storage #llm

5 min read

Mingxin Technology

Aug 11

Three Clause Types Often Missed in Compute Rental Contracts

#ai #storage #llm #performance

5 min read

Mingxin Technology

Aug 11

Energy Consumption Assessment and Optimization Paths for Data Center-Scale KV Cache Deployment

#kvcache #ai #storage #llm

7 min read

Mingxin Technology

Aug 11

Choosing Compute Rental: Focus on Three SLA Metrics, Not Unit Price

#sla #ai #storage #llm

5 min read

Mingxin Technology

Aug 11

Bandwidth and Latency Optimization Strategies for KV Cache in Edge Computing

#kvcache #ai #storage #llm

5 min read

Mingxin Technology

Aug 11

Four Key Dimensions for Evaluating the Reliability of Domestic AI Storage

#ai #storage #llm #performance

6 min read

Mingxin Technology

Aug 11

Three-Layer Adaptation of the Ascend Inference Stack: Driver, Operator, and Framework Are All Indispensable

#ai #storage #llm #performance

6 min read

Mingxin Technology

Aug 7

Can Compressed Sensing Be Used for Inference Storage Data Compression?

#ai #storage #llm #performance

4 min read

Mingxin Technology

Aug 7

NVMe-oF vs. RDMA: Performance Comparison in Inference Storage

#nvmeof #rdma #ai #storage

5 min read

Mingxin Technology

Aug 7

Storage Selection Strategy for Inference in Domestic AI Computing Centers

#ai #storage #llm #performance

5 min read

Mingxin Technology

Aug 7

How KV Cache Prefetch Cuts Storage Latency

#kvcache #ai #storage #llm

6 min read

Mingxin Technology

Aug 5

KV Cache Reuse in Multi-Turn Dialogue: 29% Throughput Gain Measured

#kvcache #ai #storage #llm

5 min read

Mingxin Technology

Aug 5

Evaluating Long-Context Inference Performance of Domestic Accelerator Cards

#ai #storage #llm #performance

4 min read

Mingxin Technology

Aug 5

Storage Bottlenecks in Real-Time Video Inference and a Tiered Acceleration Approach

#ai #storage #llm #performance

5 min read

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

Badges

Writing Debut

A Comparative Framework for KV Cache Performance on Domestic Accelerators: Measured Insights

KV Cache Storage Product Comparison: Methodology and Measured Benchmarks

KV Cache Memory vs. Hit Rate: Why Larger Caches Yield Diminishing Returns

The Boundary Between External Memory and VRAM for KV Cache in Local LLM Deployment

Optimizing Compute Rental Costs: Dynamic Scaling and On-Demand Allocation Strategies

Measured Application Analysis of Domestic KV Cache Products in AI Inference

Three Clause Types Often Missed in Compute Rental Contracts

Energy Consumption Assessment and Optimization Paths for Data Center-Scale KV Cache Deployment

Choosing Compute Rental: Focus on Three SLA Metrics, Not Unit Price

Bandwidth and Latency Optimization Strategies for KV Cache in Edge Computing

Four Key Dimensions for Evaluating the Reliability of Domestic AI Storage

Three-Layer Adaptation of the Ascend Inference Stack: Driver, Operator, and Framework Are All Indispensable

Can Compressed Sensing Be Used for Inference Storage Data Compression?

NVMe-oF vs. RDMA: Performance Comparison in Inference Storage

Storage Selection Strategy for Inference in Domestic AI Computing Centers

How KV Cache Prefetch Cuts Storage Latency

KV Cache Reuse in Multi-Turn Dialogue: 29% Throughput Gain Measured

Evaluating Long-Context Inference Performance of Domestic Accelerator Cards

Storage Bottlenecks in Real-Time Video Inference and a Tiered Acceleration Approach

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