Data Parallelism

Analysis of TPU Internal Architecture: From Chip Design to Large-Scale Clusters

by
Analysis of TPU Internal Architecture: From Chip Design to Large-Scale Clusters

Abstract This article delves into the hardware architecture and design philosophy of Google’s Tensor Processing Unit (TPU). It covers the components of a single TPU chip (including matrix multiplication units, vector processing units, and memory hierarchy) to the interconnection topology of multiple chips (such as 2D/3D toroidal connections and optical switching technology). The article discusses … Read more

Differences and Similarities Between CPU and NPU

by
Differences and Similarities Between CPU and NPU

With the rapid development of NPU, is it still necessary to study the architecture related to CPU? In fact, this is a matter of passing down knowledge, and I personally believe it is still very necessary. Through a comparison of CPU and NPU, we can see that NPU can be regarded as an extension and … Read more

Introduction to Multithreading in C++11 (Part 1)

by
Introduction to Multithreading in C++11 (Part 1)

Introduction This series is written because I found my programming skills to be lacking, so I wanted to create a blog to help myself review. I hope it can help everyone (and not mislead too much). If you have any questions, feel free to point them out in the comments. The environment for this article … Read more

C++ Code Not Running? Boost Image Processing Speed with SIMD – This Optimization is Incredible!

by
C++ Code Not Running? Boost Image Processing Speed with SIMD - This Optimization is Incredible!

This article is based on a thorough review of relevant authoritative literature and materials, forming a professional and reliable content. All data in the article is verifiable and traceable. Special note: The data and materials have been authorized. The content of this article does not involve any biased views and objectively describes the facts with … Read more

K8S Application Orchestration and Management: Job and CronJob

by

Continuing from the previous study on Controllers, this section will focus on the practical aspects of Job and CronJob. First, consider the following question: Can we run task processes directly through Pods? If we do this, how do we solve the following issues? How to ensure that the processes within the Pod terminate correctly? If … Read more

Practical Implementation of RISC-V Vector Technology in Embedded Systems

by
Practical Implementation of RISC-V Vector Technology in Embedded Systems

In the design of CPU cores, SIMD (Single Instruction Multiple Data) is a commonly used data parallel processing method that operates on multiple data elements simultaneously through a single instruction to enhance the computational efficiency and speed of the processor. SIMD has a wide range of applications, such as image processing, scientific computing, encryption and … Read more

Understanding SIMD Execution

by
Understanding SIMD Execution

SIMD has played an increasingly important role in processors in recent years, providing significant performance improvements for multimedia applications, and most computations within GPGPU warps are also performed in a SIMD-like manner. Understanding SIMD is crucial for a comprehensive understanding of architecture. This book provides a comprehensive and detailed introduction to the design concepts and … Read more