Fei has over 10 years of development experience at Tencent and Sogou. Through this account, he shares his deep insights on performance.
This article explains the concept of CPU load balancing, the hierarchical scheduler domain and group structures in a multi‑core SoC, when and how the Linux kernel performs periodic, no‑hz, and idle load‑balancing, and outlines the step‑by‑step algorithm used to migrate tasks for balanced system performance.
This article explains the concepts, components, and lifecycle of Remote Procedure Call (RPC), compares it with HTTP, and demonstrates how to use the open‑source C++ SRPC framework to build client and server applications, including code generation, serialization, compression, and protocol handling.
This article dissects Go's net package, showing how its Listen, Accept, Read and Write functions combine goroutines with epoll to deliver synchronous‑style code that avoids costly thread switches, and walks through the underlying source code and runtime mechanisms step by step.
This article explains how Redis initializes, creates a TCP listener, registers file events, runs a single‑threaded reactor loop, accepts client connections, processes commands via a command table, and sends replies, illustrating the core backend architecture of the in‑memory database.
The article explains why Linux reports "fork: cannot allocate memory" when the real issue is PID allocation failure caused by excessive threads, analyzes the kernel's do_fork and alloc_pid implementations, and shows that newer kernels still return ENOMEM for non‑memory errors.
This article explains how a CPU handles interrupts using mechanisms like the programmable interrupt controller (PIC), advanced APIC, interrupt vectors, IDT, and how interrupt affinity and CPU affinity can be configured to balance load across multiple cores, illustrating both synchronous exceptions and asynchronous interrupts.
This article provides a step‑by‑step methodology for identifying why a Redis instance becomes slow, covering baseline latency testing, slow‑log analysis, big‑key detection, expiration patterns, memory limits, fork overhead, huge‑page effects, AOF configuration, CPU binding, swap usage, memory fragmentation, network saturation, and practical remediation techniques.
This article narrates how two friends transform secret light‑bulb blink codes into binary representations, then progressively design logic gates, adders, a full‑adder, RAM, an ALU and a programmable control unit, illustrating the fundamentals of digital circuit design and automation.
This article explains the Linux SO_REUSEPORT feature introduced in kernel 3.9, how it allows multiple processes to bind and listen on the same port, the kernel's load‑balancing mechanism, and provides practical C examples and verification steps for developers.
This article provides a thorough introduction to the Domain Name System (DNS), covering its hierarchical design, query workflow, caching mechanisms, message structure, resource record types, and common security threats along with mitigation techniques, offering essential knowledge for networking fundamentals.
