---
title: "WingEdge777 (@wing-edge-777) | Hashnode"
description: "success is fleeting, failure is the constant. Read the latest articles by WingEdge777 on Hashnode."
url: "https://hashnode.com/@wing-edge-777"
---

@wing-edge-777

# WingEdge777

@wing-edge-777 · Shanghai, China · Joined April 2026

success is fleeting, failure is the constant.

[Website](https://www.wingedge777.com/en)

### [\[CUDA in Practice\] SGEMM — Beating cuBLAS: A Deep Dive into Peak-Performance Matrix Multiplication in Pure CUDA C++](https://wingedge777.hashnode.dev/cuda-in-practice-sgemm-beating-cublas-a-deep-dive-into-peak-performance-matrix-multiplication-in-pure-cuda-c?utm_source=hashnode&utm_medium=feed)

A Deep Dive into Peak-Performance Matrix Multiplication in Pure CUDA C++

27 min read · [wingedge777.hashnode.dev](https://wingedge777.hashnode.dev/cuda-in-practice-sgemm-beating-cublas-a-deep-dive-into-peak-performance-matrix-multiplication-in-pure-cuda-c?utm_source=hashnode&utm_medium=feed)

[May 29](https://hashnode.com/posts/cuda-in-practice-sgemm-beating-cublas-a-deep-dive-into-peak-performance-matrix-multiplication-in-pure-cuda-c/6a19590e52c4918e2648c084)

### [\[CUDA in Practice\] RoPE — Why Kernel Fusion in Hand-Written Operators Matters: Reducing Memory Traffic and Launch Overhead](https://wingedge777.hashnode.dev/cuda-in-practice-rope-why-kernel-fusion-in-hand-written-operators-matters-reducing-memory-traffic-and-launch-overhead?utm_source=hashnode&utm_medium=feed)

AI compilers are evolving fast. In many cases, torch.compile plus JIT optimization in PyTorch can deliver striking speedups, to the point that people often say, "hand-written operators are no longer n

8 min read · [wingedge777.hashnode.dev](https://wingedge777.hashnode.dev/cuda-in-practice-rope-why-kernel-fusion-in-hand-written-operators-matters-reducing-memory-traffic-and-launch-overhead?utm_source=hashnode&utm_medium=feed)

[May 24](https://hashnode.com/posts/cuda-in-practice-rope-why-kernel-fusion-in-hand-written-operators-matters-reducing-memory-traffic-and-launch-overhead/6a1292451a3cf7bffc816a10)

### [\[CUDA in Practice\] Matrix Transpose — From Padding to XOR Swizzle: The Art of Shared Memory Access Optimization](https://wingedge777.hashnode.dev/cuda-in-practice-matrix-transpose-from-padding-to-xor-swizzle-the-art-of-shared-memory-access-optimization?utm_source=hashnode&utm_medium=feed)

Note: Text translated by AI. Code crafted by human. Matrix transpose is one of the most fundamental operations in deep learning and high-performance computing. The deceptively simple coordinate swap

33 min read · [wingedge777.hashnode.dev](https://wingedge777.hashnode.dev/cuda-in-practice-matrix-transpose-from-padding-to-xor-swizzle-the-art-of-shared-memory-access-optimization?utm_source=hashnode&utm_medium=feed)

[Apr 9](https://hashnode.com/posts/cuda-in-practice-matrix-transpose-from-padding-to-xor-swizzle-the-art-of-shared-memory-access-optimization/69d7a9e6fa7251682eac9e6b)

### [A Deep Dive into DeviceQuery: Understanding Your GPU Hardware](https://wingedge777.hashnode.dev/a-deep-dive-into-devicequery-understanding-your-gpu-hardware?utm_source=hashnode&utm_medium=feed)

0\. Preface Before writing a single line of high-performance CUDA code, you must know your silicon. deviceQuery is often the first command a developer runs, yet its output is usually ignored. This post

9 min read · [wingedge777.hashnode.dev](https://wingedge777.hashnode.dev/a-deep-dive-into-devicequery-understanding-your-gpu-hardware?utm_source=hashnode&utm_medium=feed)

[Apr 5](https://hashnode.com/posts/a-deep-dive-into-devicequery-understanding-your-gpu-hardware/69d22b6140c9cabf44af37f6)
