指向 GPU 内存的指针
设备指针存放在主机变量中
指向 GPU 内存的指针 是 CoddyKit 上的免费 CUDA Academy 课时。 这是第 2 节课，共 4 节。 你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。 这是 CUDA Academy 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 CUDA Academy 课程共包含 4 节课。
本课时的部分内容尚未翻译，以英文显示。
A Pointer That Lives on the CPU
cudaMalloc hands you a normal C++ pointer variable. It sits in host memory, but the address inside it points into GPU memory. 🔗
The Address Means the Device
A device pointer holds a number that is only valid in the GPU's address space. To the CPU it is just an integer, not real RAM.
float* d_data; // host variable, device addressThe d_ Naming Habit
Programmers prefix device pointers with d_ and host pointers with h_. The name alone reminds you which side each one belongs to.
float* h_data;
float* d_data;Do Not Dereference on the Host
Writing *d_data in CPU code is a trap. The host cannot touch GPU memory directly, so this crashes or reads nonsense.
Kernels Dereference It Freely
Inside a kernel the same device pointer is perfectly valid. There, d_data[i] reads and writes real GPU memory as you expect.
__global__ void k(float* d) {
d[0] = 1.0f;
}Pass It Into the Kernel
You hand the device pointer to a kernel as an argument. The GPU then uses that address to find your data.
myKernel<<<grid, block>>>(d_data);Two Pointers, Two Spaces
Keep host and device pointers strictly apart. A host pointer is meaningless on the GPU, and a device pointer is meaningless on the CPU.
cudaMemcpy Bridges the Gap
To get data between the two, you copy it. cudaMemcpy takes both pointers and moves bytes across the divide for you.
Pointer Arithmetic Still Works
You may offset a device pointer on the host, like d_data + n, to point at a slice. You just must not read through it there.
float* d_tail = d_data + 100;Mixing Them Up Is Subtle
The compiler will not stop you from confusing the two. Many CUDA bugs are simply a host pointer handed where a device pointer belonged.
Discipline Beats Debugging
Clear naming and a fixed convention save hours. Decide your d_ and h_ scheme early and never break it. 🧹
Quick Check
Let us test where a device pointer can safely be dereferenced.
Recap
You learned a device pointer lives in a host variable but addresses GPU memory: usable inside kernels, off-limits to direct host access. 🎉
常见问题解答
「指向 GPU 内存的指针」课时是免费的吗？
是的 — 「指向 GPU 内存的指针」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 CUDA Academy 课程的其余内容，请升级到 CoddyKit PRO。 CUDA Academy 课程共包含 4 节课。
「指向 GPU 内存的指针」这节课中我会学到什么？
设备指针存放在主机变量中 你通过在浏览器中直接运行的动手代码来练习 CUDA Academy，全天候 AI 导师会在你学习这节课的过程中回答你的问题。
学习 CUDA Academy 需要有经验吗？
无需任何先前经验。CoddyKit 上的 CUDA Academy 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。 这是第 2 节课，共 4 节。
「指向 GPU 内存的指针」课时需要多长时间？
大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。
我能在这节 CUDA Academy 课中编写并运行代码吗？
能。每节 CUDA Academy 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。