CUDA新删除 [英] CUDA new delete

问题描述

有人可以清楚地解释如果从 __ device __ 或 __ global __ 调用新的和删除关键字的行为，代码在CUDA 4.2？

Can someone give a clear explanation of how the new and delete keywords would behave if called from __device__ or __global__ code in CUDA 4.2?

内存分配在哪里，如果它在设备上是本地还是全局？

Where does the memory get allocated, if its on the device is it local or global?

它的问题的上下文条件我试图在GPU上创建神经网络，我想要一个链接的表示（像一个链表，但每个神经元存储一个链接持有权重的连接列表，以及指向其他神经元的指针），我知道我可以在内核启动之前使用 cudaMalloc 分配，但我想内核控制如何和何时网络已创建。

It terms of context of the problem I am trying to create neural networks on the GPU, I want a linked representation (Like a linked list, but each neuron stores a linked list of connections that hold weights, and pointers to the other neurons), I know I could allocate using cudaMalloc before the kernel launch but I want the kernel to control how and when the networks are created.

谢谢！

推荐答案

C ++ new 和 code>在设备堆内存上操作。该设备允许以这种方式分配全局（即，板上）存储器的一部分。 new 和 delete 的工作方式与 device malloc code> free 。

C++ new and delete operate on device heap memory. The device allows for a portion of the global (i.e. on-board) memory to be allocated in this fashion. new and delete work in a similar fashion to device malloc and free.

您可以使用运行时API调用。

您可能还会对 C ++新建/删除示例代码

这些功能需要CC 2.0或更高版本。

CC 2.0 or greater is required for these capabilities.

这篇关于CUDA新删除的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！