无法分配字节 - NS-2 中的大规模模拟
我正在使用 NS-2 对 5000 个及以上节点进行模拟。
我遇到了以下错误：
“无法分配 32026 字节”[许多此类错误] 和 '总线错误'
我已将我的代码用于多达 500 个节点，并且运行良好。现在我正在模拟大型网络。以前有人遇到过这个错误吗？
I am running simulations using NS-2 for 5000 nodes and above.
I have encountered the following errors:
'unable to alloc 32026 bytes' [many such errors] and
'bus error'
I have used my code for up to 500 nodes and it worked fine. Now I am simulating for large networks. Has anyone encountered this error before?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
你的内存不足了。如果您使用如此小的分配量收到该消息，则表明您确实内存不足；解决这个问题并不简单。
导致此问题的一种可能是在同一系统上同时运行太多其他程序，另一种可能是内存泄漏。虽然 Tcl 本身经过了仔细的内存泄漏审核，但我不能谈论 NS2，并且您的脚本本身可能会通过保留对事物的引用时间超过应有的时间而导致“泄漏”。解决这个问题需要大量的工作。
当然，您可以采用老式的“投入更多硬件来解决问题”的方法。获取 Tcl 和 NS2 的 64 位版本，并在具有 16GB 内存的系统上运行它。这个选项比以前便宜很多，因为 64 位硬件正在成为主流（终于！），但它不是免费的……
You're running out of memory. That you're getting that message with such a small allocation indicates that you are really running out of memory; fixing this is not trivial.
One possibility that could be causing this is running too many other programs on the same system at once, and another possibility is that you have a memory leak. While Tcl itself is audited carefully for memory leaks, but I cannot talk about NS2 and your scripts could themselves be causing a “leak” by holding onto references to things for longer than they should. Sorting this out is a substantial amount of work.
Of course, you could take the good old-fashioned “throw more hardware at the problem” approach. Get a 64-bit build of Tcl and NS2 and run it on a system with, say, 16GB of memory. This option is a lot cheaper than it used to be as 64-bit hardware is going mainstream (at last!) but it isn't free…