CosMos下载(找了很久才找到的) 结合colinux与OpenMosix 让windows上使用linux集群成为可能。

geopower 2006-02-08 10:17:55

Ian Latter has reposted the CosMos v0.1.5.0
From: Bruce Knox <bknox@...>
Subject: Ian Latter has reposted the CosMos v0.1.5.0
Newsgroups: gmane.linux.cluster.openmosix.general, gmane.linux.cluster.openmosix.devel
Date: 2006-01-15 02:27:17 GMT

Ian Latter has reposted the CosMos v0.1.5.0 installation file on his
http://midnightcode.org/projects/chaos/ site.

No, this is not an updated version, but Ian has reposted it after recieving a number of requests for CosMos
which is no longer available from its original site. The CosMos setup-0.1.5.0.exe is a Windows
installation exe. I tested it when it was first released, as I recall, and had no problem using it. (Note
that Andreas Schäfer and Christian Kauhaus have more current and somewhat related work, The Harpy
Project, on running coLinux with openMosix.)

Bruce

The current file is:

CosMos v0.1.5.0 August 16, 2004 reposted at:
http://midnightcode.org/projects/chaos/code/setup-0.1.5.0.exe
http://midnightcode.org/projects/chaos/code/setup-0.1.5.0.exe.md5

below is a snippet from the original List postings to openmosix-general and openmosix-devel August 15,
2004 (with the file location updated):
...
CosMos-0.1.5.0;
Brief: A virtual (Linux and openMosix) cluster node for Win32 platforms - CosMos is the supercomputer for
your workstations.
Packages; Linux-2.4.26 kernel with openMosix-2.4.26-20040706 and openmosix-tools-0.3.6-2.
Software; http://midnightcode.org/projects/chaos/code/setup-0.1.5.0.exe
Size; 4.7Mbytes
MD5; dd8d4722526c8dfd293d29f2765aa388
or http://midnightcode.org/projects/chaos/code/setup-0.1.5.0.exe.md5

CHAOS-1.5 and CosMos-0.1.5.0 are compatible, and both can partake in the same openMosix cluster.

CosMos is experimental. There are a lot of things we still don't know about openMosix on coLinux - such as why
programs don't auto-migrate away from Windows cluster nodes. However, if you have a hard/dedicated
master/central/head node, then programs will migrate from that head node, to Win32 nodes correctly --
from experiments done to date.

To get CosMos to work on Windows using the PCAP driver, you need to specify the name of the Win32 NIC that you
are going to bridge with. I have tried to autodetect this, but the detection is not in sync with the coLinux
bridge driver, so you may need to experiment. Tips for connectivity;

-- choose "other" from the select list .. and type a name ...
-- enter names that are brief but matching your NIC (Ie; instead of "Intel(R) PRO/100 VM Network
Connection", just type "Intel").
-- if your NIC is bridged using a Microsoft bridge adapter, then the name of your NIC is actually "Microsoft"
-- many systems will default to the name "Local Area Connection".
-- If all else fails, boot a node using "Power On Debug" and see what error coLinux aborts with (it lists the
adapters on startup).

The only configuration GUI for CosMos is in the installer. If you don't want to uninstall and reinstall,
then see config.xml in the installed directory.

See the coLinux FAQs at http://www.colinux.org/ if you need more help.

My only concern with the release of these two versions is that the DHCP leasing hasn't been tested well. It
works, but I'm not certain that it retains its lease correctly. Test it before deploying it in production environments.
...
Regards,

Ian Latter
...全文
250 回复 打赏 收藏 转发到动态 举报
写回复
用AI写文章
回复
切换为时间正序
请发表友善的回复…
发表回复
内容概要:本文围绕基于Basisformer模型的时间序列锂离子电池SOC(State of Charge,荷电状态)预测展开研究,提出一种融合Transformer架构与基函数映射机制的深度学习方法,旨在提升锂电池SOC估计的精度与鲁棒性。该方法采用PyTorch框架实现,通过构建轻量化Basisformer网络,对电池运行过程中采集的电压、电流、温度等多维时序数据进行高效建模,充分挖掘数据中的长期依赖关系与非线性动态特征,从而实现高精度的端到端SOC预测。文中系统阐述了模型的设计原理、网络结构细节、训练优化策略及实验验证流程,并在公开电池数据集上开展对比实验,结果表明该方法在预测准确性、收敛速度和泛化能力方面均优于传统的LSTM、CNN-LSTM及标准Transformer模型。; 适合人群:具备一定Python编程基础和深度学习理论知识,从事电池管理系统、新能源汽车、储能系统或时间序列预测相关研究的研究生、科研人员及工程技术人员。; 使用场景及目标:①应用于锂电池健康状态监控与剩余电量精确估计,提升电动汽车与储能系统的安全性与能效管理水平;②为深度学习在工业级时序预测任务中的【锂电池SOC估计】【PyTorch】基于Basisformer时间序列锂离子电池SOC预测研究(python代码实现)落地提供可复现的技术方案与优化思路;③作为基于注意力机制的电池状态预测研究的入门与进阶参考资料。; 阅读建议:建议读者结合提供的Python代码实践操作,深入理解Basisformer中基函数编码与注意力机制的融合设计,同时可尝试在不同电池工况数据上迁移验证,进一步掌握模型调参与性能优化技巧。
内容概要:本文系统阐述了基于麻雀搜索算法(SSA)的无人机三维路径规划方法,并提供了完整的Matlab代码实现。该方法采用一种新兴的群体智能优化算法——麻雀搜索算法,针对复杂三维环境(如城市、山地等)中的无人机飞行任务,构建多目标优化模型,综合考虑路径长度、飞行高度变化、威胁规避能力及转弯角度成本等因素,以实现安全、高效、低能耗的最优路径搜索。文中不仅详细介绍了SSA算法的原理及其在路径规划中的具体应用流程,还通过与其他主流智能算法(如灰狼优化GWO、粒子群PSO、鲸鱼优化WOA等)进行对比实验,验证了SSA在收敛速度、寻优精度和鲁棒性方面的优越性能。该资源属于一系列智能优化算法在无人机路径规划中应用的研究成果之一,配套仿真模型与可运行代码,便于科研复现与教学实践。; 适合人群:具备一定Matlab编程基础,熟悉智能优化算法基本原理,从事无人机路径规划、智能控制、群体智能或相关领域研究的研究生、科研人员及工程技术人员。; 使用场景及目标:① 在复杂三维地形或城市环境中规划无人机的安全飞行路径;② 对比分析麻雀搜索算法与其他智能优化算法(如PSO、GWO、WOA等)在路径规划任务中的性能差异;③ 实现兼顾路径最短、能耗最低、安全性最高的多目标无人机航迹优化。; 阅读建议:建议结合提供的Matlab代码进行仿真实验,深入理解目标函数设计、环境建模、约束处理及算法参数调优等关键环节,可进一步拓展至动态障碍物避障、多无人机协同路径规划等更复杂的应用场景。

3,287

社区成员

发帖
与我相关
我的任务
社区描述
Linux/Unix社区 实用资料发布区
社区管理员
  • 实用资料发布区社区
加入社区
  • 近7日
  • 近30日
  • 至今
社区公告
暂无公告

试试用AI创作助手写篇文章吧