vardrag 发表于 2016-8-16 00:46:29

这是拿什么东西来算的,难道是BUG?

本帖最后由 vardrag 于 2016-8-16 00:48 编辑

亮点自寻

iwait 发表于 2016-8-16 01:10:29

设备安装数 9,299

vardrag 发表于 2016-8-16 02:03:29

iwait 发表于 2016-8-16 01:10
设备安装数 9,299

对于大学机房来说也算正常

Biofluid 发表于 2016-8-16 04:23:29

Syracuse University是雪城大学。。。人家一所大学总电脑数有这么多不奇怪啊!

vardrag 发表于 2016-8-16 11:19:59

Biofluid 发表于 2016-8-16 04:23
Syracuse University是雪城大学。。。人家一所大学总电脑数有这么多不奇怪啊!

你看看他的得分,再看看总运行时间,再比较一下你和他每个任务的所需时间。

horst1981 发表于 2016-8-16 11:41:17

平均每条结果运行时间3:30,真是神速啊,难道有神U

mjy 发表于 2016-8-16 12:08:19

贴wcg的水区问问看
https://secure.worldcommunitygrid.org/forums/wcg/listthreads?forum=41

vardrag 发表于 2016-8-16 12:12:23

mjy 发表于 2016-8-16 12:08
贴wcg的水区问问看
https://secure.worldcommunitygrid.org/forums/wcg/listthreads?forum=41 ...

洋文不行啊

vardrag 发表于 2016-8-16 12:19:38

horst1981 发表于 2016-8-16 11:41
平均每条结果运行时间3:30,真是神速啊,难道有神U

看样子他主要还是算MCM,我这里i5-4590每个任务是2小时多一点,他这3分半相当于120G的主频,效率是我的34倍,就算是神U+软件优化也不会这么快吧?

flybike007. 发表于 2016-8-16 12:27:35

设备数真多

mjy 发表于 2016-8-16 13:13:55

原来wcg已经有人讨论过了
https://secure.worldcommunitygrid.org/forums/wcg/viewthread_thread,39372_offset,0#527671

其中比较重要的回复是knreed和Shaky Jake的I took a look at the devices and results for them. They have a nice cluster up and running (based on quick google searching it looks like they are back filling a condor grid with work from us). The elapsed time being reported by the results is in the range you would expect - but they are all being reported as 0.0 cpu time. We see this once in awhile on different devices and it usually has to do with how the client was installed.

However - the important thing here is that they are returning legitimate and valid work for the program. They won't earn many badges with things this way, but since they have been contributing like this for a couple of years, I'm guessing that they are ok with that.
knreed说的是返回结果都是有效的。
I have seen this on several WUs on other machines. Alll of them list the OS as Vertual. Mine is the _2.
I have looked at a number of valid entries with 0:00 CPU time and they all have the OS of Linux 3.2.0-56-virtual.

MCM1_ 0124750_ 0605_ 2-- Linux 3.13.0-37-generic 736 Valid 7/25/16 14:35:52 7/25/16 20:59:06 5.96 112.4 / 98.9
MCM1_ 0124750_ 0605_ 1-- Linux 2.6.32-504.12.2.el6.x86_64 736 Error 7/24/16 18:41:13 7/25/16 14:34:51 0.00 0.0 / 0.0
MCM1_ 0124750_ 0605_ 0-- Linux 3.2.0-56-virtual 736 Valid 7/24/16 18:40:53 7/24/16 21:14:23 0.00 85.4 / 98.9
Shaky Jake说的是某些Linux虚拟机返回结果时时间为0

我个人估计是这个学校搞了个集群来算,九千多个设备实际上是虚拟机,而这些虚拟机返回的大部分结果时间显示为0(是bug),但实际计算时间远不止如此。

horst1981 发表于 2016-8-16 14:08:40

mjy 发表于 2016-8-16 13:13
原来wcg已经有人讨论过了
https://secure.worldcommunitygrid.org/forums/wcg/viewthread_thread,39372_off ...

这样算不算合理利用系统漏洞

horst1981 发表于 2016-8-16 14:22:57

vardrag 发表于 2016-8-16 12:19
看样子他主要还是算MCM,我这里i5-4590每个任务是2小时多一点,他这3分半相当于120G的主频,效率是我的34 ...

120G
简值黑科技。

刚上了BOINCstats上搜了一下ID,好像没有符合的。
感觉最符合的那个才13台服务器(X系列和E系列的U)。

vardrag 发表于 2016-8-16 14:28:45

horst1981 发表于 2016-8-16 14:22
120G
简值黑科技。




信息基本上都是隐藏的

horst1981 发表于 2016-8-16 14:38:18

vardrag 发表于 2016-8-16 14:28
信息基本上都是隐藏的

应该是这个,WCG排名也是22。


另外,牌子没你的好看


页: [1] 2
查看完整版本: 这是拿什么东西来算的,难道是BUG?

论坛官方淘宝店开业啦~