[E@H]出官方试用优化程序了~~~

fwjmath

在留言板上的回应是大概快了15~20%~~~
地址：http://einstein.phys.uwm.edu/power_apps.php

还有~~~不知道大家有没有注意星期三的时候Server Status那里的S5R3项目进度突然从二十几跳到了九十几~~~根据官方的说法是由于弄错了数据文件，搜索频率大于等于800Hz的WU得出的结果都没用，不过现在所有这样的WU都被删掉了~~~现在官方打算将S5R3分成两部分，一个就是现在的部分，还有大约一个星期就完了；另一个是S5R3b，专门搜索大于800Hz的频率，将会在大约一个星期后开始，这次就应该不会乌龙了~~~
[已订正]

[ 本帖最后由 fwjmath 于 2008-1-25 14:49 编辑 ]

Youth

8错8错，马上试用：）

另外，应该是S5R3而不是S3R5哦

Youth

直接用新的程序来计算老程序算了一半的任务包，下班前应该能看出来速度变化。

Tynox

Boinc会自动更新吗?

Youth

至少现在不会，如果试用结果好的话才会考虑转正吧：）

The Apps that are published here are meant for "power users", i.e. people who more or less know what they are doing with their computers and with BOINC. The instructions are short, if there are any at all, and not suitable for beginners. Support from the project team is limited.

These Apps usually trade in some features or compatibility we want to keep in the official Apps for speed, or run on rather exotic OS for which we can't provide much support. The Apps have been tested by the project so far that they run successful on at least one configuration (usually the one that they have been built on) and don't pose a danger to the project. They usually come with an app_info.xml file that uses the anonymous platform, assigns a valid application name and a serial number that allows us to easily keep track of these results in the database.

Use them at your own risk (i.e. credit).

Tynox

哦.那就不是用了.还是用正式版的吧.

fwjmath

昨天晚上糊涂了~~~

macwang

太好了，试用中。。。。。

Julian_Yuen

嗯 ，去试试看

kittyjia

真不错，正在试用

Youth

刚算好一个，因为是计算中途替换的新程序，最后的计算时间考虑到任务包本身的差异没看出太明显的差别，明天应该能再算好几个。

说是这个优化还没利用到sse之类的指令集，不知道啥时候能有百分之三五十或更多的提速：）

Julian_Yuen

能够提速的话，也算是节能减排了

Youth

91472220 36691604 25 Jan 2008 20:02:23 UTC 26 Jan 2008 10:57:31 UTC Over Success Done 40,020.52 236.18 pending
91463919 36687955 25 Jan 2008 16:27:52 UTC 26 Jan 2008 9:01:14 UTC Over Success Done 40,685.94 236.18 pending

上面这两个是优化的

91396188 36657767 23 Jan 2008 23:40:44 UTC 26 Jan 2008 2:46:40 UTC Over Success Done 44,534.99 236.18 pending
91383243 36651955 23 Jan 2008 18:21:08 UTC 25 Jan 2008 20:02:21 UTC Over Success Done 46,816.86 236.18 236.18

上面这两个是部分优化的，下面是原来官方程序的，还是能看出一些优化的效果的，即使考虑到任务包本身完成时间就不固定

91364692 36643586 23 Jan 2008 6:47:17 UTC 25 Jan 2008 11:36:10 UTC Over Success Done 49,157.48 236.18 pending
91353418 36638375 23 Jan 2008 0:46:08 UTC 23 Jan 2008 23:40:44 UTC Over Success Done 49,894.25 236.18 pending
91276159 36602186 22 Jan 2008 13:02:42 UTC 23 Jan 2008 18:21:08 UTC Over Success Done 50,664.28 236.18 pending
91276153 36594464 22 Jan 2008 6:03:43 UTC 23 Jan 2008 6:47:17 UTC Over Success Done 52,402.64 236.18 236.18
91276150 36592329 21 Jan 2008 20:12:24 UTC 23 Jan 2008 0:46:06 UTC Over Success Done 52,835.83 236.18 236.18
91272257 36602185 21 Jan 2008 14:29:11 UTC 22 Jan 2008 13:02:42 UTC Over Success Done 51,172.33 236.18 236.18
91249048 36591697 20 Jan 2008 17:27:16 UTC 22 Jan 2008 6:03:43 UTC Over Success Done 60,873.95 236.18 236.18
91238723 36587014 20 Jan 2008 11:00:58 UTC 21 Jan 2008 20:12:24 UTC Over Success Done 60,956.64 236.18 236.18
91213971 36575604 19 Jan 2008 22:15:18 UTC 21 Jan 2008 14:29:11 UTC Over Success Done 55,287.02 236.18 236.18
91195254 36567019 19 Jan 2008 16:33:27 UTC 20 Jan 2008 17:27:16 UTC Over Success Done 54,596.09 236.18 pending
91165351 36540637 18 Jan 2008 22:32:15 UTC 20 Jan 2008 11:00:56 UTC Over Success Done 51,980.39 236.18 236.18
91163941 36552996 18 Jan 2008 17:33:41 UTC 19 Jan 2008 22:15:18 UTC Over Success Done 53,250.88 236.18 236.18
91136806 36442795 18 Jan 2008 7:53:42 UTC 19 Jan 2008 16:33:27 UTC Over Success Done 46,498.81 236.18 236.18
91119002 36531799 17 Jan 2008 20:46:50 UTC 18 Jan 2008 22:32:15 UTC Over Success Done 49,501.88 236.18 236.18

前朝遗少

试验中。

jintan

这个怎么用呢？？？