Updates on Tsinghua Nano Tech
log in

Advanced search

Message boards : News : Updates on Tsinghua Nano Tech

Author Message
wenjing wu
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 13 Sep 10
Posts: 161
Credit: 751,216
RAC: 0
Message 597 - Posted: 24 May 2012, 11:10:08 UTC
Last modified: 26 May 2012, 5:55:41 UTC

基于志愿者的反馈信息,我们解决了应用程序的bug,目前的LAMMPS应用程序更稳定!
请支持清华大学纳米技术研究组!根据文档,选择Tsinghua Nano Tech Research应用,即可以接收LAMMPS作业。

由于目前LAMMPS应用还处于更大规模的测试阶段,我们需要更多的志愿主机的参与,来发现和解决应用程序中的bug。如果在测试过程您的主机花了几个小时运行了一个任务,但是出现了“计算错误”,我们会根据服务器上的记录给您补偿相应的积分。所以请不要犹豫,继续支持LAMMPS吧!

Based on feedbacks from our volunteers, we have fixed a few bugs in the application,and now we are sending out more jobs!Please support Scientists from Tsinghua Nano Tech research group!To do so, please follow document to check "Tsinghua Nano Tech Research" application to crunch LAMMPS jobs!

Notice: We are still trying to identify and fix the bugs in the application, so
some of the jobs may run on your hosts for hours and end up with a computation error, however in this case, we can still compensate you with the equivalent credit by the records from server! So please do not hesitate to try it :)
____________
加油!CAS@home!我们帮助科学家跟时间赛跑!
Go CAS@home! We help scientists to race against time!

astroWX
Send message
Joined: 27 Sep 10
Posts: 29
Credit: 147,739
RAC: 0
Message 599 - Posted: 24 May 2012, 20:31:11 UTC

LAMMPS is not the option I see. Rather it offers "Run only the selected applications [ ] Tsinghua Nano Tech Research" which changes to
"Run only the selected applications (all applications)" after clicking 'Update Preferences.' I assume that all refers to LAMMPS. (Also checked the box to accept test tasks.)

So far, only SCT 5.14 tasks received.
____________
Greetings from the US Pacific Northwest.

wenjing wu
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 13 Sep 10
Posts: 161
Credit: 751,216
RAC: 0
Message 600 - Posted: 25 May 2012, 2:18:35 UTC - in response to Message 599.
Last modified: 25 May 2012, 2:20:33 UTC

You are right! It is Tsinghua Nano Tech Research! Thanks for poiting it out!
As you can check from the monitoring page, there are available jobs, but LAMMPS application generate different results on different platforms due to float point calculation,so we use a strategy which is to send different instances of the same job to identical platforms(with the same OS and CPU architecture), so your host may not get suitable Instancese of jobs right away!



LAMMPS is not the option I see. Rather it offers "Run only the selected applications [ ] Tsinghua Nano Tech Research" which changes to
"Run only the selected applications (all applications)" after clicking 'Update Preferences.' I assume that all refers to LAMMPS. (Also checked the box to accept test tasks.)

So far, only SCT 5.14 tasks received.

____________
加油!CAS@home!我们帮助科学家跟时间赛跑!
Go CAS@home! We help scientists to race against time!

Dirk Broer
Send message
Joined: 20 May 12
Posts: 7
Credit: 88,013
RAC: 0
Message 603 - Posted: 25 May 2012, 8:56:03 UTC - in response to Message 600.

Can we expect a GPU version of the LAMMPS application in the future? GPU architecture is suited very well for floating points calculation, and up to 10 or more times faster. If it is programmed in OpenCL it might be easier to develop a nVidia and a Ati version of it, as CUDA is nVidia only and CAL/Brook is Ati only (and obsolete).

wenjing wu
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 13 Sep 10
Posts: 161
Credit: 751,216
RAC: 0
Message 604 - Posted: 25 May 2012, 10:51:19 UTC - in response to Message 603.

Yes, it is possible! LAMMMPS includes a bunch of user packages which can be used for different molecular simulations, and it seems that some of the packages already have GPU versions! The plan of LAMMPS on CAS@home is to support different research groups, and we will choose the GPU version if the user package used by a specific research group has the implementation of GPU version.

Can we expect a GPU version of the LAMMPS application in the future? GPU architecture is suited very well for floating points calculation, and up to 10 or more times faster. If it is programmed in OpenCL it might be easier to develop a nVidia and a Ati version of it, as CUDA is nVidia only and CAL/Brook is Ati only (and obsolete).


____________
加油!CAS@home!我们帮助科学家跟时间赛跑!
Go CAS@home! We help scientists to race against time!

Eugene.C
Send message
Joined: 26 Mar 12
Posts: 2
Credit: 26,598
RAC: 0
Message 605 - Posted: 25 May 2012, 11:25:49 UTC

我在计算lammps时。CPU占用是满的,但是进度的数字不会涨,暂停一下,再继续,进度的数字会变化。这样我不知道到底什么时候能算完,是不是计算程序和BOINC manager之间的通讯还不太好?

astroWX
Send message
Joined: 27 Sep 10
Posts: 29
Credit: 147,739
RAC: 0
Message 606 - Posted: 25 May 2012, 21:42:08 UTC - in response to Message 600.

... As you can check from the monitoring page, there are available jobs, but LAMMPS application generate different results on different platforms due to float point calculation,so we use a strategy which is to send different instances of the same job to identical platforms(with the same OS and CPU architecture), so your host may not get suitable Instancese of jobs right away!

Good to know. Thanks for the feedback, Wu Wenjing. Appreciated.
____________
Greetings from the US Pacific Northwest.

wenjing wu
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 13 Sep 10
Posts: 161
Credit: 751,216
RAC: 0
Message 608 - Posted: 26 May 2012, 3:40:38 UTC - in response to Message 605.

如果你计算的是batch_23之前的作业的话,可能会出现进度条没有变化。如果是batch_23,有的任务需要2个小时,所以你可能不能立即看到这个进度条的变化。这个进度显示所依赖的文件是由LAMMPS的用户提供的,他们根据模拟的进度实时计算出来的。BOINC Manager 是根据这个文件来显示的。

我在计算lammps时。CPU占用是满的,但是进度的数字不会涨,暂停一下,再继续,进度的数字会变化。这样我不知道到底什么时候能算完,是不是计算程序和BOINC manager之间的通讯还不太好?


____________
加油!CAS@home!我们帮助科学家跟时间赛跑!
Go CAS@home! We help scientists to race against time!

Eugene.C
Send message
Joined: 26 Mar 12
Posts: 2
Credit: 26,598
RAC: 0
Message 609 - Posted: 26 May 2012, 6:23:46 UTC - in response to Message 608.

多谢回复,没注意是batch多少的,昨天没有手动中止任务,它自己算完了。后来查看服务器上任务状态是计算错误,知道是属于正常情况,所以不担心了。继续支持你们的工作。

wenjing wu
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 13 Sep 10
Posts: 161
Credit: 751,216
RAC: 0
Message 610 - Posted: 26 May 2012, 11:09:44 UTC - in response to Message 609.

是的,那个计算错误是由于版本v1.15中的一个bug,我昨天已经解决了bug,并发布了一个新的版本v1.16。作业batch 25中没有再出现这个计算错误的问题了。对于之前计算完成了,但是在GUI上显示为计算错误的主机,我们正在想办法统计出来,进行积分的补偿! 谢谢支持!

多谢回复,没注意是batch多少的,昨天没有手动中止任务,它自己算完了。后来查看服务器上任务状态是计算错误,知道是属于正常情况,所以不担心了。继续支持你们的工作。


____________
加油!CAS@home!我们帮助科学家跟时间赛跑!
Go CAS@home! We help scientists to race against time!

ztxforever
Send message
Joined: 28 May 12
Posts: 2
Credit: 123,330
RAC: 0
Message 611 - Posted: 29 May 2012, 23:48:38 UTC

2012/5/30 4:38:21 CAS@home Aborting task batch_27_184_4: exceeded disk limit: 48.16MB > 47.68MB
2012/5/30 4:38:32 CAS@home Computation for task batch_27_184_4 finished

检查Log发现昨天晚上有关LAMMPS的一条信息,提示超过磁盘限制,我不清楚是否是属于Bug,所以提交上来,或许会有帮助。

wenjing wu
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 13 Sep 10
Posts: 161
Credit: 751,216
RAC: 0
Message 612 - Posted: 30 May 2012, 2:47:40 UTC - in response to Message 611.
Last modified: 30 May 2012, 2:49:13 UTC

是的,您提供的信息非常有用,我们的作业提交系统中对磁盘空间的估计有bug,我们昨天已经发现了这个问题,已经解决了!
Yes, there was a bug in our job submission system on the estimation of required disk space from BOINC client, this happens in batch 27 and 28. We already fixed this bug yesterday as we spotted it.

2012/5/30 4:38:21 CAS@home Aborting task batch_27_184_4: exceeded disk limit: 48.16MB > 47.68MB
2012/5/30 4:38:32 CAS@home Computation for task batch_27_184_4 finished

检查Log发现昨天晚上有关LAMMPS的一条信息,提示超过磁盘限制,我不清楚是否是属于Bug,所以提交上来,或许会有帮助。

____________
加油!CAS@home!我们帮助科学家跟时间赛跑!
Go CAS@home! We help scientists to race against time!

hwt
Send message
Joined: 27 Nov 10
Posts: 3
Credit: 30,956
RAC: 0
Message 615 - Posted: 3 Jun 2012, 5:25:30 UTC - in response to Message 597.

我想问一下,我已经勾选Tsinghua Nano Tech这个项目,但为什么收不到这个项目的任务?

eruda
Send message
Joined: 5 Aug 10
Posts: 4
Credit: 473,968
RAC: 0
Message 616 - Posted: 4 Jun 2012, 8:17:18 UTC
Last modified: 4 Jun 2012, 8:17:32 UTC

Did anyone notice the 'Database Error' on top of the forum?

wenjing wu
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 13 Sep 10
Posts: 161
Credit: 751,216
RAC: 0
Message 617 - Posted: 5 Jun 2012, 2:53:15 UTC - in response to Message 615.

这个很正常,因为这个时间段服务器上可能没有要发出的作业,所以客户端自然就收不到作业了。可以看这个页面,如果Application 的Unsent的 任务数目是0的话,就说明服务器上没有要发送的作业了。因为服务器上的作业也是由清华的用户通过我们的作业提交系统提交的,有的时间段,作业会比较多,有的时间段,作业比较少。

我想问一下,我已经勾选Tsinghua Nano Tech这个项目,但为什么收不到这个项目的任务?


____________
加油!CAS@home!我们帮助科学家跟时间赛跑!
Go CAS@home! We help scientists to race against time!

wenjing wu
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 13 Sep 10
Posts: 161
Credit: 751,216
RAC: 0
Message 618 - Posted: 5 Jun 2012, 2:54:39 UTC - in response to Message 616.

Yes, One of the database table was crashed, it is fixed now. Thanks!

Did anyone notice the 'Database Error' on top of the forum?


____________
加油!CAS@home!我们帮助科学家跟时间赛跑!
Go CAS@home! We help scientists to race against time!


Post to thread

Message boards : News : Updates on Tsinghua Nano Tech