[独立平台] [生命科学类] Folding@Home

rexts217

原帖由 涼宮ハルヒ 于 2008-6-20 18:48 发表
9800GTX 4+2的供电，上800/2000长期没问题
88GTS512 3+1的供电，核心电压也低750/1830差不多了。。。

真的能上800?

涼宮ハルヒ

原帖由 rexts217 于 2008-6-20 20:01 发表


真的能上800?

恩，基本上只要不碰到雷的话 基本都能上800的

原帖由 sakeki 于 2008-6-20 20:01 发表
为何返厂呢？

某天直接点不亮了。。

Youth

最好能开新帖讨论项目的具体问题，这样更方便大家查看新闻帖，也方便问题的讨论：）

sakeki

原帖由 Youth 于 2008-6-20 20:42 发表
最好能开新帖讨论项目的具体问题，这样更方便大家查看新闻帖，也方便问题的讨论：）

OK

vmzy

June 24, 2008
GPU2 update

The GPU2 client has been out for a while now for our newest platform, NVIDIA, and I wanted to give an update.  We're making great progress on several fronts of the beta testing of this client, with improvements to the CPU utilization and visualization (which currently is pretty much broken) coming soon.  We are also working to support multi-gpu configurations.  These are our highest non-science priorities.  On the science-side, we're scrubbing the GPU clients to make sure the results make sense.  GPU programming is challenging for many reasons, especially due to reduced precision and the complexity of using lots of threads in flight, and so it's important to make sure the results are accurate.  So far, the results look promising.  Once the GPU2 cores are  completely validated and these client issues  are addressed, we'll take the client out of beta and make a push to get an even greater adoption of this new client platform.


KNOWN ISSUES FOR GPU2/NVIDIA

Viewer doesn't work (coming soon).  This will require a core upgrade, which is in the works.

GTX280 driver version.  For pre-GTX280 cards, we recommend version 174.55 of the CUDA driver.   We recommend 177.35 for GTX cards.

CPU usage can be strange (we're looking into this).  The CPU utilization can spike on certain machines.  We have an idea for what's the issue and Scott LeGrand at NVIDIA is working on a fix.

UNSTABLE_MACHINE error if too many EUE's (not really a bad thing -- a true feature of the client).  If you see this error in your client log, it means that there is some problem with your configuration.

大意：
GPU2/NVIDIA已经发布了，我们当前正在处理多GPU设置，CPU利用率，图形界面，以及计算精度方面的问题。

已知问题：

图形界面不能用。需要升级计算内核。

有些机器的cpu占用有问题。当前NVIDIA的Scott LeGrand 正在着手解决此事。

如果你遇到停工，那么意味着您的机器设置有问题。

shouldbe

希望尽快达到SMP和单进程CPU客户端的水平。争取做到不用理会GPU2客户端，随便使用GPU，看HDTV、玩游戏。这样就完美了。

vmzy

June 28, 2008
Folding@home's End User License Agreement (EULA) and why it's there

There's been some recent discussion of aspects of Folding@home's EULA and what constitutes a violation of the EULA.  One key issue is whether donors (end users of the FAH software) can make modifications to the code in order to make enhancements, such as multi-gpu support for NVIDIA.  While such modifications may be made with the very best of intentions, these modifications do not go through our QA (either at Stanford or our relevant partners such as NVIDIA, ATI, or Sony).  This is a big problem with 3rd party client modifications is that even very subtle modifications can create problems for the science involved.  Since there's no way for donors to QA end user modifications, we cannot support their use and must remind everyone that any modifications to any of the client binaries is a true EULA violation.

The bottom line is that even though some modifications may be done with the very best of intentions (eg to help the Folding@home project, never to hurt it), we can't tell whether those intentions translate into true help for the project without QA.  Also, perhaps more importantly, it's hard to say that "some violations of our EULA" or "some modifications to the client code" are ok.  There's no way to know whether something is ok until it goes through QA.

Finally, it's been my experience with FAH that donors can often have very good ideas on how to improve the FAH software.  I have a suggestion as an alternative to 3rd party releases of client modifications.  Please go to the folding community forum (http://foldingforum.org) and make a post with the jist of your idea and someone from the FAH development team will get in touch with you.  If we're already working to QA a fix for that type of change (eg multi-gpu support), it would make more sense for us to finish our existing QA than to start fresh with some other solution (since QA of a new approach would delay the release of that feature).  We are always looking for beta testers for new functionality and so that's another way donors can help the project and get access to early features.

For example, we are actively working on several key issues right now on the GPU2/NVIDIA client, including multi-gpu support and better desktop performance, and hope to have something to release very soon (days timescale, assuming that the fixes get past QA without problems).  If you're interested in beta testing some of these features, please make a post in the forum.

I wanted to thank everyone for their help to improve FAH clients and the FAH experience in general and hope we can work together to get the needed fixes in.  We are grateful for all the help we get from FAH donors.  However, I hope everyone can appreciate the need for QA, especially radically new and complex clients like GPU2, and that we can find a way to work together to get the fixes needed.

大意：
解释了为什么不鼓励私自修改客户端的原因。鼓励有想法和希望参加内测的用户到官方论坛留言。如果不出意外，官方很快会发布新版的GPU2/NVIDIA客户端，解决多GPU支持及性能等问题。

vmzy

June 30, 2008
Assignment server code upgrade
We've upgraded our code in our main Assignment server to improve some load balancing issues.  This upgrade also has new code for how the url for new core downloads are sent.  We have tested this code on less important AS's first, but if you start seeing problems with core downloads, please make a post in our main forum (http://foldingforum.org), ideally with some of the log with -verbosity 9 to show the core download url.
大意：
我们调整了AS（任务分配）服务器的代码，以解决负载均衡问题。现在可以根据访问者位置，发送不同的下载地址。如果你遇到问题，请到官方论坛发帖，最好贴上加上-verbosity 9参数的详细日志，我们好知道究竟是哪个下载地址出了问题。

NVIDIA core upgrade to 1.06

We're rolling out an automatic core upgrade for GPU2/NVIDIA clients to v1.06.  Please post in the forum if you're having any new problems with your GPU2 client after this (although the 1.06 core has been tested very thoroughly so far).
大意：
NVIDIA内核会自动更新至1.06，如果您遇到问题请到官方论坛发帖告知。

vmzy

July 08, 2008
FAH/SMP Q & A
There was a good question in the forum that I thought others would be curious to hear:

From Vijay's blog entries it would seem that the SMP client has some fundamental advantages over running multiple singlecore client, but I can't really think of how that might be. Do you know of some architectural overview of how the MPI stuff is being used in this context?


We could just run multiple independent clients, but this would be throwing away a lot of power.  What makes an SMP machine special is that it is more than just the sum of the individual parts (CPU cores), since those cores can talk to each other very fast.  In FAH, machines talk to each other when they return WUs to a central server, say once a day.  On SMP this happens once a millisecond or so (or faster).   That 86,000,000x speed up in communication can be very useful, even if there isn't 100% utilization in the cores themselves.  

The easy route would have been to run multiple single-CPU FAH-cores (this is what other projects do), but that would be a big loss for the science, as this throws away a very, very powerful resource (fast interconnects between CPUs).  Indeed, it is this sort of fast interconnect which makes a supercomputer "super", since the CPUs in supercomputers (eg BlueGene) are pretty slow, but the communication between cores is very, very fast.  


We've done a lot to develop algorithms for FAH-style internet connections between CPUs, but there are some calculations which require fast interconnects, and that's where the FAH/SMP client is particularly important.  By allowing us to do calculations that we couldn't do otherwise, the science is pushed forward significantly (and we thus reward SMP donors with a points bonus due to this extra science done and the extra hassle involved in running the SMP client).

I guess it remains to be seen if we can pull off MPI on FAH to the point where it works effortlessly, but so far Lin and OSX look pretty good, so we're close.  The A2 core should hopefully seal the deal.  Now, the main task is getting Windows/SMP behaving well ...

大意：
解释了，SMP客户端为什么速度快的原因。它充分利用了多核cpu的内核间高速通信功能。smp并不是2个客户端独立运算2个WU，而且协同计算一个WU。使得SMP的计算速度大于2个客户端的计算速度。

[ 本帖最后由 vmzy 于 2008-7-9 13:27 编辑 ]

vmzy

July 09, 2008
Modification to Assignment Server code

We've made a modification to how the Assignment Server (AS) code works.  We've done some initial testing and are now releasing it to the backup AS (assign2.stanford.edu).  If that checks out, we'll release it to the main AS.  The change involves how we assign SMP WUs.  If you're seeing something strange there (eg SMP WUs' to non-SMP clients or vice versa), please let us know in the forum (http://foldingforum.org).
大意：
修改了任务分配代码，现在已经放到备用服务器上进行公测，如果您遇到任务分配问题（例如把，SMP任务发给普通客户端，或者把普通任务发给SMP客户端）请到论坛发帖告诉我们。

vmzy

新的AS代码，貌似有大问题，负载均衡好像配置错了。
今天早上连接服务器失败率高达97%，而且老是分配到一个已经没有任务的服务器上。用了将近3个小时才下到新WU。郁闷啊！

vmzy

果然是新代码的负载均衡有问题，现在官方已经恢复到上一版代码了，现在任务下载应该正常了。
官方效率不错，虽然美国是凌晨，但是BUG上报后很快就解决了。

vmzy

July 12, 2008
Sharing FAH results -- raw data now available

We've been working behind the scenes for a while to find the right way to share the raw data results from Folding@home.  We've partnered with the Simbios National Biomedical Computing Center to provide data for download.  If you're curious, check out our first data set project page

https://simtk.org/home/foldvillin

We'll be releasing more data as time goes on.  Our hope is that by making the raw data openly available, this will greatly supplement the published results.
大意：
公开结果原始数据。
译者注：这下某些‘有识之士’该闭上您老的‘臭嘴’了吧！

shouldbe

有识之士的“臭嘴”？他不公开数据就总会有人说。既然是利用了大家的计算能力，就应该把计算结果公开。作为一个donor，有权利要求公开数据。译者也许不在乎这种权利，但不等于别的donor不在乎。当这样的donor，就算被人说骂做臭嘴也心安理得。

vmzy

July 27, 2008
Collection server code update

For a while, the collection servers (CS's) haven't been working well.  We've been looking into why.  A few weeks ago we overhauled server-CS interaction and added some more CS's to help the load.  We have continued to look into CS issues and have some new ideas.

With these new ideas, we've made an update to how the collection server works.  We think this will help clients send back results significantly.  This update code has been placed on 2 CS's right now (171.64.122.76 and 171.64.122.86) and we're going to watch to see if this improves the situation.  We have a few other tweaks that might help as well, but we want to try this one first.
大意：
修改了CS（结果回收服务器）代码，现已部署到171.64.122.76 和 171.64.122.86上了，看看能不能解决任务上传不稳定的问题。
July 27, 2008
Issues with maintenance on GPU2 server

One of our servers (171.64.122.74) had some issues during maintenance which delayed it's coming back on line.  Due to this, 171.64.122.74 was back on line on Saturday (PST) instead of Friday evening, due to an issue found during maintenance.  Luckily this maintenance is a once a year thing, so it won't be needed for a while.  It's back up now.

Due to this server being off line, the backup server (171.64.65.20) took quite a hit.  To avoid this, we're moving to longer WU's (8-24 hours) and looking to add an additional server so the servers take less of a hit.  Longer WU's alone will make a HUGE difference, since the servers would get hit much less frequently and thus the load would be much smaller.  That should greatly improve these issues.

The current situation is that all servers are up and running, but there's a lot of clients trying to access them.  We'll keep an eye on them throughout the weekend to make sure they're going well.
大意：
GPU2服务器171.64.122.74出了点小问题，导致例行维护的时间比预计时间长。
此间备用服务器承受了很大的压力，为此我加大了任务包的计算量（8-24小时），并新增了一台服务器。
现在一切恢复正常了，我们会继续密切关注这些服务器的工作情况。

译者注：还好我上面的帖子没变成口水贴，引发‘洪水’，看来以后说话要注意些了，自我检讨下。