[独立平台] [生命科学类] Folding@Home

vmzy

MARCH 14, 2013
GPU QRB update

The Quick Return Bonus (QRB) gives more points when WUs are completed quickly.  This helps keep the points in line with the science.  Now with the GPU core maturing, our plan is to treat all WUs identically, i.e. benchmark on a single benchmark machine (SMP) and use those points. Now that we can do just about any calculation on any piece of hardware, it's strange to benchmark them separately. That wasn't the case before where the capabilities of the GPU and SMP cores were very different.

With the new GPU core (17), we'll have that matching capability. Our plan is to introduce QRB to GPUs with the rollout of production core 17 WUs.
大意：
GPU（17内核任务）开始实现奖励分制度。

GPU core progress & general design philosophy

We often have to make difficult decisions on what hardware to support in the future, including adding new platforms or removing existing ones.  Removing existing platforms always leads to a lot of disruptive change for donors, so we try to do this as rarely as we can.  In particular, in the GPU1 to GPU2 transition, there was a big change done quickly, which was extremely hard on donors.

For GPUs in particular, the central issue is that in general, GPU technology keeps on progressing and GPU manufacturers come up with new ways to do things which make the old ones obsolete. So, it's probably safe to say that until hardware design innovation changes, *eventually* older GPUs will become obsolete for FAH. We try to keep as many GPUs working as long as possible, but eventually it becomes a losing battle, as we have only a fixed number of programmers and more GPU types (even from a given vendor, say different CUDA capable levels) require more programmers to keep up, and eventually we run out of resources.

Right now, there's a division with the Fermi cards. Fermi and later cards have powerful new capabilities that the older cards do not have. So, I can imagine eventually we'll run out of tricks to support the older cards. I can't predict when that will be, as it depends on lots of things, but I can say we're trying hard to support everything as long as we can.

One of the biggest issues is that scientific needs can change based on where the science takes us and that's particularly hard for us to predict.  We'll try to let donors know as soon as we can if there will be any changes, but some donors questions prompted this blog post so that at least donors have some sense of how our decisions are made internally.
大意：
由于显卡（N卡）的架构、普及范围以及对cuda编程语言的实现程度不同，有时我们不得不放弃对一些老旧显卡的支持。因为我们的编程人员也不足，维护那么多内核需要极大的人力资源开销。
所以我们内部做了个艰难的决定，今后逐步放弃对N卡中非fermi卡的支持。请大家见谅。当然这个放弃要根据科研的需要逐步进行，不会突然取消的。

vmzy

MARCH 19, 2013
FAHBench 1.0

We’ve released FAHBench 1.0, with a new slick GUI that should make it much more accessible to new comers. Click on the FAHBench link above or the image below to try it out! Don’t worry, it maintains backwards compatibility with the old command line interface.

More info at http://fahbench.com/


大意：
发布FAHBench 1.0。对界面做了改动，现在更人性化了。

vmzy

MARCH 25, 2013
Peptoid structure prediction

Guest post from Dr. Greg Bowman, UC Berkeley

Prof. Vince Voelz’s lab has published an exciting paper on their recent successes with predicting the structures of protein-like molecules called peptoids (here).  Peptoids are similar to proteins but with a rearrangement in their chemistry (see example below).  Their similarity to proteins allows peptoids to function like proteins.  However, the alteration in peptoid chemistry relative to proteins effectively makes them invisible to parts of the immune system designed to recognize foreign proteins.  Therefore, peptoids are an attractive option for drug design.  To fully realize this potential, we need to be able to predict the structures of peptoids and design them to perform specific functions.  The Voelz lab’s work demonstrates that computer simulations can provide this sort of information by presented predicted structures of a number of peptoids along with experimental structures confirming the accuracy of their predictions (see example below).



Peptide vs. peptoid chemistry.  In peptoids, a group of atoms (called R) is moved from a carbon to an adjacent nitrogen (N).

  

An example of one of the Voelz lab's predicted structures (in green) overlaid with the experimental structure (in white).
大意：
类胨结构预测。
类胨的结构和蛋白质类似，但是他们可以绕过免疫系统，对某些细胞实现精确打击。这个特性使它们有潜力成了新药。为此我们对它进行了计算机模拟研究并和实验数据进行了比对。

金鹏

MARCH 31, 2013
Sneak peak at OpenMM 5.1: about 2x increase in PPD for GPU core 17

We have been aggressively working on OpenMM (the key code used in the FAH GPU cores), creating new algorithms to increase performance on NVIDIA and AMD GPUs.  The results have been pretty exciting.  With OpenMM 5.1 (vs OpenMM 5.0, used in the current core 17 release), we are getting about a 2x speed up on typical FAH WU calculations, which will lead to an automatic 2x increase in PPD once this software is out of beta testing and integrated into core 17.  

There's a lot of testing to do and it's very possible that these numbers will change, but the results were so exciting that I wanted to give donors a heads up.  Here's some numbers that we're seeing:

OpenCL running on the GTX 680: The first 2 columns are nanoseconds per day (i.e. how much science gets done in a GPU day) and the 3rd column is the speedup of 5.1 over 5.0.

Type of Calculation	OpenMM 5.0	OpenMM 5.1	Speedup
Implicit hbonds	92	134	1.46
Implicit hangles	153	209	1.36
RF hbonds	31.4	78.1	2.49
RF hangles	58	113.0	1.95
PME hbonds	19.6	41.5	2.12
PME hangles	37.3	66.9	1.79

OpenCL running on a Radeon HD 7970:  The first 2 columns are nanoseconds per day (i.e. how much science gets done in a GPU day) and the 3rd column is the speedup of 5.1 over 5.0.

Type of Calculation	OpenMM 5.0	OpenMM 5.1	Speedup
Implicit hbonds	87	120	1.38
Implicit hangles	96	104	1.09
RF hbonds	33.5	83.5	2.49
RF hangles	51.8	90.2	1.74
PME hbonds	21.8	49.3	2.26
PME hangles	34.6	63.0	1.82

Note that "PME hbonds" is likely the most common calculation that we plan to run in the near term with core 17.   We're very excited about the way this is shaping up and think that donors would be curious to know where this is going.  

大意：
OpenMM 5.1 性能剧透
当前gpu的主要工作在编写新算法提高opencl的运算效率。目前已经取得了突破性进展。平均提速达2倍以上。具体数据见表（注：当前core17主要进行的是 PME hbonds 运算）。
待内侧完成，开始公测新版core17内核时，你会发现gpu的ppd瞬间核爆的。

vmzy

APRIL 02, 2013
Would you like to learn more about the Folding@home community and your contribution to it?

Today, we have a guest blog post by  Vickie Curtis, a research student in the UK's Centre for Research in Education and Educational Technology.  She's working with the Folding@home team to glean more feedback from donors.

Would you like to learn more about the Folding@home community and your contribution to it?  I am a doctoral student at the Institute for Educational Technology at the Open University in the UK.  I am looking at how digital technologies are changing the way scientists interact with members  of the wider public, and I am particularly interested in online 'citizen science' projects such as  Folding@home.

Folding@home is one of the longest-running and most successful online citizen science projects, and it would be great to know a little more about why people contribute to the Folding@home community, their views about the project, and about these types of project in general.  I have prepared an online survey for participants, which should take about 10-15 minutes to complete.  The feedback will be shared with the Folding@home team and may help them to make improvements to the project.  I will also share the findings with you via the website and blog.

All the information you supply will be kept on a secure server and not passed to any third parties.  If you would like to take part, please follow the link below.

http:/www.survey.bris.ac.uk/open/foldingathome
大意：
Vickie Curtis是英国教育研究中心的博士生。她想利用FAH研究下数字技术是如何改变科学家与民众之间的交互的。请志愿者们花10-15分钟填份调查表吧。

金鹏

by kasson » Wed Apr 03, 2013 11:57 pm

We are pleased to announce the publication of some of our recent folding results. Per Larsson in my lab is the lead author on this publication, and it details how influenza proteins interact with membranes (and change shape i.e. refold in the membranes). We also look at how some important mutations change this interaction and what this might tell us about critical influenza protein function.

This work involved a series of F@H projects, notably those in the 6050-6099 range and in the 71XX range. We are also continuing this investigation and hope to have more results in the future.

Thanks for your support of Folding@Home!

Article freely available at: http://www.ploscompbiol.org/article/info%3Adoi%2F10.1371%2Fjournal.pcbi.1002950

Summary reproduced from the article:
Membrane fusion is a common process critical to both cellular function and infection by enveloped viruses. Influenza is a particularly useful model system for studying fusion because the fusion reaction is accomplished by a single protein, hemagglutinin. Furthermore, mutations to the membrane-inserted portion of hemagglutinin have been identified that do not detectably alter the rest of the protein but can either arrest fusion halfway or block it entirely. For influenza at least, it seems that the membrane-inserted hemagglutinin peptide plays a critical role in promoting fusion, perhaps by increasing the local disorder of lipid bilayers. However, we lack a mechanistic understanding sufficient to predict the activity of fusion peptide mutants from their sequence. Here, we have used lipid tail protrusion as a way to measure how much fusion peptides disorder their surrounding bilayer; we see a strong relationship between lipid tail protrusion and the ability of fusion peptide mutants to promote lipid mixing between membranes. Our simulations also predict that this lipid tail protrusion is much more common when the peptides adopt a kinked helix structure than when they are straight or hairpin-like. We therefore hypothesize that, while all three types of structure likely undergo conformational exchange, the kinked helix structure is most active in promoting fusion.
大意：
发表有关膜融合研究结果的论文。主要涉及任务：6050-6099和71XX。
论文摘要：
在细胞运转和病毒感染过程中膜融合是很常见的一个工序。之所以挑选流感病毒来进行研究，是因为它只有一种蛋白质——红血球凝集素，可以进行膜融合。（译注：剩下的太难翻译了，我放弃）


貌似出成果了，鸟文看不懂，烦请V版翻译

vmzy

APRIL 22, 2013
Folding@home survey still open –– please let us know how we're doing

Here's a guest post from Vickie Curtis, a Research Student at UK's Centre for Research in Education and Educational Technology.

I am a doctoral student at the Institute for Educational Technology at the Open University in the UK.  I am looking at how digital technologies are changing the way scientists interact with members of the wider public, and I am particularly interested in online 'citizen science' projects such as Folding@home.

A few weeks ago we launched an online survey to learn a little more about why people contribute to the Folding@home community, their views about the project, and about ‘citizen science’ projects in general.  We’ve had a great response so far, but would like to keep the survey open for a couple more weeks so that we can capture the views of participants who haven’t yet had a chance to take part (we would love to hear from more women who contribute to Folding@home).

The survey should take about 10 minutes, and the feedback will eventually be shared with you via the website and blog.  All the information you supply will be kept on a secure server and not passed to any third parties.  If you would like to take part, please follow the link below.

Many thanks to those who have already contributed!

http://www.survey.bris.ac.uk/open/foldingathome
大意：
上次的问卷调查准备延期几周，希望大家（尤其是女性）踊跃参加。

金鹏

http://www.equn.com/forum/forum.php?mod=redirect&goto=findpost&ptid=37316&pid=499582

proteneer


大家好，我是雨桐. 我是FAH的GPU核心开发人员。今天我想给大家介绍一下我们最近几个月的工作进展.

GPU核心开发有三个主要模块：
首先，OpenMM是一个分子动力模拟的开源库，FAHBench和FAH都采用了这个库。其次，FAHBenching是我们的GPU标准测量程序，它支持所有于OpenCL兼容的硬件设备。让我们高兴的是，FAHBench最近刚被加入了Anandtech的GPU测试包。核心17是大家所熟悉的，所有FAH客户端用来做科学计算的核心程序。图上这些箭头是想说明，我们所有的开发模块都是相互联系的。

现在让我们来看看核心17的开发过程。

去年10月，我们开始构思核心17。我们想达到三个主要的目标：新的核心应该更快，更稳定，并能够支持隐式溶剂模型（Implicit solvent）以外更多的模拟计算。由于核心15、16已经太过冗杂，我们决定重零开始构建核心17。

11月我们开始重写核心15,16中已有的核心计算功能。到1月，我们的计算服务器，分配服务器，和客户端都已经完成修改，支持核心17。同时我们首次组建了一个内部测试团队，这个团队采用freenode上的IRC聊天室，给我们提供实时的测试反馈。

今年2月，核心17在1000多个GPU上发布了beta版本。从中我们获得了宝贵的反馈。这个版本的核心在NVIDIA的设备上并没有获得很好的加速，在AMD上还不错；这个版本经常崩溃；还存在一些bug。于是我们又开始重新设计改进我们的核心。

今年4月，我们对OpenMM(库)增加了很多优化和纠错，第一次开发了linux版本的GPU核心，同时我们的测试团队已经成长到30多人。

终于，我们得到了现在的核心17。

我们现在能支持更多的模拟，从隐式溶剂模型到分子数超过10万以上的大型系统。核心17增加了系统稳定性，代码扩展性更好，并第一次提供了linux系统支持，并且速度变得非常快。你们一定会好奇，核心17到底有多么快？让我们来看一组数据：
新的核心在GTX Titan上从每天5万点加速到每天12万点
在GTX 680上，从每天3万点加速到8万点
在AMD7970上，从每天1万加速到11万点
在AMD7870上，则从每天5千加速到5万点

我们并没有怠惰，我们已经开始着手设计支持Intel设备，比如i7，集成显卡，和Xeon Phi。我们计划逐步增加FAH上的计算项目，让我们组里的研究人员可以添加更多有趣的研究。当然，增加计算速度还是我们不变的需求。

回到最初这幅图，我想谈谈你们如何能参与帮助FAH。
如果你是一个程序员，欢迎你参加开源的OpenMM扩展开发，OpenMM这个月底会发布到github上；
如果你是热情爱好者，希望你能使用FAHBen测试新型硬件，并且加入我们在freenode上的内部测试团队；
如果你是捐赠者，希望你能多宣传FAH，让更多人，更多机器，参与进来。

结束前，我想感谢一些人。右边是我们的内部测试团队，他们积极的提供测试反馈。没有他们，我们无法这么快完成开发。左边是pande组里的成员，Joseph和Peter和我一样，是核心开发人员。Diwakar和TJ帮助我们建立了很多计算项目。Christian和Robert总是支持并提供有用的反馈。

最后一件事，这周我会在reddit上做一个公开问答，地址是reddit.com/r/folding.
欢迎大家来看看，和我们聊聊天。

关于FAH Core 17更新的一些介绍，有中文字幕。欢迎大家去YouTube上看。 http://www.youtube.com/watch?v=rXK6g6TDqxg

vmzy

MAY 08, 2013
Stats server is down and undergoing maintenance

A key FAH server is down right now and stats updates have been suspended until it is back up.  As always, stats are kept on the Work Servers (WS's) so even if an update hasn't been run, the points are being accumulated as WUs come in, so it's only an issue of updating the database for donors to see.

We don't have an ETA on this right now, but our team is working on it.
大意：
统计服务器挂了。目前攻城师正在努力修复中。统计页面暂停更新，不过数据库里的积分不会少的，请大家放心。

金鹏

May 24, 2013
Stats system speed update We've been working to streamline the stats system update to minimize downtime for donors.  We now are able to update stats without taking the web pages off line, so stats updates will continue every hour, but the stats web pages on our site will continue to be available.

Posted at 09:10 AM | Permalink

大意：
优化统计更新流程。现在统计可以每小时更新一次，而且统计时不会影响统计页面的访问。

wpf999

金鹏 发表于 2013-5-26 10:46
May 24, 2013
Stats system speed update We've been working to streamline the stats system update to m ...

以前也是每小时更新一次，但是更新期间网页无法访问。现在改进了，在更新数据期间，网页也可访问

金鹏

JUNE 29, 2013

Welcome to FahCore 17!

We are proud to announce that our latest GPU core, FahCore 17, was recently moved from beta to advanced testing, the last quality assurance step before a full release. As we previously mentioned, this core is a significant step for us. FahCore 17 is a complete overhaul from our previous GPU cores. It brings a cleaner and more streamlined codebase, new serialization mechanisms that allow us to set up diverse simulations, and improved stability. Its use of OpenCL has united our development, allowing the single core to run on both Nvidia and AMD cards, and theoretically any OpenCL-capable device. It is also our first GPU core to run natively in Linux, although we are only supporting Nvidia GPUs there for the time being as we wait for AMD's Linux drivers to mature a bit more. Overall, this core sets a strong foundation for the future of GPU core development.

On AMD cards, FahCore 17 is about 10 times faster than the old GPU cores, and on Nvidia it's about twice as fast. This is mainly due to its OpenMM 5.1 base, which contains many optimizations which deliver a significant speedup. One optimization in particular that we are waiting for is CUDA JIT, a just-in-time compiler that Nvidia may be introducing into its drivers in the coming future. Not only will this technology allow us to offer support for the CUDA platform with FahCore 17, but the JIT compiler is likely to deliver a massive speedup. For the time being, we continue to work at finding additional optimizations on our end. We have also successfully tested FahCore 17 with extremely large proteins (500,000+ atoms), which are on par with the ones used by "bigadv" CPU projects.

To run FahCore 17, you need a Fermi GPU or better and Windows or Linux, or a AMD HD5000 or better and Windows. It also currently requires proprietary drivers from these vendors. You can test FahCore 17 by adding the "client-type = advanced" setting into the extra core options in the V7 client, as in the Configuration FAQ. Another excellent resource is the GPU FAQ which describes why GPUs are so helpful to us.

We'd like to thank all the alpha testers on FreeNode's #fah IRC channel, as well as the beta testers on foldingforum.org, who have all helped us bring the core to this point!

Posted at 12:40 PM | Permalink

大意：
FahCore 17由小公测（beta）转至大公测（advanced）。之前说过Core 17是一次质的飞跃。它不仅更快、更强、更合理，而且将以前A、N两个内核合并为一个。理论上支持所有支持OpenCL的硬件。并且N卡可以直接上linux跑了（A卡由于官方驱动不支持，没法跑）。
Core 17在A卡上提速到10倍，N卡上提速到2倍。Core 17的下一个大更新就是CUDA JIT功能，不但能让Core 17原生支持CUDA，而且能带来速度的大提升。目前我们的工作是细化代码，并且对超大蛋白质成功进行了测试（之前只能在CPU的bigadv上跑）。
FahCore 17，目前支持Windows或Linux平台的Fermi及以上级N卡，Windows平台的HD5000及以上级A卡。当然别忘了装最新的显卡驱动。只要安装V7客户端并加入client-type = advanced参数，你就可以跑FahCore 17了。

caozhonghua

金鹏 发表于 2013-6-30 10:15
JUNE 29, 2013 Welcome to FahCore 17!We are proud to announce that our latest GPU core, FahCore 17, w ...

参数需要改了？

vmzy

JULY 08, 2013
New FAH web site

We've been working behind the scenes on a revamp of our web site.  It went live today (http://folding.stanford.edu/home).  This is part of our larger plan to make FAH more friendly and easy to use, especially to non-experts.  

With that said, we're now thinking about next steps to make FAH more fun and appealing to experts, such as computer enthusiasts and gamers.  We're in the early stages of deciding what would be useful there.  If you have ideas, please do give us some feedback on our forum at this thread:

http://foldingforum.org/viewtopic.php?f=16&t=24532
大意：
网站改版，新网站地址为：http://folding.stanford.edu/home
我们想让网站更人性化，对电脑发烧友及玩家更有吸引力，如果你有什么好的建议，欢迎到官方论坛发帖告知，谢谢。

vmzy

JULY 11, 2013
A peek into Core 17 benchmarking

Our primary goal with benchmarking is "equal points for equal work."  However, making this process consistent over lots of different types of WUs and different types of hardware is tricky.  We had an internal discussion about the PPD for two projects (7810 and 8900) recently and we thought donors might find these details interesting.  

We were working to rebalance the points to make the PPD consistent, but just doing that over the wide range of hardware is difficult.  Check out the graph below which shows the PPD on the y-axis and donor GPUs sorted along the x-axis by typical PPD.  The dark line shows averages and the gray area shows error bars (variation between WUs for a given project on the same GPU type).

What we see is that our protocol balanced the PPD on the low end, but on the high there is both bigger variation (more shaded areas) and also bigger differences on the very highest power GPUs.  In these situations, we usually go with our protocols, but this time, given all the analysis we did on it, I thought it would be interesting for donors to see these sorts of details.

It's these sorts of variations which leads to PPD fluctuations, so perhaps the main lesson here is that even with our protocols and plans, it's really hard to be consistent over all the different hardware, even when we're talking about just GPUs and just 2 projects.

大意：
Core 17积分那些事儿
关于积分我们的首要目标是实现“同工同酬”，为此我们研究了下7810 和 8900任务的得分，如图所示。
不同的设备，得分波动实在太大了，要想平衡积分那是相当困难的啊，不过这个可以有，我们会加紧研究现有数据，争取能尽量平衡积分系统。