[独立平台] [生命科学类] Folding@Home

vmzy

AUGUST 21, 2013
Planned server room maintenance: Monday, August 26

We will be moving many key Folding@home servers on Monday August 26.  While much of FAH will be up, certain key systems will be down for a few hours starting in the morning (pacific time) of Monday August 26.  We'll give more updates as we move along on Monday.
大意：
周一早上，几台关键服务器搬家，届时部分服务器会下线几个小时。

vmzy

AUGUST 23, 2013
Reminder about aging cores (eg Core11, Core78)

One aspect which has dominated FAH for a decade is the continual push for new scientific approaches.  This manifests itself in terms of new scientific cores.  For example, the new GPU core (Core17) has brought huge speed improvements (especially to AMD GPUs) and involved a complete rewrite.

A negative consequence of this continuous push for improvement and progress is that newer cores often are restricted to more advanced hardware.  To help utilize as much power as is available to FAH, we tend to continue projects with older cores, but eventually, the science they can do becomes too limiting, and we must retire them.

This is the eventual fate for all cores, but is most certainly an issue sooner for certain cores, especially cores11 and 78.  While we don't have any specific end dates for either, we'd like to remind donors that those cores are reaching "end of life" status, and when they are retired, certain older hardware (eg CPUs that don't support SSE2 or older GPUs) won't be supported by FAH.  

The bottom line is that we're working to delay that as long as possible, but this post is a heads up that our support of those cores won't last a lot longer.  If I had to guess, I'd say probably within a year or so they would be retired, maybe 2 years if the existing projects need additional data.  

As always, we'll give donors more information as we know it and try to give a more specific end date when we know it.
大意：
致我们终将逝去的core
老的计算内核计算效率低，而且计算速度慢（老cpu专用core），有些时候严重拖慢项目进度。所以我们决定逐渐放弃对老core（例如cores11、78）、老cpu（不支持sse2）、老显卡的支持。
当然放弃的日期要看项目需要，也许1年、也许2年。让我们为这些曾经为FAH立下过汗马功劳的core致敬。

金鹏

AUGUST 27, 2013
UPDATE 4: Planned server room maintenance: Monday, August 26
Looks like everything is up, except for a single server (VSP07) and its VM's associated with it.  This server is serving Core11 GPU clients, so those are off line at the moment.  Our sysadmins are working on this now.

Posted at 10:12 AM | Permalink

AUGUST 26, 2013
UPDATE 3: Planned server room maintenance: Monday, August 26
The networking seems to be up, but there are a few issues.  They've got something basic going now and will resolve the remaining issues in the morning.  We are running a stats update right now.

Posted at 06:32 PM | Permalink

UPDATE 2: Planned server room maintenance: Monday, August 26
Looks like the servers are up and healthy, but there is an issue with the networking which central IT is working on.  For now, we are in a holding pattern until the networking has been resolved.

Posted at 05:44 PM | Permalink

AUGUST 26, 2013
UPDATE: Planned server room maintenance: Monday, August 26
The maintenance is moving along, but certain key machines are down during the move, most notably the main AS, GPU AS, and some key stats systems.  We expect the maintenance to be completed in an hour or two.

Posted at 12:50 PM | Permalink

vmzy

Adaptive Sampling Movie
September 5, 2013 by Vijay Pande
Matthew Harrigan and Robert McGibbon in the Pande Group have recently made a movie which highlights how MSM Adaptive sampling works in a simple yet relevant example (a 2 dimensional Muller potential).   Adaptive sampling works by first running several relatively short trajectories, then building an MSM model, then using the MSM to run new trajectories.

In this movie, we compare the adaptive sampling approach to the more traditional approach of a single, long trajectory, and find that long trajectories can get stuck in minima and the adaptive sampling approach can quickly explore the whole space.
http://www.youtube.com/embed/oh-xqR79QM0

大意：
有关Adaptive Sampling算法的介绍视频
相对于传统方法，新算法更适合长轨迹运算，而传统算法很容易算死。（译注：让大包来的更猛烈些吧）

vmzy

Introducing the Chodera lab at the Memorial Sloan-Kettering Cancer Center
September 14, 2013 by john Chodera ·
The Chodera laboratory at the Memorial Sloan-Kettering Cancer Center (MSKCC) has just received our new Folding@Home server hardware, allowing us to tap into the 10+ PFLOP/s of computing resources generously provided by the F@H Donor Community.  We’re excited to join this vibrant community and bring some F@H resources to bear on various problems in cancer we’re currently working on.

Our mission

Our mission is to completely redesign the way that therapeutics—especially anticancer drugs—are designed using computers, graphics processors (GPUs), distributed computing, robots, and whatever technology we can get our hands on.  We are striving to make the design of new cancer drugs much more of an engineering science, where state-of-the-art computer models quantitatively and accurately predict many aspects of drug behavior before they are synthesized.  We certainly won’t get there overnight—lots of hard work is needed to improve our algorithms, forcefields, and theory.   But by tapping into the enormous computing resources of F@H, we can more rapidly make predictions and then test them in our laboratory (with robots!) to quickly make improvements through learning from each cycle of prediction and validation.

Our server hardware

We are installing some heavy-duty servers to serve up F@H work units mainly to GPU users.  Our main server is a Dell R820 with 4x Intel Xeon E5-4620 2.2GHz processors and 256GB of RAM, which will also allow us to analyze incoming data “on the fly” to adaptively and intelligently make use of F@H Donor computing resources.  Our backup server is a Dell R620 with 2x Intel Xeon E5-2690 2.9GHz processors, also with 256GB of RAM.  These servers are both connected via SAS to a Dell MD3260 disk array containing 60x 4TB disks, providing 240TB of raw storage, or 192TB of usable storage with RAID6 to guard against data loss.

Expect more updates from us soon as we roll out our first big F@H projects.  We’ll tell you more about the science, the challenges, the technology, and our hope to transform the way cancer is treated and drugs are designed to treat important diseases.  For now, we’ve been busy moving into our new laboratory space and run our first robotic experiments!


Members of the Chodera lab conduct their first wetlab experiment in the Chodera lab at the Memorial Sloan-Kettering Cancer Center: isothermal titration calorimetry measurements of protein-ligand binding affinities.
大意：
介绍下Memorial Sloan-Kettering癌症中心的Chodera实验室
Memorial Sloan-Kettering癌症中心的Chodera实验室收到了FAH的服务器，开始接入FAH网络。利用FAH的庞大算力进行癌症研究。

目标：利用一切新技术创新癌症疗法（特别是研发抗癌新药）。利用FAH的算力和先进的算法和实验室的自动化测试设备，我们可以更快的进行药物研发迭代（FAH模拟完立即进行实验室验证，然后改进）。

服务器：主服务器是Dell R820，备用是Dell R620，磁盘阵列是Dell MD3260（240TB空间）

目前一切还在进行中，后面会不断更新进展，让FAH的第一个大型应用类项目来的更猛烈些吧。

vmzy

http://foldingforum.org down for service
September 17, 2013 by Vijay Pande ·
The http://foldingforum.org/ forum is currently down for servicing.  In particular, we are adding https support for more secure logins to the forum.  There will be a DNS change associated with the changes and DNS can take some time to update.  Until then, you can access the forum now with https (i.e. at https://foldingforum.org ).
大意：
官方论坛停机维护，连接加密，今后可以使用https协议访问论坛了。

vmzy

Core16 GPU core approaching end of life
September 28, 2013 by Vijay Pande ·
Our projects which use Core16 have completed and this means that Core16 is reaching its end of life, as the Core17 GPU core has some key features needed in new projects.  We do not have an ETA for the last Core16 WU, but it is likely soon, i.e. in a week or two.

One outstanding issue for Core17 is that while it has much greater performance for most AMD GPUs, it has been reported that PPD is slower for 5xxx class AMD GPUs.  We have been brainstorming fixes for this issue internally, but we have no solution for improving this unfortunately.  We are continuing to discuss this internally to see what we can do.
大意：
A卡Core16即将下线。
目前Core16的包基本上都处理完了，预计一两周内正式下线。
Core17在A卡上性能很强（不过5xxx上效率很低，但目前没有找到合适的解决方法）。

金鹏

October 10, 2013 by Vijay Pande ·
We were very excited to see that Michael Levitt is a co-recipient of the 2013 Nobel Prize in Chemistry.  Dr. Levitt has been a mentor of mine since I got to Stanford in 1999 and also has been an informal advisor for Folding@home over the years.  There’s a nice summary of his work in a recent Stanford News article:

http://news.stanford.edu/news/2013/october/michael-levitt-science-100913.html

In a nutshell, Dr. Levitt did pioneering work in using Molecular Dynamics to study proteins, a method which is at the heart of most of Folding@home’s calculations.

Filed Under: Uncategorized ·

金鹏

Planned server room shutdown October 22 at 6am pacific time

October 11, 2013 by Vijay Pande ·

One of our main server rooms is going through planned maintenance on Tuesday, October 22 starting at 6:00 AM pacific time.  The outage is expected to be less than one hour, so power should be restored by 7:00 AM. However, additional time will be required to bring all computer systems back online and doing a cold restart some times leads to new issues.  We’ll give more updates as we know them.

Filed Under: Uncategorized ·

大意：10月22日18时，机房例行维护。届时服务器预计会暂停1小时左右。

vmzy

Stats manually turned off in preparation for early Tuesday morning shutdown
October 22, 2013 by Vijay Pande ·
One of our primary server rooms will be shutdown for service tomorrow morning.  Most of FAH will be operational, but we’re taking the stats update off line now so there won’t be complications tomorrow morning.  WUs returned will still get credit, but the credit will not show up until the stats update is re-activated.
大意：
为迎接机房例维，暂停统计服务器。

Planned maintenance complete, server room back online
October 22, 2013 by Vijay Pande ·
Our planned maintenance is now complete and that server room is now back online.  The stats update has recommenced.
大意：
例维顺利完成，机房归位，统计服务器恢复正常。

vmzy

Changes in Core17: update and move to full fah
November 7, 2013 by Vijay Pande ·
With FAH GPU Core17 taking over the lion share of our calculations, with the older cores (especially core15 and core16) phasing out (as previously announced), we’re rolling out core 17 to full fah (not just adv).  This should hopefully resolve the issue that many GPU donors have been seeing in terms of not getting any work, now that Core15 and Core16 jobs are at end of life.

We’ve also released a new update to Core17, version 0.0.52, which is a minor update, with the following changes:

Core17 Release Notes:

-Much more responsive exiting

-Fixed bugs related to visualization for next version of the FAHViewer

-Disabled some internal debug code

-For Windows NVIDIA Users running on a single GPU, we’ve enabled a new feature called temperature control that pauses the core internally. This is disabled by default.

-To enable temperature control, a pair of additional arguments are required:

-tmax (maximum temperature threshold in Celsius), must be less than 110

-twait (minimum wait duration in seconds), must be at least 900

-if temperature > tmax, then the GPU waits for at least twait seconds

-these can be passed in via extra-core-args in the config.xml file :

<slot id=’0′ type=’GPU’>

<extra-core-args v=’-tmax=80 -twait=900′/>

</slot>

-Removed dependency of glib.
大意：
Core17转正（现在不用加adv参数了，同时Core15 和 Core16下线），并更新至0.0.52版。
主要更新：
单N卡，可以打开温度上限功能，达到指定温度停算，不过改功能默认关闭。如果要打开，参数为tmax（温度限制，不得高于110°），twait（暂停时间，最少900秒）

vmzy

Under the hood at Berkeley
November 13, 2013 by Greg Bowman ·
The Bowman lab at the University of California, Berkeley recently purchased its first Folding@home server and I thought I would share a little bit about what’s under the hood.  The main purpose of this machine is to serve and store Folding@home work units, so it is equipped with a dozen 4TB drives.  Two drives are dedicated to redundant storage (RAID-6 for experts who are curious to know), leaving us with 40 TB of storage space for collecting data.  Many of the projects I run take 10-100GB of disk space, so conservatively we’re now prepared to run 400 projects!  The machine also has four Gigabit Ethernet connections for sending/receiving data.  Another highlight is a dozen 2.0 GHz cores and 64GB of memory, allowing us to run some of our analysis locally.  The bulk of our analysis will still be conducted on clusters at Berkeley though, where we can parallelize many of the steps across hundreds of processors.
大意：
Berkeley的Bowman实验室入伙
加了一台强力服务器，40T硬盘（约支持400个子项目），强力cpu集群（可以用于本地结果分析）。

金鹏

by schwancr » Tue Dec 03, 2013 5:54 am

projects 7810 and 7811have been completed, so we will no longer be assigning more WUs.

Thanks for folding!

7810/11公测包目前圆满完成,木有包了

金鹏

New GPU Project 7627

by jadeshi » Fri Dec 13, 2013 9:38 am

This project is a continuation of Project 7625, a permutant of the protein CheY. It consists of a new set of starting structures in order to better sample the entire conformational landscape. Project description copied from original posting:

Projects 7624, 7625, 7626 - CheY Permutant Folding

These projects study the effect of chain topology on the folding of CheY. Circular permutants studied by the Matthews lab at U. Mass have revealed non-trivial dependence on connectivity of the backbone. These new projects are designed to tease out the physical origin of this dependence.

For nVidia Fermi GPUs (GPU3) on Windows machines
Points = 5187
timeout = 12.6
deadline = 17.9
atoms = 1980

These WUs will require a new version (v2.20) of Core15.

These projects aim to study the folding of CheY, which, if successful, will be the largest protein folded in silico. CheY is well-studied experimentally, and exhibits many interesting scientific features. Due to it's wide study in the lab, we hope that it will provide a test case for FAH methodology, and give us good ideas about how to improve our techniques. This is an exciting project, since it represents something that is really only possible on FAH combined with the power of GPUs.
==============

These WUs will require a new version (v2.20) of Core15.

搞错了吧?老版本都是2.25了,新版应该是2.30的CORE15?

vmzy

Changes to the bigadv threshold
December 17, 2013 by Peter Kasson ·
We have a policy of periodically re-evaluating the bigadv program, including the threshold required to run bigadv projects.
It is the intent of bigadv to match large and resource-intensive work units with some of the most powerful machines used by FAH donors.  This “most powerful” line naturally advances with computing power.  To date, bigadv has been a CPU-based program, and with increasing numbers of CPU cores and power of those cores, we have decided to lay out a roadmap of bigadv threshold changes for the next several months.
Feb 17 (two months from today):  bigadv threshold will become 24 cores
Apr 17 (four months from today):  bigadv threshold will become 32 cores
We want to give advance notice of these changes, and we recognize that change is not always welcome or comfortable.  We should also emphasize that the science performed by donor machines is valuable in all parts of the FAH project, and part of the change in bigadv threshold is because we would like to encourage moderately powerful machines to help boost the capabilities of non-bigadv SMP projects where we do a lot of this science.
We also recognize that core count is not the most robust metric of machine capability, but given our current infrastructure it is the most straightforward surrogate to evaluate.
Thank you once again for your generous participation in the Folding@Home project!
大意：
BA上调cpu数下限。计划2月17号，调成24T。4月17号调成32T。