找回密码
 新注册用户
搜索
楼主: vmzy

[独立平台] [生命科学类] Folding@Home

[复制链接]
发表于 2008-5-27 23:51:22 | 显示全部楼层
原帖由 shouldbe 于 2008-5-27 21:03 发表
当初买8800GTS还是对的。至少魔兽从来没有让我失望。这回不魔兽的时候就FAH。希望PPD能够重新回到4000以上。不过电费会很恐怖。我住到新家刚半年电表度数已经是住进来快2年的邻居的2倍了。夏天一来,再开空调就更恐怖 ...

8800GTS配备的外排式散热器就足以说明它能顶住
回复

使用道具 举报

发表于 2008-5-28 10:50:33 | 显示全部楼层
G92的88GTS发热不高的
只是原装散热器为了追求静音,导致在待机模式下风扇转速设定的过于保守。对于我这种追求静音的人来说当然是好事。
如果要压温度的话,推荐用软件手动吧风扇转速调制50-60%
回复

使用道具 举报

 楼主| 发表于 2008-5-28 12:22:46 | 显示全部楼层
May 27, 2008
More info about the GPU1 to GPU2 transition

There's been several questions regarding the GPU1 client and why we decided to shut it down.  I hope I can shed some light here at least on why we're doing what we're doing, such that even if people disagree with our decisions, they can at least see where we're coming from.  

Some people have asked "why shutdown the client if it's working?" The bottom line here is that the GPU1 results are no longer scientifically useful. It's pretty clear now that DirectX (DX) is not sufficiently reliable for scientific calculations. This was not known before (and some people wouldn't believe this until we proved it). With the GPU1 results, we can now show what the limitations are pretty decisively.

GPU1 also did help us a lot in terms of developing its successor and what's needed to run GPU's in a distributed computing fashion.  The good news here is that GPU2 is behaving very well, on both ATI and NVIDIA hardware, and this is a direct result of what we've learned with GPU1 WU's.  In the end, however, GPU1 will not be able to help us understand protein misfolding, Alzheimer's Disease, etc due to this unresolvable limitations. We could keep GPU1 live just crunching away in its current form, but that would be wasting people's electricity at this point, as we've learned everything we can learn from those cards can do.

In the past, we had a somewhat similar shutdown situation, i.e. when QMD core projects stopped. In that case, donors were left hanging since we didn't give any warning for stopping QMD projects. We did try (perhaps unsuccessfully) to handle the GPU1 situation better than QMD. In QMD, we stopped needing that core and so we stopped the calculation without warning, not realizing the impact that would cause. With GPU1, we gave a several month warning (indeed, note that GPU1 is still actively running, so all of this is information in advance to shutting down GPU1). We tried to avoid the QMD situation by giving advance warning, but it looks like donors would like even more advance warning. However, there's limits to how much in advance we know the situation ourselves.

Indeed, the knowledge that it made sense to end GPU1 came reasonably recently to us.  We have been working on CAL for a while and it seemed that CAL might be a solution, but we only knew until we got some testing "in the wild."  DirectX (DX -- what GPU1 is based on) works much better in the lab than in the wild, and it was possible that CAL behaved that way too.  After seeing that CAL behaved well in the wild, it became clear that the GPU1 path was obsolete.  However, this is a relatively recent finding and we made the announcement about the situation relatively shortly thereafter.

It was a tough decision. Some suggested we just leave GPU1 running, even though people's electricity really would be going to waste, other than generating points. I didn't think that was a good idea. We did know it would be a tough PR hit, but when people talk about the history of FAH, I want to make it clear that we're here to address AD and other diseases, not just running calculations for the sake of points and nothing more (which has been the critique of some other distributed computing projects).

So, what's the right thing to do?  I guess it comes to this: would GPU1 donors be happier if we just keep GPU1 servers running, doing with no scientific value for points? We could do that, at a cost of taking away personnel from improving existing clients, keeping existing servers going, etc for the sake of keeping GPU1 running. However, that's not what FAH is for and I think it's important that FAH not devolve into a big points game, losing sight of why we're doing what we're doing.

PS Some further discussion can be found here .

大意:
因为GPU1的DirectX底层无法提供对高精度科学计算的支持,所以我们要关闭它。我们是为了科学而计算,不是为了计算而计算,希望大家理解。

评分

参与人数 1基本分 +10 维基拼图 +5 收起 理由
BiscuiT + 10 + 5

查看全部评分

回复

使用道具 举报

发表于 2008-5-29 11:23:08 | 显示全部楼层
这好像就是回答我的问题的http://www.equn.com/forum/thread-18161-1-2.html

看来已经有人在FAH官方论坛上发表看法了。不过我觉得Pande最CD的地方在于让某些donor花了大量的money买了好显卡,结果却发现原来大家都犯了一个大错误——算出来的东西由于精度不够根本不能用!于是只好放弃。这个玩笑真的开大发了。不要哪天突然他又蹦出来说,我们的XXX包由于设计错误,算出来的结果不能用。或者ATI/NVIDIA的client设计不完善,算出来的结果不能用。那大伙辛辛苦苦的算这些东西干啥呢?那些电费就这样打了水漂了?我对Pande小组的研究态度是否严谨打了一个大问号。我觉得既然有这样的前车之鉴,Pande小组就不应该再仓促的发布新的Client。或者至少要声明新的Client的结果的实用性有多少,是试验性质的,还是实用性的。这样对大家有个交待。不要又象GPU1那样马后炮。这很伤donor的心。要知道很多人坚持计算出了PPD以外,还有一个“正当”的理由,就是在为科学做贡献。要不花那么多电费做什么?
回复

使用道具 举报

发表于 2008-5-29 13:02:24 | 显示全部楼层

回复 #139 shouldbe 的帖子

研究的对象就是未知的事情。
对于未知的事情必然有不确定性,既然发现原先的方法/硬件有缺点有局限就应该予以更新或者淘汰,抱着不放会影响研究进程。

你之前提到的问题已经回复过了,FAH官方没有要求donor去花大量的money买好显卡。
既然你愿意为他们研究贡献自己的力量/财力,那么是不是应该同时有面对研究失败的良好心态呢?
回复

使用道具 举报

 楼主| 发表于 2008-5-29 15:03:13 | 显示全部楼层
GPU1一直都是Beta版,测试版,本来就是为了测试用的,测试不通过就取消,没什么问题啊!
回复

使用道具 举报

 楼主| 发表于 2008-5-30 09:57:55 | 显示全部楼层
May 29, 2008
New FAH map: FAH donors in May 2008

We've been looking into ways to map FAH donors.  Here's an update from May 2008.  We're looking for ways to automate this process.  Updates are kept on our web site map page.
大意:
更新了,贡献者全球分布图。

注:一看图就知道为什么,中国队,飞流直下三千尺了。

GPU2 points & benchmark machine (and points in general)

I've talked about this in the Folding Community Forum (FCF), but I thought it's worth reminding people (especially NVIDIA-based folders new to the project).  As the GPU2 code matures, we may need to change the points per day (PPD) and/or the nature of the benchmark machine.  The PPD for the bechmark machine may go up or may go down depending on what change we decide in the end.  The main problem is that both the NVIDIA and ATI code is undergoing optimizations, which means PPD will change in complex ways.  For example, if the NVIDIA code doesn't change, but the ATI code gets optimized and our benchmark is based on an ATI card, then NVIDIA PPD will go down, even if the NVIDIA code doesn't change.  

Also, how efficient these cards are depends on the protein we run on a given card.  To address this issue, we will likely switch the benchmark machine to run a card whose PPD does not change so variably with protein, although that may be hard to find (we'll see).

Anyway, just a heads up -- the PPD we see today is not necessarily what you'll see later.  Your PPD may increase or decrease depending on many complex factors (code optimization, benchmark machine, protein simulated, etc), and I expect there may be more points variability with GPUs than with other platforms (isn't life on the bleeding edge fun).  However, due to the great performance of the GPU's right now (which will likely only get better), the GPU PPD would be significantly greater than a typical CPU client.

I should also remind donors that we've been talking about rebalancing points in general, to try to get SMP, GPU, PS3, and classic points in line with the scientific productivity done by the clients.  As long as more points = more science, donors can optimize for points and that will automatically optimize for maximum scientific output.
大意:
我们在GPU正式发布会调整积分策略,以使得分平衡。

[ 本帖最后由 vmzy 于 2008-5-30 14:39 编辑 ]

评分

参与人数 1基本分 +20 维基拼图 +5 收起 理由
BiscuiT + 20 + 5

查看全部评分

回复

使用道具 举报

发表于 2008-5-31 13:22:07 | 显示全部楼层
我想到科学研究的不确定性导致研究失败的问题了。但还是不平.......

看了地图只能让我感到科学无国界,而科学成果是国界的。donor的数量也许从侧面反应了这个问题——直接受益者更愿意付出。
回复

使用道具 举报

 楼主| 发表于 2008-6-2 11:32:26 | 显示全部楼层
May 31, 2008
ATI GPU2 core issue

Due to a bug in ATI GPU2 core version  1.01, the core has a bug in how it stores trr files.  This makes the files too large, leading to corruption, leading to no points server side.  We released a bug fix core yesterday (version 1.02) and it looks like that fixes the problem.

We are doing an automatic core upgrade today (Saturday) at 9:30am pacific time.  If you want to fix this before then, you could manually delete your core.  The client will then automatically download a new one (v 1.02).

However, if you see that the WU is still in the queue, you will need to delete the WU, either by using qfix or deleting the work directory (which should only be done in between WU's, not while one is running).
大意:
ATI GPU2 1.01版计算内核有BUG,会造成结果文件异常,导致白算。
我们已经发布了1.02版新内核,如果你想手动升级,请把计算内核程序手动删除。FAH会自动下载新版内核程序。

评分

参与人数 1基本分 +20 维基拼图 +8 收起 理由
BiscuiT + 20 + 8

查看全部评分

回复

使用道具 举报

发表于 2008-6-2 13:41:48 | 显示全部楼层
NVIDIA的GPU2客户端什么时候出来啊?
回复

使用道具 举报

 楼主| 发表于 2008-6-2 21:47:46 | 显示全部楼层
官方正式开了NVIDIA GPU2的子论坛,估计不久就要开始公测了,请大家拭目以待。
子论坛地址:http://foldingforum.org/viewforum.php?f=43
回复

使用道具 举报

发表于 2008-6-2 23:04:34 | 显示全部楼层
NVIDIA GPU2的客户端不知道要不要专门针对CUDA的驱动。。
回复

使用道具 举报

发表于 2008-6-3 11:27:13 | 显示全部楼层
貌似现在驱动中就包含CUDA
回复

使用道具 举报

 楼主| 发表于 2008-6-5 11:03:43 | 显示全部楼层
June 04, 2008
GPU1 retirement: Friday June 6 at 9am pacific time

As previously discussed, we are retiring GPU1.  We earlier discussed possibly doing this on June 2.  We've pushed this back a little, but will do so on Friday June 6 at 9am pacific time.  We'd like to take this opportunity to thank all the GPU1 donors who have pushed this technology very far, leading to the next generation GPU2 clients we have developed now.
大意:
原定于6月2日关闭GPU1,我们现在决定稍微延后一下,将于6月6日关闭。感谢大家曾经做出的贡献。

评分

参与人数 1基本分 +10 维基拼图 +5 收起 理由
BiscuiT + 10 + 5

查看全部评分

回复

使用道具 举报

发表于 2008-6-8 19:34:19 | 显示全部楼层
期待N卡客户端……
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 新注册用户

本版积分规则

论坛官方淘宝店开业啦~

Archiver|手机版|小黑屋|中国分布式计算总站 ( 沪ICP备05042587号 )

GMT+8, 2024-6-13 09:38

Powered by Discuz! X3.5

© 2001-2024 Discuz! Team.

快速回复 返回顶部 返回列表