[BOINC] [物理类] AQUA@home[已结束]

BiscuiT

项目主页：http://aqua.dwavesys.com/

AQUA (Adiabatic QUantum Algorithms) is a research project, based at D-Wave Systems, Inc., whose goal is to predict the performance of superconducting adiabatic quantum computers on a variety of hard problems arising in fields ranging from materials science to machine learning.  


About AQUA@home

D-Wave's AQUA (Adiabatic QUantum Algorithms) is a research project whose goal is to predict the performance of superconducting adiabatic quantum computers on a variety of hard problems arising in fields ranging from materials science to machine learning. AQUA@home uses Internet-connected computers to help design and analyze quantum computing algorithms, using Quantum Monte Carlo techniques. You can participate by running a free program on your computer.

AQUA@home is based at D-Wave Systems Inc., Burnaby, British Columbia, Canada.

绝热量子算法，目标是研究超导绝热量子计算机在材料和机械科学上的各种问题。
志愿者将帮助项目使用蒙特卡罗技术设计和分析量子计算算法。

Youth

呵呵，刚准备过来发帖，慢了一步：）

转一个官方论坛上相关的帖子过来吧：

有人提问：

Hi,
before I DL some executables to my computer I'd like to get some answers about this project:

- Are you trustworthy?
- What will happen with the results?
- What will be crunched at all?

主要是关于项目方的性质，研究结果的处理，计算内容。

项目方的回答1：

AQUA is developed by a number of scientists at D-Wave Systems Inc. (http://www.dwavesys.com). D-Wave is a research company that is building a Quantum Computer. We need to crunch a lot of data to simulate the behaviour of quantum systems (we are dealing with exponential complexity). That is why we decided to ask for help from the BOINC volunteer community.

All our results will be published in scientific journals or on-line, so everybody will share the results of the computations. This was one of the first decisions we made when thinking about starting AQUA@home.

The current application you are running computes the minimum gap between the best answer and the second best answer in an optimization problem. This information is needed so that we can know how long an adiabatic quantum computer should run to compute the best answer. The details are complicated, but a good place to start is Geordie Rose's blog: http://dwave.wordpress.com/.

-Kamran

项目是由D-Wave系统公司的一群科学家发起的，该公司主要是研究量子计算机的，因为有大量的数据要模拟，所以决定发起这个boinc项目。
项目所有的研究结果将发表在科学杂志上，这是我们发起这个项目时就做出的决定。
最后一段是研究内容。

项目方的回答2：

Hi Saenger,

Kamran answered most of your questions, here is some more info:

1. Yes! We have had good relations with a series of academic and business partners over about 10 years since the company was founded. If you go to arxiv.org and search on author=M.H.S. Amin and author=A.J. Berkley you will find most of our scientific publications. We are financed about 1/2 from governments (primarily Canada) and 1/2 from equity investors (primarily Draper Fisher Jurvetson and Goldman Sachs). We've raised about $70M in total. The code being run was co-developed by us and Dr. Peter Young from UCSC, an expert in monte carlo.

2. They will be published; the "punchline" is a plot of size of the problem being solved (number of variables in a spin glass problem) vs. the median runtime for our adiabatic quantum optimization algorithm.

3. The computation is of a particular correlation function that produces the energy gap between ground and first excited states in the quantum computer system being simulated. This energy gap can be used to calculate the runtime of the algorithm. You can see this paper for a model of what we are calculating and how: http://arxiv.org/abs/0803.3971 .

Hope this was useful & welcome to the project!

可以在arxiv.org上搜索到项目方发表的大部分论文，公司的经费一半来自政府（主要是加拿大），一半来自权益投资者（？），当前运行的代码是由项目方和UCSC的Peter Young博士开发的。
第二段和最后一段还是关于结果的处理和项目的计算内容。

Youth

目前的任务包：期限短，一天不到；内存占用小，2M左右；计算时间还行一个10分钟不到就收工，另一个26分钟36%。

BiscuiT

开始还以为是研究液体的。。缩写太让人捉摸不透。。（AQA部就好了吗。。

Youth

AQUA@home has seen a very healthy increase in the number of its volunteers! Everybody involved in this project wishes to thank you for your contributions, whether they be CPU cycles, or suggestions for making this a better site. We will continue to imrove all aspects of this project and we rely on you to continue to support it.

项目方已经看到志愿者的数量在稳步增加，感谢所有人在多方面做出的贡献，我们将继续改进我们的项目，希望大家能继续支持！

luoyinbazi

有一台计算机正在运行中，因为这个项目内存占用极小（2M）。

tokey213

是啊是啊 我的任务管理器里显示只占了292K内存

BiscuiT

http://bits.blogs.nytimes.com/20 ... f-quantum-computer/

December 22, 2008, 6:21 am

D-Wave Arms ‘Smoking Gun’ Proof of Quantum Computer

By Ashlee Vance

D-WaveSystems, a Canadian start-up, claims to have concocted a way of provingthat its quantum computer is actually a quantum computer. But to pulloff the proof, it needs thousands of people to volunteer spare time ontheir personal computers.


D-Wave chief technology officer Geordie Rose in front of the company’s latest quantum computer.


While founded in 1999, D-Wave remains a start-up because it has yetto ship a computer that customers can actually buy. Much of thecompany’s work is locked in the theoretical realm of quantum computing– long considered a sort of black arts backwater of the computing fieldfull of promise but thin on near-term practicality.


In a bid to prove that it has a computer capable of quantum effectsrather than just a regular computer, D-Wave has released some simulation software.The software centers on an algorithm designed to show off a quantummachine’s ability to solve complex problems more quickly than atraditional machine.


To crank through the simulation software, D-Wave is looking to spread the computing work across thousands of computers. It has embraced a type of distributed computing model similar to the SETI@home project where people download software that scans for signs of extraterrestrial intelligence as a background job on their PCs.

Once D-Wave collects the results of the simulations and processesthe information, it will compare the simulation against an actual runof its latest quantum computer, which should be completed in the nextfew months. If the figures match up, then D-Wave will have shown thatit really has a quantum machine, silencing skeptics, according toGeordie Rose, the company’s chief technology officer.

“Either the hardware will agree with the prediction or it won’t,”Mr. Rose said. “If it doesn’t agree, then you don’t have a quantumcomputer. On the other hand, if the data tracks the predictions, youhave smoking gun evidence that what you built is a quantum computer.”

Part of the quantum computing premise revolves around quantum bits,or qubits. These are similar to the 1 or 0 bits derived fromtransistors used in normal computers except that quantum mechanics,which works off the up and down spin of a nucleus, can hold more valuessimultaneously. So, where a pair of bits could represent just onechoice out of 11, 00, 01 and 10, a pair of qubits could store all ofthese values. (The D-Wave simulation is for a machine with 128 qubits.)

The upshot of all this is that quantum computers should be capableof processing very complex problems with thousands of variables notpractical for standard machines. And — voila — a whole new world ofcomputing opens up.


D-Wave, which has received plenty of venture capital funding, ispitching the volunteer effort as being in the name of furtheringscience. It should be noted, however, that D-Wave hopes to be afor-profit company and could rent time on existing computers to churnthrough its simulations faster.


Perhaps the volunteers donating their PCs should ask for stockoptions to go along with each set of data they send back to the company.

xbc555

早上下了个“阿瓜”，显示要24小时。
用赛扬2.0算了一天到下午4点左右进度93%.......噩梦开始了....
  机子死了，重启，“惊喜”ING==！任务进度归零，从头开始了.....
   

   这是怎么个情况呢？ 难道这个项目不设置存盘点的？

Youth

我之前也测试过一个任务，至少在前十几分钟内是没存过盘的，重启就从头开始了。

luoyinbazi

这个任务好像是不设存盘点的，我已经遇到过了，现在我的不是7×24的电脑，不运行这个项目

xbc555

明天拼了，非把这个包算完不可！

luoyinbazi

任务包变大了，P4 2.8G算一个包要24小时了！中间不能关机和中断，否则从0%开始！任务需要在72小时内完成！
不是7×24小时的，特别是机子比较慢的同志参与还是要考虑一下的。
不过有一点要说明的是：每个24小时的包，可以得分400.

luoyinbazi

任务包再次变大了，P4 2.8G 的CPU运算时间，已经从24小时，变到96小时，中间不能中断！
任务期限是20天

更正：经过几天的运算发现，每个WU接到手的时候，会显示需要96小时，但是实际运行20小时左右就可以完成。

[ 本帖最后由 luoyinbazi 于 2009-1-8 20:58 编辑 ]

Mr.Cancel

I‘am come back