Nutritious Rice for the World
проект стартовал 12 мая 2008
возможно, для очень быстрых процессоров будет выгоднее по очкам, чем медицинские субпроекты WCG, поскольку поощрение идет за количество найденных вариантов рисовых белков (или чего-то там) за 8 часов вычислений...
у меня AMD 2.4 ГГц, все-таки в пересчете на час работы, за рис давали меньше, чем FAAH
к тому же, одно и тоже задание рассылается 19 раз и раздаются очки только после того, как вернется минимум 14 результатов по этому заданию - предполагается, что у всех они будут отличаться, так что это не дубли, а скорее "всеми пальцами в небо"
Вопрос: зачем считать задание (одно!!) целых 19 раз, если можна для проверки посчитать три раза?
недоверчивые или ждут чудес
На самом деле проект очень хорошо для старых машин:
1. Очень поднимается статистика кол-ва сданых WU.
2. Даже при очень небольшом времени на кранч машина успеет выполнить WU, потому что считает всегда 8 часов реального времени.
А что касается большого лимита WU, так лучше перестраховаться в начале проекта, чем потом иметь возможно неверные результаты. Также, может быть, там используются вычисления/моделирования, которые не всегда дают один и тот же ответ (например вероятность чего-то учитывается). Тогда только при достаточно большом количестве результатов можно указать на правильный ответ.
Вот как примерно выглядит графика в проекте
Молекула вертится, типа как в розетте
Посчитал 100 процессорных дней
С этого дня вю Nutritious Rice for the World считаются по 6 часов (были по 10)
В шапке добавлен раздел "Как работает рассчетное ядро"
а золотую медальку за сколько дней дают?
у меня уже 77 процессорных дней на проекте просчитано, а так до сих пор серебрянная, хотя на других проектах уже была бы золотая точно.
90 дней - стандартный срок
Официальной новости еще нет, но на https://secure.worldcommunitygrid.org/forums/wcg/viewthread?thread=24606 что "фаза Б" проекта ("фаза А" закончилась недавно) закончится в Августе 2009! После этого возможно будет пауза для разработки новой версии клиента которая будет считать практические данные
Самый полезный проект
гречка самый полезный проект.
ещё nutritious lard может быть...
nikelong, та тю.
раком болеет 1 из 100000 а жрать хотят 1 из 1.
Один из одного хотят жрать рис? Не не не ...
даешь Nutritious Potatoes For the World!
картофель это семейство паслёновых. та ну нах его есть...
Статус проекта, апдейт
To visualize the work of this project at the protein level, each protein is drawn below as a color dot on a 200x200 grid. The color indicates the status of each protein. Approximately 65,000 proteins are now being processed in this project. Enough work will go into the processing to generate 100,000 three dimensional structures for each protein.
Не знаю кто считает этот проект но скоро финиш(где-то до апреля будут WU) , просчитано 92%.
Организовал соревнование до 1 апреля
Через 3 недели в проекте Nutritious Rice For The World планируют начать отгрузку новых заданий
Примерно до 25 марта -- 1 апреля должны закончиться ВЮ из этой фазы проекта
А я и не знал что и тут соревнования проходят
Уже 4-й день считаю WCG (весь) всеми своими компами... в Nutritious Rice For The World успел слить 10 резальтов на сумму 10,721 - так что - пусть знают наших
Жаль бразильцев не успеем обставить...
Team Name Current Score
1 BRASIL - BRAZIL@GRID 4,112,954
2 Ukraine 2,713,958
Последняя пачка ВЮ загружена в GRID. На следующей неделе задания этой фазы закончатся. Всем спасибо и приятного аппетита
Buck, POINTs в WCG равны 7 кредитам BOINC
это такой архаизм на офсайте, который тянется при переводе grid.org на платформу BOINC
Обновление статуса проекта.
Apr 1, 2010
We have begun to analyze the terabytes of results that have been generated through the generous efforts of the volunteers.
Now comes the difficult part of sifting through the data to find the best models. The folding algorithm is noise and there will be many inaccurate models. We need to find the best models from the almost 7 billion models generated. This should take approximately 3-6 months using our fastest methods. After identifying the most accurate models, we then will use the information to figure out what functions these proteins perform in the rice organism. This involves comparing the structure and sequence to known proteins and is also a time consuming process. The plant genomes are not nearly as well studied as the human and mammalian genomes which makes the process all the more difficult.
We are also developing faster and more accurate technologies to examine the data. As we have mentioned in the forums, a gpu-accelerated version of the simulation process has already been developed which is several orders of magnitude faster and more accurate. We have and are extending that technology to the analyses of the model structures. We have also http://protinfo.compbio.washington.edu/mfs/.
We are applying for funding to support these and other efforts to analyze the mountain of data that has been generated during this process. We too are volunteers, and it is our hope that our combined efforts in the NRW project will help develop rice strains that will make a difference in fighting malnutrition and feeding the world’s people. Finally, as the project comes to an end, we want to thank everyone for their generous contributions to this endeavor, especially those that volunteered their computers and time to generate the data. We really appreciated it.
Tentative future plans are to resubmit an application to the IBM to apply the Protinfo algorithm to proteins encoded by 1000 plant transcriptomes generated by the http://www.onekp.com/. This work in progress. Thus the efforts of the WCG volunteers and the results of this study will have a broader impact beyond rice proteomics.
All good things must come to an end and this is the case with the Nutritious Rice for the World project as the final results came in earlier this week.
We are analyzing the results and this will take some time. All of the scientists involved in this project are volunteers as well and we will be analyzing the data to identify proteins and genes that may be useful in breeding better rice strains. We are also applying for funding to further develop some of the technologies such as gpu acceleration of the process and sophisticated techniques that recognise structure and sequence patterns or signatures to identify the function of the protein.
There are some tentative plans to perhaps apply the software to other plant genomes.
More details are avaialble at http://protinfo.compbio.washington.edu/rice/status.html
On behalf of the scientific team involved I'd like to thank the IBM team for their support and enthusiasm for the project. Ensuring that the software could run properly on such a diverse set of hardware is quite an amazing feat and IBM deserves recognition for the time and resources that they have devoted.
However, none of this would have been possible without the help of the thousands of volunteers who donated their computers to this project. We consider it a great honour that you have allowed us onto their home machines to run our simulations these past years.
Thank you so much for your efforts.
Ха, а мне только сегодня пришло письмо от WCG о завершении этого проекта
World Community Grid is pleased to announce that as a result of the generous contribution of computing power from our members, the Nutritious Rice for the World project finished on April 6, 2010.
Now that the first step is finished, stay tuned to learn what insights the researchers find as they analyze the data.
For more information please go to News and Updates.
We still need your help with six (6) other ongoing projects! World Community Grid continues to run the following projects: FightAIDS@Home, Help Conquer Cancer, Help Cure Muscular Dystrophy - Phase 2, Help Fight Childhood Cancer, Human Proteome Folding - Phase 2, and Discovering Dengue Drugs - Together, Phase 2. All of these important projects need your computer time.
If you only had Nutritious Rice for the World selected for active projects, then you will start receiving work from the other active projects. To modify your project selection criteria, please go to your My Projects page.
The World Community Grid Team
This is a thread to update members about how the results of the Nutritious Rice for the World project are being used.
We are evaluating the model structures sent back evaluate which are most likely to be representative of the real structures. These will be posted and accessible on the website. This in itself is a difficult and time consuming process. We have developed and are developing gpu-based software to allow us to do this faster and use more accurate techniques. A paper on the technical aspects of the project should be out shortly.
After the initial analysis, we will be collaborating with other rice researchers to focus on genes of interest and analyze the models in depth to attempt to ascertain the function of these proteins.
More details as they develop...
Nutritious Rice for the World Scientist
(добавил в шапку)
Sorry about not updating for so long.
The lab has received some funding and I am now working full-time on NRW rather than in my spare time. I am very happy to be back in Seattle.
We are about to receive some new CPU/GPU servers to analyze the data and there should be something soon. We are also applying for funds to really upgrade our cluster in anticipation of the 1000 plant proteome (1KP) project.
To make things clear, the GPUs are being used to analyze the data internally. We do have a GPU-aware protein folding client as well. If World Community Grid is ready to go with GPUs we are ready to utilize GPUs in the 1KP project. If not, we can proceed using CPUs with fewer candidate structures per sequence.
Thanks for you patience. I will be updating this thread much more regularly as we get results.
NRFW predicts the structures of rice proteins.
These structures can be compared with proteins with known function. Similarity in structure implies similariy of function. Similarity of sequence to proteins of known function also implies similarity of function to those proteins. Unfortunately with rice, most sequences are dissimilar to anything that has been studied. However combining both structural and sequence similarity information allows us to assign the function of a rice protein/gene more accurately.
Geneticists/breeders use that information to develop better strains of rice as in the projects that you mentioned. So these are not competing projects but projects that can use the information in NRFW.
IRRI is also helping us focus on examining proteins/genes that they believe are most interesting for their work.
We can run the exact same client on both GPU/CPU but ideally thats something that we might want to change. It makes more sense to leverage the capabilities of GPUs to do some pre-analysis as well as generating structures.
The ATI cards are just as capable in terms of double precision math - it's just that the early acceptance of CUDA means that ATI owners were left out. As you may know, we don't need double precision so we can gain an additional 2-5x speedup and we have tested the folding software on ATI GPUs. OpenCL is supported by both NVIDIA and ATI.
I agree wholehearted with your sentiments and if anything, you may be understating the amount of useful GPUs out there. Even stock and budget GPUs for movie playing are becoming quite powerful.
Nutritious Rice for the World Scientist
хехе. велкам 2 гпу эра?
он не сравнил по скорости с коре дуо например.
апдейт статуса проекта!
Everything about this project is open - results, data, code.
In the case of NRW, adequate methods to analyze the data from 10 billion noisy protein models simply dd not exist - so honestly, the last thing I am worried about is being scooped. This is one of the few advantages of being at the bleeding edge.
I was much more worried that the new methods being developed actually work and that the results from NRW are as useful as possible so that similar approaches in the future will be funded. It is very hard to convince funding agencies, especially these days, to fund anything that hasnt been thoroughly tested and "guaranteed" to give something useful. This is one the many disadvantages of being at the bleeding edge.
However, this is starting to come together finally as the new methods are benchmarking well on our test sets. We've just had a paper accepted in Bioinformatics on one a new GPU-optimised algorithm that we are using to choose the best structures. I should have the pipeline up and running in the next week and we will start putting up structures on our website soon.
More to come...
Отлично, будет уже 2 проекта на гпу Ожидаем..
A paper was published in the journal Bioinformatics, which describes the use of Graphics Processing Units (GPU's) to accelerate computations comparing protein structures.
“Accelerated protein structure comparison using TM-score-GPU”
Lay Person Abstract:
As part of the analysis of the computed results of the Nutritious Rice for the World project, the researchers need to be able to compare protein structures and efficiently compute a similarity score. A scoring method based on “Template Modeling”, known as TM-score, provides significantly better results than the “root-mean-square-deviation” method, but requires much more computer processing time. To solve this problem the researchers developed a version of the TM-score algorithm which makes use of Graphics Processing Units (GPU's) which are found in newer video hardware, used particularly on gaming computers to enhance the visual experience. Using GPU's they were able to run millions of protein comparisons about 70 times faster. The paper describes how they accomplished this and they offer the software freely to other scientists, who may be able to use it for their research.
Motivation: Accurate comparisons of different protein structures play important roles in structural biology, structure prediction and functional annotation. The root-mean-square-deviation (RMSD) after optimal superposition is the predominant measure of similarity due to the ease and speed of computation. However, global RMSD is dependent on the length of the protein and can be dominated by divergent loops that can obscure local regions of similarity. A more sophisticated measure of structure similarity, Template Modeling -score, avoids these problems, and it is one of the measures used by the community-wide experiments of critical assessment of protein structure prediction to compare predicted models with experimental structures. TM-score calculations are, however, much slower than RMSD calculations. We have therefore implemented a very fast version of TM-score for Graphical Processing Units (TM-score-GPU), using a new and novel hybrid Kabsch/quaternion method for calculating the optimal superposition and RMSD that is designed for parallel applications. This acceleration in speed allows TM-score to be used efficiently in computationally intensive applications such as for clustering of protein models and genomewide comparisons of structure.
Results: TM-score-GPU was applied to six sets of models from Nutritious Rice for the World for a total of 3 million comparisons. TM-score-GPU is 68 times faster on an ATI 5870 GPU, on average, than the original CPU single-threaded implementation on an AMD Phenom II 810 quad-core processor. Availability and implementation: The complete source, including the GPU code and the hybrid RMSD subroutine, can be downloaded and used without restriction at http://software.compbio.washington.edu/misc/downloads/tmscore/ . The implementation is in C++/OpenCL.
Access to Paper:
To view the paper, please http://bioinformatics.oxfordjournals.org/content/28/16/2191.full.pdf.
Just to let people know.
The way that science works is that publications are months if not years behind the current work. For example, the methodology described in the paper has already been incorporated into a new method for choosing the best protein structure using TM-score as a similarity measure. This has already been applied to the rice protein structures and we have already used the best structures to predict the function of the rice genes.
But to write the results up as papers, we have to do the control calculations and benchmarking. Then after writing the manuscripts, it goes to internal reviews, external reviews, revisions, proofs etc before it gets published.
However, we will be putting up the results on our website soon - right after we do the benchmarks.
Invision Power Board
© Invision Power Services