vfrex
10-10-2006, 04:59 PM
For years now, folding@home has been the default distributive computing project that I think of, and I think that applies for most other people. Its cumulative computing power is still growing between faster processors and now the GPU client coming into existence. The competition between teams/tech sites keeps many people involved, and many others joining.
As of October 2006, more than 178,000 CPUs were actively participating in Folding@Home (active CPUs are defined as those returning work units within the last 50 days), with over 1,700,000 CPUs registered. This level of participation makes the Folding@home distributed supercomputer one of the most powerful supercomputers in the world, capable of a sustained computational level of over 190 teraFLOPS.
http://en.wikipedia.org/wiki/Folding@home
However, there are other valid projects that aren't getting enough attention. So, the majority of the rest of this post will be about other DC projects that I might consider participating in, and that I think other bjorn members might as well. If others are interested, maybe we could set up a team.
One example, rosetta@home has teams and works on windows, linux, and OSX:
Rosetta researchers are rely on a technique known as distributed computing, which pools the resources of idle computers everywhere. The project already has ~70,000 active PCs (Jun-2006) for ~38 TeraFLOPS sustained cumulative processing power, but is still actively seeking new participants to reach the 150 TeraFLOPS computing speed mark.
Rosetta also works on proteins, but here is an explanation of the difference:
I know Baker and Ranganathan and their work very well and (like the rest of the protein community) find their work very important and impressive. However, Rosetta@home and Folding@Home are addressing very different problems.
Rosetta only predicts the final folded state, not how do proteins fold (and Rosetta has nothing to do with protein misfolding). Thus, those methods are not useful for the questions we're interested in and the diseases we're tackling (Alzheimer's Disease and other aggregation related diseases).
Also, one should note that accurate computational protein structure prediction is still very challenging compared to what one can do experimentally, whereas the information obtained from Folding@home on the nature of folding and misfolding pathways matches experiment (eg with quantitative validation in rates, free energy, etc) and then goes beyond what experiment can tell us in that arena. While Rosetta has gone a long way and is a very impressive project, given the choice between a Rosetta predicted structure and a crystal structure, one would always chose the crystal structure. I bet that will be changing due to their great efforts, but that may still be a ways off for that dream to be realized.
So, both are valuable projects IMHO, but addressing very different questions. I think there are some misunderstandings out there, though. Some people think FAH is all about structure prediction (which it is not -- that's Rosetta's strength) and some think Rosetta is about misfolding related disease (which it's not, that's Folding@Home's strength). Hopefully this post helps straighten some of that out.
Then there is Grid.org, which is searching for possible cancer drugs. The downside to this project is that there is only a client for Windows. That said, most people here run Windows anyway. It features teams as well. Also, it is backed by a company, and has gone through a few projects so far, so it is quite stable and successful. Further, I like the idea that the projects here are more discrete than folding@home.
http://en.wikipedia.org/wiki/United_Devices_Cancer_Research_Project
Here is a list of other DC projects: http://en.wikipedia.org/wiki/List_of_distributed_computing_projects
Here are a few from that list that I considered noteworthy:
http://www.childhooddiseases.org/download.html - searching for drugs to treat Tuberous Sclerosis Complex
http://www.d2ol.com/ - "The Drug Design and Optimization Lab (D2OL)™ works to discover drug candidates against Anthrax, Smallpox, Ebola and SARS and other potentially devastating infectious diseases." This one also supports teams and Windows, Linux, Solaris, OSX.
http://www.stephenbrooks.org/muon1/ - The experiment is called the Neutrino Factory, scheduled for construction some time from 2015. Its primary aim is to fire beams of neutrinos (fundamental particles) through the Earth's interior to detector stations on different continents. They're doing this to measure whether they change type en route (there are 3 types of neutrino) and data from this in turn will allow them to determine the neutrino's mass more accurately.
The reason they want to do this is that the neutrino is just about the most common particle in the universe (billions pass through your body every second) and its mass will influence such things as the evolution of the universe and the exact way matter was first formed in the 'big bang'. In fact, neutrinos make up one quarter of the types of matter particles specified in the current 'standard model' of physics. In order to progress to more advanced theories of physics, it is often crucial to know the properties of particles to high precision in order to distinguish between the slightly different predictions of alternative theories.
Actually the entire Neutrino Factory complex (estimated to cost at least $1.9bn) will have several scientific aims. The neutrinos are used for fundamental physics experiments, but the proton beam that is produced at the start (this hits the target rod at the beginning of the simulation you download) is also going to be used in experiments such as neutralising radioactive waste by transmuting the radioactive elements into stable ones, and providing a high intensity source of neutrons for 3D atomic microscopy. The muon beam that will be coming from the end of the section we are optimising can also be used as the basis for a 'muon collider', a machine that could produce the highest-energy collisions between fundamental particles ever made artificially.
You are simulating the part of the process where the proton beam hits the target rod and causes pions to be emitted, which decay into muons. These would then proceed to a storage ring and decay into electrons and the neutrinos that are used for experiments. This is a fairly critical part of the apparatus, which catches the pions and confines some of them into a beam while they decay. The efficiency of this dictates that of the entire machine because it is built of a lot of acceleration stages 'in series' with each other. Whether the project eventually gets funded to be built depends on what levels of performance can be achieved with the designs generated during the present R&D. However, users of this program have already doubled the estimated efficiency of one stage and more are to be optimised in the future.
http://www.netdimes.org/ - Not so much a CPU cruncher, but this might be something a community of tech geeks should be interested in:
How does the Internet look like? How does it evolve?
DIMES is a distributed scientific research project, aimed to study the structure and topology of the Internet, with the help of a volunteer community (similar in spirit to projects such as SETI@Home).
Due to the way the Internet is engineered, distributing the Internet mapping effort is very important, and the only efficient way to measure the Internet structure is by asking you to participate. What we ask is not so much your CPU or bandwidth (which we hardly consume), but rather, your location. The more places we'll have presence in, the more accurate our maps will be. Understanding the structure and function of the Internet is an important research task, that will allow to make the Internet a better place for all of us.
The DIMES agent performs Internet measurements such as TRACEROUTE and PING at a low rate, consuming at peak 1KB/S. The agent DOES NOT send any information about its host's activity/personal data, and sends ONLY the results of its own measurements. Aside from giving a good feeling, running the DIMES agent will also provide you with maps of how the Internet looks from your home (currently) and will (in the future) provide you with a personalized 'Internet weather report' and other user-focused features.
Rosetta seems nice being cross platform, and in that it is attacking proteins from a different angle than F@H. There are a few projects that are attempting the same as Rosetta, although using different methods. Rosetta is one of the most accurate from that bunch. From the quote from Professor Vijay Pande above about the differences between F@H and Rosetta@H, While Rosetta has gone a long way and is a very impressive project, given the choice between a Rosetta predicted structure and a crystal structure, one would always chose the crystal structure. I bet that will be changing due to their great efforts, but that may still be a ways off for that dream to be realized.
I'm going to continue editing this post over the next couple of hours with more details about other projects and stuff. What I'd like to see though for now though, is discussion about the main point. That is, whether you guys feel that more emphasis should be placed on other projects. Could our idle CPU cycles be used more effectively than throwing them at what everybody else is anyway.
As if you needed more reading material, here is a nice writeup of the differences and goals of some of these projects: http://www.hyper.net/dc-howto.html#diffs
As of October 2006, more than 178,000 CPUs were actively participating in Folding@Home (active CPUs are defined as those returning work units within the last 50 days), with over 1,700,000 CPUs registered. This level of participation makes the Folding@home distributed supercomputer one of the most powerful supercomputers in the world, capable of a sustained computational level of over 190 teraFLOPS.
http://en.wikipedia.org/wiki/Folding@home
However, there are other valid projects that aren't getting enough attention. So, the majority of the rest of this post will be about other DC projects that I might consider participating in, and that I think other bjorn members might as well. If others are interested, maybe we could set up a team.
One example, rosetta@home has teams and works on windows, linux, and OSX:
Rosetta researchers are rely on a technique known as distributed computing, which pools the resources of idle computers everywhere. The project already has ~70,000 active PCs (Jun-2006) for ~38 TeraFLOPS sustained cumulative processing power, but is still actively seeking new participants to reach the 150 TeraFLOPS computing speed mark.
Rosetta also works on proteins, but here is an explanation of the difference:
I know Baker and Ranganathan and their work very well and (like the rest of the protein community) find their work very important and impressive. However, Rosetta@home and Folding@Home are addressing very different problems.
Rosetta only predicts the final folded state, not how do proteins fold (and Rosetta has nothing to do with protein misfolding). Thus, those methods are not useful for the questions we're interested in and the diseases we're tackling (Alzheimer's Disease and other aggregation related diseases).
Also, one should note that accurate computational protein structure prediction is still very challenging compared to what one can do experimentally, whereas the information obtained from Folding@home on the nature of folding and misfolding pathways matches experiment (eg with quantitative validation in rates, free energy, etc) and then goes beyond what experiment can tell us in that arena. While Rosetta has gone a long way and is a very impressive project, given the choice between a Rosetta predicted structure and a crystal structure, one would always chose the crystal structure. I bet that will be changing due to their great efforts, but that may still be a ways off for that dream to be realized.
So, both are valuable projects IMHO, but addressing very different questions. I think there are some misunderstandings out there, though. Some people think FAH is all about structure prediction (which it is not -- that's Rosetta's strength) and some think Rosetta is about misfolding related disease (which it's not, that's Folding@Home's strength). Hopefully this post helps straighten some of that out.
Then there is Grid.org, which is searching for possible cancer drugs. The downside to this project is that there is only a client for Windows. That said, most people here run Windows anyway. It features teams as well. Also, it is backed by a company, and has gone through a few projects so far, so it is quite stable and successful. Further, I like the idea that the projects here are more discrete than folding@home.
http://en.wikipedia.org/wiki/United_Devices_Cancer_Research_Project
Here is a list of other DC projects: http://en.wikipedia.org/wiki/List_of_distributed_computing_projects
Here are a few from that list that I considered noteworthy:
http://www.childhooddiseases.org/download.html - searching for drugs to treat Tuberous Sclerosis Complex
http://www.d2ol.com/ - "The Drug Design and Optimization Lab (D2OL)™ works to discover drug candidates against Anthrax, Smallpox, Ebola and SARS and other potentially devastating infectious diseases." This one also supports teams and Windows, Linux, Solaris, OSX.
http://www.stephenbrooks.org/muon1/ - The experiment is called the Neutrino Factory, scheduled for construction some time from 2015. Its primary aim is to fire beams of neutrinos (fundamental particles) through the Earth's interior to detector stations on different continents. They're doing this to measure whether they change type en route (there are 3 types of neutrino) and data from this in turn will allow them to determine the neutrino's mass more accurately.
The reason they want to do this is that the neutrino is just about the most common particle in the universe (billions pass through your body every second) and its mass will influence such things as the evolution of the universe and the exact way matter was first formed in the 'big bang'. In fact, neutrinos make up one quarter of the types of matter particles specified in the current 'standard model' of physics. In order to progress to more advanced theories of physics, it is often crucial to know the properties of particles to high precision in order to distinguish between the slightly different predictions of alternative theories.
Actually the entire Neutrino Factory complex (estimated to cost at least $1.9bn) will have several scientific aims. The neutrinos are used for fundamental physics experiments, but the proton beam that is produced at the start (this hits the target rod at the beginning of the simulation you download) is also going to be used in experiments such as neutralising radioactive waste by transmuting the radioactive elements into stable ones, and providing a high intensity source of neutrons for 3D atomic microscopy. The muon beam that will be coming from the end of the section we are optimising can also be used as the basis for a 'muon collider', a machine that could produce the highest-energy collisions between fundamental particles ever made artificially.
You are simulating the part of the process where the proton beam hits the target rod and causes pions to be emitted, which decay into muons. These would then proceed to a storage ring and decay into electrons and the neutrinos that are used for experiments. This is a fairly critical part of the apparatus, which catches the pions and confines some of them into a beam while they decay. The efficiency of this dictates that of the entire machine because it is built of a lot of acceleration stages 'in series' with each other. Whether the project eventually gets funded to be built depends on what levels of performance can be achieved with the designs generated during the present R&D. However, users of this program have already doubled the estimated efficiency of one stage and more are to be optimised in the future.
http://www.netdimes.org/ - Not so much a CPU cruncher, but this might be something a community of tech geeks should be interested in:
How does the Internet look like? How does it evolve?
DIMES is a distributed scientific research project, aimed to study the structure and topology of the Internet, with the help of a volunteer community (similar in spirit to projects such as SETI@Home).
Due to the way the Internet is engineered, distributing the Internet mapping effort is very important, and the only efficient way to measure the Internet structure is by asking you to participate. What we ask is not so much your CPU or bandwidth (which we hardly consume), but rather, your location. The more places we'll have presence in, the more accurate our maps will be. Understanding the structure and function of the Internet is an important research task, that will allow to make the Internet a better place for all of us.
The DIMES agent performs Internet measurements such as TRACEROUTE and PING at a low rate, consuming at peak 1KB/S. The agent DOES NOT send any information about its host's activity/personal data, and sends ONLY the results of its own measurements. Aside from giving a good feeling, running the DIMES agent will also provide you with maps of how the Internet looks from your home (currently) and will (in the future) provide you with a personalized 'Internet weather report' and other user-focused features.
Rosetta seems nice being cross platform, and in that it is attacking proteins from a different angle than F@H. There are a few projects that are attempting the same as Rosetta, although using different methods. Rosetta is one of the most accurate from that bunch. From the quote from Professor Vijay Pande above about the differences between F@H and Rosetta@H, While Rosetta has gone a long way and is a very impressive project, given the choice between a Rosetta predicted structure and a crystal structure, one would always chose the crystal structure. I bet that will be changing due to their great efforts, but that may still be a ways off for that dream to be realized.
I'm going to continue editing this post over the next couple of hours with more details about other projects and stuff. What I'd like to see though for now though, is discussion about the main point. That is, whether you guys feel that more emphasis should be placed on other projects. Could our idle CPU cycles be used more effectively than throwing them at what everybody else is anyway.
As if you needed more reading material, here is a nice writeup of the differences and goals of some of these projects: http://www.hyper.net/dc-howto.html#diffs