shai_
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 05:05:34 AM |
|
anyone knows how to GPU fold in Linux ? they mention: "Folding on AMD GPUs is problematic in linux due to poor OpenCL driver suppport from AMD, so GPU folding on AMD hardware is not supported at this time. This could change at any time with a newer driver. Linux GPU folding is still in beta testing." https://folding.stanford.edu/home/guide/linux-install-guide/
|
|
|
|
DearPassin46
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 80
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 05:14:01 AM |
|
Pay attention to you for a long time, but the progress is too slow, at least GRC is worked out, a little disappointed you too slow.
|
|
|
|
LiteMine
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 05:15:04 AM |
|
You can't wait 5 days? ![Roll Eyes](https://bitcointalk.org/Smileys/default/rolleyes.gif)
|
|
|
|
r3animation
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 05:35:00 AM |
|
I think we will receive a email which let us IPO ![Angry](https://bitcointalk.org/Smileys/default/angry.gif) Anyone understand the above? Or is he being mega-stupid?
|
|
|
|
lopalcar
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 10:19:34 AM Last edit: May 09, 2014, 10:31:19 AM by lopalcar |
|
The problem I see with your suggestion of the algorithm is that it hurts the effectiveness of the science behind it. Theoretically an algorithm could be written that could fold protein by itself so we wouldn't use f@h but still fold proteins however there are 3 problems with this.
1) F@H has been working on what they have been doing for 10 or more years and they have a very nice system in place for validation and all of that. Plus there points system runs off of comparing what your machine can do compared to a test system which results in increased costs for the curecoin devs who will need to set this up. Getting a validated system is quite hard but you would need a near perfect one in place before you started otherwise people could cheat the system by submitting invalid units - a problem that F@H has experienced a few times in the past I believe. Also I believe F@H is not fully open source - because people could exploit the code and use it to submit invalid work units so this code would need to be written from scratch.
2) The folding of random problems by itself would be useless - we would need a list of what proteins to fold and these are normally developed in Stanford laboratories and then submitted to the network. The cost in research time and equipment to determine what proteins to fold is probably close to if not in the millions of dollars something which I doubt the developers can afford. However, without this step done beforehand it would not be possible to start the coin.
3) The whole point of F@H is to get protein results which are then taken back to the lab and tested out to confirm how useful they are - the computing portion is done to lower the number of possible choices because testing every possible combination would take forever and not be effective. As a result more work would need to be done in order to do the actual science or the protein folding is again useless.
And what you said about the Asics I disagree - I think the network will manage itself. 20% of the coins is still a decent amount and I think it would attract Asics from people who have Asics that are no longer effective for Bitcoin mining (such as first generation Bitcoin Asics from Avalon or Butterfly labs). A lot of these people will/can turn to Bitcoin and then the profitability of the network will manage itself.
I understand that the network could become screwed by a few Asic miners but I also believe that as the coin will attract enough Asic miners to keep it secure.
I like your idea of removing the Asics from the equation and sticking with a F@H algorithm but I believe that the time (years) and the money (close to millions) would probably take so much time that it is not worth it and probably very difficult to get backing for.
Also to address your question about the primes the reason they are able to do it is because the prime computing doesn't really build off any outside research and the code for how to calculate prime numbers on a laptop is pretty widely available so you just need to modify it some - it is a much simpler calculation and it is a lot easier to confirm whether or not someone has submitted a valid answer.
Hope this helped - I don't know it all perfectly but that is based of my knowledge and research someone else on the project probably has more to add/update with this.
Thanks, Cameron
Well,you say more or less the same as I think, I know all of this which you are saying and see many problems in it: 1) F@H didn't received many cheating trys simply because they aren't playing with money, lets see what happens when they system becomes attacked day after day for people wanting to earn more curecoins... 2) I know that fold a random aminoacid sequence wouldn't have any interest, but there are more ways to to make this more open sourced allowing to any researcher to submitt workunits to the system, not only stanford (maybe an organization where every researcher can join and subbmit workunits via BOINC?). For example, anyone could upload a sequence to http://robetta.bakerlab.org/ and then they can send that workunit to BOINC via resetta project. I think this is better than work only for stanford. 3) It's discussed in 2) there are many other alternatives to f@h used for foldin, docking and analize protein interactions which are open source and will allow to more people do use the "folding" network for their own, I think this is the aim of a decentralized crypto, isn't it? About asics, I mean that you need 100 miners of first gen to compete with one new miner, someone with 5 of this miners could break the network easy having in mind that many people expect to use block erupters usb for mining this. And about prime numbers example I know what you say, for that which I said that this will require much work and knowledge, but anyways, we have the alternative that I said before using BOINC.
|
|
|
|
laughingbear
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 12:22:59 PM |
|
The problem I see with your suggestion of the algorithm is that it hurts the effectiveness of the science behind it. Theoretically an algorithm could be written that could fold protein by itself so we wouldn't use f@h but still fold proteins however there are 3 problems with this.
1) F@H has been working on what they have been doing for 10 or more years and they have a very nice system in place for validation and all of that. Plus there points system runs off of comparing what your machine can do compared to a test system which results in increased costs for the curecoin devs who will need to set this up. Getting a validated system is quite hard but you would need a near perfect one in place before you started otherwise people could cheat the system by submitting invalid units - a problem that F@H has experienced a few times in the past I believe. Also I believe F@H is not fully open source - because people could exploit the code and use it to submit invalid work units so this code would need to be written from scratch.
2) The folding of random problems by itself would be useless - we would need a list of what proteins to fold and these are normally developed in Stanford laboratories and then submitted to the network. The cost in research time and equipment to determine what proteins to fold is probably close to if not in the millions of dollars something which I doubt the developers can afford. However, without this step done beforehand it would not be possible to start the coin.
3) The whole point of F@H is to get protein results which are then taken back to the lab and tested out to confirm how useful they are - the computing portion is done to lower the number of possible choices because testing every possible combination would take forever and not be effective. As a result more work would need to be done in order to do the actual science or the protein folding is again useless.
And what you said about the Asics I disagree - I think the network will manage itself. 20% of the coins is still a decent amount and I think it would attract Asics from people who have Asics that are no longer effective for Bitcoin mining (such as first generation Bitcoin Asics from Avalon or Butterfly labs). A lot of these people will/can turn to Bitcoin and then the profitability of the network will manage itself.
I understand that the network could become screwed by a few Asic miners but I also believe that as the coin will attract enough Asic miners to keep it secure.
I like your idea of removing the Asics from the equation and sticking with a F@H algorithm but I believe that the time (years) and the money (close to millions) would probably take so much time that it is not worth it and probably very difficult to get backing for.
Also to address your question about the primes the reason they are able to do it is because the prime computing doesn't really build off any outside research and the code for how to calculate prime numbers on a laptop is pretty widely available so you just need to modify it some - it is a much simpler calculation and it is a lot easier to confirm whether or not someone has submitted a valid answer.
Hope this helped - I don't know it all perfectly but that is based of my knowledge and research someone else on the project probably has more to add/update with this.
Thanks, Cameron
Well,you say more or less the same as I think, I know all of this which you are saying and see many problems in it: 1) F@H didn't received many cheating trys simply because they aren't playing with money, lets see what happens when they system becomes attacked day after day for people wanting to earn more curecoins... 2) I know that fold a random aminoacid sequence wouldn't have any interest, but there are more ways to to make this more open sourced allowing to any researcher to submitt workunits to the system, not only stanford (maybe an organization where every researcher can join and subbmit workunits via BOINC?). For example, anyone could upload a sequence to http://robetta.bakerlab.org/ and then they can send that workunit to BOINC via resetta project. I think this is better than work only for stanford. 3) It's discussed in 2) there are many other alternatives to f@h used for foldin, docking and analize protein interactions which are open source and will allow to more people do use the "folding" network for their own, I think this is the aim of a decentralized crypto, isn't it? About asics, I mean that you need 100 miners of first gen to compete with one new miner, someone with 5 of this miners could break the network easy having in mind that many people expect to use block erupters usb for mining this. And about prime numbers example I know what you say, for that which I said that this will require much work and knowledge, but anyways, we have the alternative that I said before using BOINC. I look forward to your coin, when will it be out?
|
|
|
|
cameronpalte
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 03:37:09 PM |
|
anyone knows how to GPU fold in Linux ? they mention: "Folding on AMD GPUs is problematic in linux due to poor OpenCL driver suppport from AMD, so GPU folding on AMD hardware is not supported at this time. This could change at any time with a newer driver. Linux GPU folding is still in beta testing." https://folding.stanford.edu/home/guide/linux-install-guide/Just looked it up - yes as of now if you have an AMD Gpu you can not fold on Linux. I'll see if I can talk to them about that because a lot of people are currently folding scrypt coins with Linux AMD rigs that are becoming obsolete so it would be nice to have a system that would allow these people to switch over. However, if you want to fold as of now you need to either switch to Nvidia GPUs or get a Windows operating system.
|
|
|
|
ChristianVirtual
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 71
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 03:41:13 PM |
|
1) F@H didn't received many cheating trys simply because they aren't playing with money, lets see what happens when they system becomes attacked day after day for people wanting to earn more curecoins... 2) I know that fold a random aminoacid sequence wouldn't have any interest, but there are more ways to to make this more open sourced allowing to any researcher to submitt workunits to the system, not only stanford (maybe an organization where every researcher can join and subbmit workunits via BOINC?). For example, anyone could upload a sequence to http://robetta.bakerlab.org/ and then they can send that workunit to BOINC via resetta project. I think this is better than work only for stanford. 3) It's discussed in 2) there are many other alternatives to f@h used for foldin, docking and analize protein interactions which are open source and will allow to more people do use the "folding" network for their own, I think this is the aim of a decentralized crypto, isn't it? I kind of share the fears regarding your point 1; hopefully not happen. As for Open source: the base of FAH is open source, every one can get it from here https://simtk.org/home/openmm. Researcher could build their own infrastructure based on that. I have no problem that FAH keeps their system closed . It's their decision to ensure integrity of the results created during the process.
|
|
|
|
cameronpalte
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 03:46:22 PM |
|
Well,you say more or less the same as I think, I know all of this which you are saying and see many problems in it: 1) F@H didn't received many cheating trys simply because they aren't playing with money, lets see what happens when they system becomes attacked day after day for people wanting to earn more curecoins... 2) I know that fold a random aminoacid sequence wouldn't have any interest, but there are more ways to to make this more open sourced allowing to any researcher to submitt workunits to the system, not only stanford (maybe an organization where every researcher can join and subbmit workunits via BOINC?). For example, anyone could upload a sequence to http://robetta.bakerlab.org/ and then they can send that workunit to BOINC via resetta project. I think this is better than work only for stanford. 3) It's discussed in 2) there are many other alternatives to f@h used for foldin, docking and analize protein interactions which are open source and will allow to more people do use the "folding" network for their own, I think this is the aim of a decentralized crypto, isn't it? About asics, I mean that you need 100 miners of first gen to compete with one new miner, someone with 5 of this miners could break the network easy having in mind that many people expect to use block erupters usb for mining this. And about prime numbers example I know what you say, for that which I said that this will require much work and knowledge, but anyways, we have the alternative that I said before using BOINC. 1) There are currently some places that offer small rewards for folding. These rewards can be pretty substantial but aren't a lot because of costs if you use regular folding. However, if you were able to cheat the system the rewards would become significantly more worthwhile. I do agree that F@H will probably get more cheat attempts as the coin comes hout, however, as of now they have quite a good system in place that would keep someone from folding random amino acids. I'll do more research into the details of their system to see how useful it is. 2) You could setup an oganization - except boinc has strict rules and generally you can't submit work units by themselves. And also, I don't think a system like this at least initially will be able to generate enough computing jobs for everyone mining the coin to be able to have something to do. As said before a major problem that I see with making it open source is that their won't be enough work units for everyone mining the coin to have something. The bitcoin network is currently 160x as large as Stanford's F@H network which is like 45 petaflops. If we manage to add say another 45 petaflops then now they are theoretically getting results twice as fast which means they need to produce work units twice as fast. For a smaller network we may multiply it in size by say 50x and they may not be able to handle the demands. I understand how one asic could get a few percent of the overall power especially in the beginning but I'm confused as too how that would crash the network. I think Boinc could work however, Boinc would still have the same limitations you talked about with F@H and the calculations would still be integrated in a 80-20% system. I don't think they would change anything if they switched to boinc. And as ChristianVirtual just touched on - part of the reason F@H is closed source for some parts is so that people don't get access to those parts and try to submit fake work units.
|
|
|
|
ChristianVirtual
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 71
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 03:50:17 PM |
|
anyone knows how to GPU fold in Linux ?
Works well with nvidia cards under Ubuntu. I run two GTX780 and one 660TI. Just need to make sure the "right" driver versions are installed. Original NV driver, not open source. And newer cards don't like older driver and vice versa.
|
|
|
|
Vorksholk
Legendary
Offline
Activity: 1713
Merit: 1029
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 05:02:35 PM |
|
Hey everyone! To consolidate the above information into one explanation:
1.) Curecoin, as cygnusxi mentioned, is a hybrid PoW/PoS coin. The traditional PoW was, of course, SHA256D for Bitcoin. Like Bitcoin, Curecoin uses SHA256D, and the PoS first implemented by PPCoin.
2.) People have expressed concern over the possibility of people cheating Folding@Home. As pointed out higher up in this thread, parts of F@H that have to do with work formatting are closed-source in an attempt to eliminate the possibility of people faking workunits. However, no downloadable software is truly 'closed source', as at some level the software has to actually run, and to do so must be expressed in machine code/assembly. Folding@Home has server-side integrity checks of work units which aim to also eliminate attempts at cheating. Of course people will try, most will fail. As the project moves on, exploits will be found and fixed. Any project like this will suffer similar attacks, and at this point it is a small concern.
3.) The folding pool can be adapted to support other projects. Right now, Folding@Home is the project we chose to integrate first with. However, in the future, other projects can be added/removed and tracked. Our back-end works by aggregating statistics from Folding@Home, and doing a proportional payout. In a simple example, if User A got 20,000 Points on Tuesday and User B got 40,000 Points on Tuesday with no other users doing any folding, User A would get 33.3% of the Curecoins for that day's payout, and User B would get 66.6% of the Curecoins that day. If we were to add an additional project, we would need to find a conversion of points between the projects (use a benchmark machine or two, and test how many points per day that machine could get with each project for both CPU work and GPU work, and strike a balance by using a proportion to convert points from both projects into a normal point system). If a benchmarking machine could get 8,000 points per day on, say, Rosetta@Home while getting 12,000 PPD on F@H, then each Rosetta@Home point would be worth 3/2 of a F@H point, so if one person got 8000 R@H points and another got 12000 F@H points, they would receive equal payout, as they used approximately equal computational power for the same amount of time.
While we are not currently working on implementing other projects, the possibility is certainly there, and the project is not, by definition, tied to Folding@Home.
4.) Folding@Home is a major distributed computation network, and therefore has the resources to ensure that the project can continue to grow without causing problems.
People have also expressed concern over the proof-of-work being SHA256D. To summarize, SHA256D is a very light hashing algorithm that high-performance ASICs have been created and distributed for. Most other algorithms currently require the use of a GPU or CPU. ASICs are unable to do folding, while GPUs and CPUs are slow at SHA256D, so the logical choice is to allow ASICs to secure the network by doing hashing, while GPUs and CPUs can do the work they are great at--folding.
|
|
|
|
laughingbear
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 05:14:34 PM |
|
People are going to have doubt and concerns, and debate these endlessly until they see everything in action. If everything works as intended this has huge potential. You would be a fool not to be enthusiastically involved in this, at least till a problem is found.
|
|
|
|
lopalcar
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 07:35:25 PM |
|
Well,you say more or less the same as I think, I know all of this which you are saying and see many problems in it: 1) F@H didn't received many cheating trys simply because they aren't playing with money, lets see what happens when they system becomes attacked day after day for people wanting to earn more curecoins... 2) I know that fold a random aminoacid sequence wouldn't have any interest, but there are more ways to to make this more open sourced allowing to any researcher to submitt workunits to the system, not only stanford (maybe an organization where every researcher can join and subbmit workunits via BOINC?). For example, anyone could upload a sequence to http://robetta.bakerlab.org/ and then they can send that workunit to BOINC via resetta project. I think this is better than work only for stanford. 3) It's discussed in 2) there are many other alternatives to f@h used for foldin, docking and analize protein interactions which are open source and will allow to more people do use the "folding" network for their own, I think this is the aim of a decentralized crypto, isn't it? About asics, I mean that you need 100 miners of first gen to compete with one new miner, someone with 5 of this miners could break the network easy having in mind that many people expect to use block erupters usb for mining this. And about prime numbers example I know what you say, for that which I said that this will require much work and knowledge, but anyways, we have the alternative that I said before using BOINC. 1) There are currently some places that offer small rewards for folding. These rewards can be pretty substantial but aren't a lot because of costs if you use regular folding. However, if you were able to cheat the system the rewards would become significantly more worthwhile. I do agree that F@H will probably get more cheat attempts as the coin comes hout, however, as of now they have quite a good system in place that would keep someone from folding random amino acids. I'll do more research into the details of their system to see how useful it is. 2) You could setup an oganization - except boinc has strict rules and generally you can't submit work units by themselves. And also, I don't think a system like this at least initially will be able to generate enough computing jobs for everyone mining the coin to be able to have something to do. As said before a major problem that I see with making it open source is that their won't be enough work units for everyone mining the coin to have something. The bitcoin network is currently 160x as large as Stanford's F@H network which is like 45 petaflops. If we manage to add say another 45 petaflops then now they are theoretically getting results twice as fast which means they need to produce work units twice as fast. For a smaller network we may multiply it in size by say 50x and they may not be able to handle the demands. I understand how one asic could get a few percent of the overall power especially in the beginning but I'm confused as too how that would crash the network. I think Boinc could work however, Boinc would still have the same limitations you talked about with F@H and the calculations would still be integrated in a 80-20% system. I don't think they would change anything if they switched to boinc. And as ChristianVirtual just touched on - part of the reason F@H is closed source for some parts is so that people don't get access to those parts and try to submit fake work units. Well, using BOINC and a POS blockchain like nxt or something symilar to exocoin (which once the block is found the miners stop till the next block need to be generated "I think") would allow us to get rid of the PoW or nearly reduce it to cero, I think that a p2p verification of the workunits should be possible, if each workunit is distributed to many workers, we only need to compare them and see if they have symilar results, this could be done by all nodes because I don't expect that comparing solutions would require many cpu time "and assuming most wost workers are honest ofcourse, I think that a system to track the workunit procedence could be developed to in case one corrupt worker is sending many fake results from same place" I'm saying boinc because I think it accept a wider range of works. About work demand in the beginings, this should be easy to solve, simply spaming for example sequences of proteins which doesn't have a 3d model yet and build one via abinitio, this use many cpu time, and I'm sure that for gpus are other friendly things to do and easy to spam to the network. The last think I dislike or don't understan is how the coins will be distributed, will the pool do that? so then, the technology is in the pool, not in the coin?, is simply a bitcoin clone centraliced in hands of the pool owners which will have the wallet where 80% of the coins will go and the pool software will distribute them? I would apreciate an explanation about this issue, because these couldn't even get considered a cryptocurrency. And please, don't think I'm intending to annoy, I really love this project, but think it could be done better, I'm not a coder like I think, only have ideas and can make the algorithms but don't know how to give them to the computer ![Sad](https://bitcointalk.org/Smileys/default/sad.gif)
|
|
|
|
cameronpalte
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 08:02:20 PM |
|
Well, using BOINC and a POS blockchain like nxt or something symilar to exocoin (which once the block is found the miners stop till the next block need to be generated "I think") would allow us to get rid of the PoW or nearly reduce it to cero, I think that a p2p verification of the workunits should be possible, if each workunit is distributed to many workers, we only need to compare them and see if they have symilar results, this could be done by all nodes because I don't expect that comparing solutions would require many cpu time "and assuming most wost workers are honest ofcourse, I think that a system to track the workunit procedence could be developed to in case one corrupt worker is sending many fake results from same place" I'm saying boinc because I think it accept a wider range of works. About work demand in the beginings, this should be easy to solve, simply spaming for example sequences of proteins which doesn't have a 3d model yet and build one via abinitio, this use many cpu time, and I'm sure that for gpus are other friendly things to do and easy to spam to the network. The last think I dislike or don't understan is how the coins will be distributed, will the pool do that? so then, the technology is in the pool, not in the coin?, is simply a bitcoin clone centraliced in hands of the pool owners which will have the wallet where 80% of the coins will go and the pool software will distribute them? I would apreciate an explanation about this issue, because these couldn't even get considered a cryptocurrency. And please, don't think I'm intending to annoy, I really love this project, but think it could be done better, I'm not a coder like I think, only have ideas and can make the algorithms but don't know how to give them to the computer ![Sad](https://bitcointalk.org/Smileys/default/sad.gif) They could have used something like nxt or exocoin but I don't think these coins existed when they started curecoin development. However, I'm not sure for their reason for this. A p2p verification could be acceptable but I doubt stanford would accept those. A system could be made, however, I don't think it would work. Spamming wouldn't work because they could be validated as Vorksholk addressed. The coins will be distributed based of how many points you get for the 80% and I believe the 20% is distributed the same way Bitcoins are distributed. I'm glad you love the project you have some good ideas ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) .
|
|
|
|
lopalcar
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 08:24:44 PM |
|
Well, using BOINC and a POS blockchain like nxt or something symilar to exocoin (which once the block is found the miners stop till the next block need to be generated "I think") would allow us to get rid of the PoW or nearly reduce it to cero, I think that a p2p verification of the workunits should be possible, if each workunit is distributed to many workers, we only need to compare them and see if they have symilar results, this could be done by all nodes because I don't expect that comparing solutions would require many cpu time "and assuming most wost workers are honest ofcourse, I think that a system to track the workunit procedence could be developed to in case one corrupt worker is sending many fake results from same place" I'm saying boinc because I think it accept a wider range of works. About work demand in the beginings, this should be easy to solve, simply spaming for example sequences of proteins which doesn't have a 3d model yet and build one via abinitio, this use many cpu time, and I'm sure that for gpus are other friendly things to do and easy to spam to the network. The last think I dislike or don't understan is how the coins will be distributed, will the pool do that? so then, the technology is in the pool, not in the coin?, is simply a bitcoin clone centraliced in hands of the pool owners which will have the wallet where 80% of the coins will go and the pool software will distribute them? I would apreciate an explanation about this issue, because these couldn't even get considered a cryptocurrency. And please, don't think I'm intending to annoy, I really love this project, but think it could be done better, I'm not a coder like I think, only have ideas and can make the algorithms but don't know how to give them to the computer ![Sad](https://bitcointalk.org/Smileys/default/sad.gif) They could have used something like nxt or exocoin but I don't think these coins existed when they started curecoin development. However, I'm not sure for their reason for this. A p2p verification could be acceptable but I doubt stanford would accept those. A system could be made, however, I don't think it would work. Spamming wouldn't work because they could be validated as Vorksholk addressed. The coins will be distributed based of how many points you get for the 80% and I believe the 20% is distributed the same way Bitcoins are distributed. I'm glad you love the project you have some good ideas ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) . I know they doesn't exist then, but yes some months ago, I think that was enbought time to do something, anyways, lets see how this work, we always have time to make a better option ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) And ofcourse, the p2p validation is getting rid off stanford, the aim, in my opinion, is to create a researchers computational network where everybody could send their works Could you quote this issue about spaming "unusefull" work to the network? Yes, but the 80% of coins where they come from? Are generated by the network every day and distributed to miners attending to coin code or this is done by the pool? Are they premined and stored in a wallet in hands of pool owners and then distributed to miners? I don't understand yet ![Embarrassed](https://bitcointalk.org/Smileys/default/embarrassed.gif) A white paper or something symilar would be great, in more than one year that I have been watching this project this could be done, couldn't it?
|
|
|
|
cameronpalte
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 09:14:18 PM |
|
I know they doesn't exist then, but yes some months ago, I think that was enbought time to do something, anyways, lets see how this work, we always have time to make a better option ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) And ofcourse, the p2p validation is getting rid off stanford, the aim, in my opinion, is to create a researchers computational network where everybody could send their works Could you quote this issue about spaming "unusefull" work to the network? Yes, but the 80% of coins where they come from? Are generated by the network every day and distributed to miners attending to coin code or this is done by the pool? Are they premined and stored in a wallet in hands of pool owners and then distributed to miners? I don't understand yet ![Embarrassed](https://bitcointalk.org/Smileys/default/embarrassed.gif) A white paper or something symilar would be great, in more than one year that I have been watching this project this could be done, couldn't it? Here is the quote from Vorsholk about validating work units: 2.) People have expressed concern over the possibility of people cheating Folding@Home. As pointed out higher up in this thread, parts of F@H that have to do with work formatting are closed-source in an attempt to eliminate the possibility of people faking workunits. However, no downloadable software is truly 'closed source', as at some level the software has to actually run, and to do so must be expressed in machine code/assembly. Folding@Home has server-side integrity checks of work units which aim to also eliminate attempts at cheating. Of course people will try, most will fail. As the project moves on, exploits will be found and fixed. Any project like this will suffer similar attacks, and at this point it is a small concern.
Yes lets see how they work. I think they idea of researchers submitting their own work units is a good idea - however, there would need to be a way to validate first whether the work units are completed and secondly if there are enough work units it is much more difficult coding wise and raw resource wise. I'm guessing a white paper could be done in the year - I'm sure they will release one at some time. That would definitely help. As for trying to answer your questions about the coins. The coins aren't premined and stored - that wouldn't work well. The way they work I believe is similar to Bitcoins. So let me take a stab at this. Coin generation is like Bitcoin. So let's say a block is generated every 10 minutes containing 25 coins - like Bitcoin. Once those 25 coins are generated it is split 20% and 80%. 20% of the coins is split over the Asics similar to how coins are currently split with Bitcoins so the blocks go to one person who did the most mining and there will be pools for this and such. The 80% of the coins will split proportionally based on how many points are generated. For example if during this time 5 people submit work units for 1000 - 2000 - 2000 - 3500 - 1500 points. Therefore person 1 would get 10% of the coins person 2 gets 20% person 3 gets 20% person 4 gets 35% person 5 gets 15%. These percents are for the 80% total coins. This shows how the 80% of coins are generated with each block.
|
|
|
|
laughingbear
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 09:20:14 PM |
|
The 80% of the coins will split proportionally based on how many points are generated. For example if during this time 5 people submit work units for 1000 - 2000 - 2000 - 3500 - 1500 points. Therefore person 1 would get 10% of the coins person 2 gets 20% person 3 gets 20% person 4 gets 35% person 5 gets 15%. These percents are for the 80% total coins. This shows how the 80% of coins are generated with each block. How will all this be verifiable? Transparency is a MUST
|
|
|
|
ahmed_bodi
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 09:21:23 PM |
|
CureCoin SHA256 Pool will be up for prereg soon. Sadly it wont be on my new platform as its not ready yet, but atleast the current modified MPOS + stratum should be rock solid.
Ahmed
|
Bitrated user: ahmedbodi.
|
|
|
laughingbear
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 09:21:33 PM |
|
Also, when is the release? Ive lost a lot of sleep already.
|
|
|
|
cameronpalte
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 09, 2014, 09:33:09 PM |
|
The 80% of the coins will split proportionally based on how many points are generated. For example if during this time 5 people submit work units for 1000 - 2000 - 2000 - 3500 - 1500 points. Therefore person 1 would get 10% of the coins person 2 gets 20% person 3 gets 20% person 4 gets 35% person 5 gets 15%. These percents are for the 80% total coins. This shows how the 80% of coins are generated with each block. How will all this be verifiable? Transparency is a MUST I agree transparency is a must - it'll probably be just as transparent as Bitcoin mining if not more. Plus F@H publishes its own stats so you could validate how many points people are getting. However, the verification would probably be similar to Bitcoin verification.
|
|
|
|
|