Is it possible to have an account where we select projects to be given to our account (ala World Community Grid). It's not about research preferences (fight cancer or fight aids) but about selecting the right project for your PC. I rarely click on the BOINC app because most of the time I leave it on it's own, but sometimes I do.
Recently I was shocked to find that my BOINC have spent 23 hours each, on two projects (that's 46 hours combined) with another 36 hours to go (each), and 12 hours left or just one day left or something. I can't remember what the project was. It is sad because of how I use my computer, it better to compute pulsar research (two hours max) or graviational wave (6 hours) or ce-low cpu (1 hour and a half).
My PC is not always on, and on working day is on for shoter hours compared of weekend (mostly on from morning till late night). I don't want to crunch on projet that take long to complete, and will be out of end date anyway.
CE must control which projects are running, otherwise we can't offer the grid for hire. We have a strict ethical policy, by written contract with the charities, so you can be sure it will never be doing anything dodgy.
We have avoided projects with huge work units, for the same reasons you describe. Look on the bright side though - original SETI@home WUs used to take my 486 PC at least two weeks to finish, and climateprediction.net WUs used to take 6 months...!
Charity Engine does not charge for non completed units to my knowledge and currently we are not granting entries for these incomplete units, but I have been tasked to figure out how Charity Engine might best handle these incomplete task to make everyone happy.
The current incomplete units are redistributed to other computers if they are not completed by the deadline as this is the design of the system to ensure the work is completed. As a participant I feel you should be rewarded to some extent for participation even though there wasn't success... though this reward needs to be balanced to ensure there isn't an incentive to cause tasks to fail.
With the full management model we have we have a better opportunity to pair computer characteristics of performance to the work units and project needs. We will be able to attach you to so there are the optimal completion of tasks and adjust this as needed. We have a significant amount of development to do to make this type of efficiency possible. The current return competition rate is extreemly high compared to the incomplete units so there is very little impact on overall performance.
Would it be possible for Charity Engine to start pushing out some work units for Rosetta@home please? I'd like at least some of my CPU time to be used for more diverse medical research.
We disabled Rosetta@home a while back as some users were experiencing some memory use issues. We wanted to try and diagnose those issues before reinitializing for general use again so users are not experiencing any negative performance to their machines. We would like to enable Rosetta@home for users as soon as possible.
Can someone tell me why openMalaria seems to run amok on my machine?
These tasks regularly barge their way into the list of applications and start using up all of the processing at the detriment of the Einstein project's applications.
I have 100% of the processor as the max usage, but the Einstein project apps are repeatedly paused because of the influx of Malaria Control apps, and as soon as a number of those are finished, the next horde flood in and again beat the Einstein apps into a corner where they sit paused for hours on end.
I would be keen to see the Einstein apps that are 70%+ finished (which is usually in excess of six hours) allowed to see it through to the end so that the results can be sent, rather than having to wait it out while there is another malaria outbreak in my current applications.
One for the tech guys, I think. Shouldn't behave as you say - although malaria is set to higher resource share than einstein because we think it's the worthier project.
Still, it should definitely let nearly-complete WUs finish.
Try dropping CPU usage to 80% in the meantime, see if it makes a difference.
I have noticed the same thing happening, as soon as malaria takes control, no other project work units down load, not even the CE work!! I have supended Malaria as I thougth the CE work units were kind of the whole point of this project......
The short answer is the Charity Engine software will sort the processing balance out over time by itself as it adapts to how often your computer runs.
Well this behavior could be a factor of a few things. First the software tracks a debt for each project that your computer runs. Meaning if (hypothetically) you have a 25% share for malariacontrol.net and 75% share for Einstein@home your computer over the time will process proportionally that much for each project. Though the computer only downloads enough work to satisfy a certain amount of time. As it processes tasks for one project other projects begin to count up that time that they are not also processing. So when their debt becomes too large they take over the priority for processsing.
Additionally depending on how close the work units are to their deadline they might take priority and show running, high priority as their status so they can try to complete by their deadline.
Thus if Einstein is processing often and malariacontrol is out of work units to distribute to your computer it will accrue a high debt and at the time it downloads new units these will take priority and begin processing once downloaded even if Einstein@home is 70% of the way done.
This is all handled by the scheduler in Charity Engine. This scheduler is under development and studied for its impact on performance and throughput. In the newer versions of the software called BOINC which Charity Engine is built upon there is a different way these units are handled to balance work to try to get more task completed for each project.
How well each scheduler handles on a particular computer depends on usage patterns. How often it is on, how often it is in use, how often it has Internet connection, etc. Each scheduler has a tradeoff and yet is under development to find the optimal scheduling technique.
@Matt also in regards to CE project tasks specifically... we have a 5 Gigabyte cap on how much the project can use so people who have low Internet caps are not getting high overage charges as a result. This too could play into the not recieving work as well as if there is work available. CE has the highest resource amount by default so when work is available it should run no matter what. Especially if you have a multi core cpu... which most people do these days.
Now I have cleared all off the Malaria work units both Einstein and CE are running. I was going days without seeing any CE workunits when Malaria was running.... I dont think I would have reached my 5gb cap this early in the month...
You'll probably hit the 5GB cap on the CE job within a couple of days - it's a distributed web-crawler app running for our first customer, Datafiniti. Indeed, it's because of that app that we have the cap.
In a few months, those sort of apps will be on unmetered bandwidth only. That will increase performance immensely and be even better for the users too. Needs a fair bit of extra development, but we're on it.
Eg. my own home PC is allowed 120GB per month, but unlimited between midnight-8am. On it's 35Mb/s connection, it could download 240GB every night without me even noticing. As things stand, it does just 5GB then stops like everyone else... :/
Just to let you know CE work units are still being downloaded today...... so I dont think its the 5gb cap thats the problem...... Sounds like Im complaining.... Im not, you are all doing a great job at getting this project up and running and keeping it all flowing nicely.
as Linux user I belong mostly to the surplus capacity of CE, and run most of time your chosen humanitarian projects. A few days ago I've noticed, that we are running actually a combination of PrimeGrid and Rosetta with an Ressource Shares of 100% (PrimeGrid) and 25% (Rosetta).
That has astonished me a little bit, because I can not really understand why the crunching of numbers shall be the worthier humanitarian project, but i don't want to start a debate about this, no panic.
From my point of view I give you some cores and you do what you think is best with them. No problem with that.
What i want to suggest is, that it would be nice when CE would explain those decisions for the surplus capacity projects to us in future. I think it would be a nice idea to share your thoughts about the supported projects and their goals, to make us understand your decisions and keep us informed and participating.
Is it possible to have an account where we select projects to be given to our account (ala World Community Grid). It's not about research preferences (fight cancer or fight aids) but about selecting the right project for your PC. I rarely click on the BOINC app because most of the time I leave it on it's own, but sometimes I do.
Recently I was shocked to find that my BOINC have spent 23 hours each, on two projects (that's 46 hours combined) with another 36 hours to go (each), and 12 hours left or just one day left or something. I can't remember what the project was. It is sad because of how I use my computer, it better to compute pulsar research (two hours max) or graviational wave (6 hours) or ce-low cpu (1 hour and a half).
My PC is not always on, and on working day is on for shoter hours compared of weekend (mostly on from morning till late night). I don't want to crunch on projet that take long to complete, and will be out of end date anyway.
CE must control which projects are running, otherwise we can't offer the grid for hire. We have a strict ethical policy, by written contract with the charities, so you can be sure it will never be doing anything dodgy.
We have avoided projects with huge work units, for the same reasons you describe. Look on the bright side though - original SETI@home WUs used to take my 486 PC at least two weeks to finish, and climateprediction.net WUs used to take 6 months...!
Yeah, but to have expiring WU, will that equal as a loss to the hiring side? They pay for the time spent, but didn't get any result.
Charity Engine does not charge for non completed units to my knowledge and currently we are not granting entries for these incomplete units, but I have been tasked to figure out how Charity Engine might best handle these incomplete task to make everyone happy.
The current incomplete units are redistributed to other computers if they are not completed by the deadline as this is the design of the system to ensure the work is completed. As a participant I feel you should be rewarded to some extent for participation even though there wasn't success... though this reward needs to be balanced to ensure there isn't an incentive to cause tasks to fail.
With the full management model we have we have a better opportunity to pair computer characteristics of performance to the work units and project needs. We will be able to attach you to so there are the optimal completion of tasks and adjust this as needed. We have a significant amount of development to do to make this type of efficiency possible. The current return competition rate is extreemly high compared to the incomplete units so there is very little impact on overall performance.
Would it be possible for Charity Engine to start pushing out some work units for Rosetta@home please? I'd like at least some of my CPU time to be used for more diverse medical research.
We disabled Rosetta@home a while back as some users were experiencing some memory use issues. We wanted to try and diagnose those issues before reinitializing for general use again so users are not experiencing any negative performance to their machines. We would like to enable Rosetta@home for users as soon as possible.
Thank you for your interest. :)
Can someone tell me why openMalaria seems to run amok on my machine?
These tasks regularly barge their way into the list of applications and start using up all of the processing at the detriment of the Einstein project's applications.
I have 100% of the processor as the max usage, but the Einstein project apps are repeatedly paused because of the influx of Malaria Control apps, and as soon as a number of those are finished, the next horde flood in and again beat the Einstein apps into a corner where they sit paused for hours on end.
I would be keen to see the Einstein apps that are 70%+ finished (which is usually in excess of six hours) allowed to see it through to the end so that the results can be sent, rather than having to wait it out while there is another malaria outbreak in my current applications.
One for the tech guys, I think. Shouldn't behave as you say - although malaria is set to higher resource share than einstein because we think it's the worthier project.
Still, it should definitely let nearly-complete WUs finish.
Try dropping CPU usage to 80% in the meantime, see if it makes a difference.
I have noticed the same thing happening, as soon as malaria takes control, no other project work units down load, not even the CE work!! I have supended Malaria as I thougth the CE work units were kind of the whole point of this project......
The short answer is the Charity Engine software will sort the processing balance out over time by itself as it adapts to how often your computer runs.
Well this behavior could be a factor of a few things. First the software tracks a debt for each project that your computer runs. Meaning if (hypothetically) you have a 25% share for malariacontrol.net and 75% share for Einstein@home your computer over the time will process proportionally that much for each project. Though the computer only downloads enough work to satisfy a certain amount of time. As it processes tasks for one project other projects begin to count up that time that they are not also processing. So when their debt becomes too large they take over the priority for processsing.
Additionally depending on how close the work units are to their deadline they might take priority and show running, high priority as their status so they can try to complete by their deadline.
Thus if Einstein is processing often and malariacontrol is out of work units to distribute to your computer it will accrue a high debt and at the time it downloads new units these will take priority and begin processing once downloaded even if Einstein@home is 70% of the way done.
This is all handled by the scheduler in Charity Engine. This scheduler is under development and studied for its impact on performance and throughput. In the newer versions of the software called BOINC which Charity Engine is built upon there is a different way these units are handled to balance work to try to get more task completed for each project.
How well each scheduler handles on a particular computer depends on usage patterns. How often it is on, how often it is in use, how often it has Internet connection, etc. Each scheduler has a tradeoff and yet is under development to find the optimal scheduling technique.
@Matt also in regards to CE project tasks specifically... we have a 5 Gigabyte cap on how much the project can use so people who have low Internet caps are not getting high overage charges as a result. This too could play into the not recieving work as well as if there is work available. CE has the highest resource amount by default so when work is available it should run no matter what. Especially if you have a multi core cpu... which most people do these days.
Now I have cleared all off the Malaria work units both Einstein and CE are running. I was going days without seeing any CE workunits when Malaria was running.... I dont think I would have reached my 5gb cap this early in the month...
Hi Matt,
You'll probably hit the 5GB cap on the CE job within a couple of days - it's a distributed web-crawler app running for our first customer, Datafiniti. Indeed, it's because of that app that we have the cap.
In a few months, those sort of apps will be on unmetered bandwidth only. That will increase performance immensely and be even better for the users too. Needs a fair bit of extra development, but we're on it.
Eg. my own home PC is allowed 120GB per month, but unlimited between midnight-8am. On it's 35Mb/s connection, it could download 240GB every night without me even noticing. As things stand, it does just 5GB then stops like everyone else... :/
Morning Mark,
Just to let you know CE work units are still being downloaded today...... so I dont think its the 5gb cap thats the problem...... Sounds like Im complaining.... Im not, you are all doing a great job at getting this project up and running and keeping it all flowing nicely.
Yo Marc,
as Linux user I belong mostly to the surplus capacity of CE, and run most of time your chosen humanitarian projects. A few days ago I've noticed, that we are running actually a combination of PrimeGrid and Rosetta with an Ressource Shares of 100% (PrimeGrid) and 25% (Rosetta).
That has astonished me a little bit, because I can not really understand why the crunching of numbers shall be the worthier humanitarian project, but i don't want to start a debate about this, no panic.
From my point of view I give you some cores and you do what you think is best with them. No problem with that.
What i want to suggest is, that it would be nice when CE would explain those decisions for the surplus capacity projects to us in future. I think it would be a nice idea to share your thoughts about the supported projects and their goals, to make us understand your decisions and keep us informed and participating.
Best,
Dennis