Text message alerts  |
| We do now send text messages to your cell phone when there is an alert.
When a domain does not respond to pings for more than five minutes a text message is send. Text messages don't come free however, you have to purchase them on this page. Text messages don't work without a cell phone number so you'll have to add that to your account. Finaly you also have to enable them on your checks page. |
|
| User accounts |
| We are still working on migrating services to this new host. Configuration changes are still being made in an effort to block spam issues which have plagued the site. We hope to have these issues resolved quickly. |
|
| CSB project has a new home |
| The Center For Structural Biology's BOINC project is being migrated to a new and improved BOINC server. |
|
| New Version |
| The work available for the V1.6x version of the application will be run down as of today in preparation for a new V1.70 version which introduces an atmosphere of nutrients into the simulation. New work should be available within a week |
|
| Email alerts |
| Email alerts have been turned on.
When a domain does not respond to pings for more than five minutes an alert is emailed. When it comes back up after that another alert is sent. |
|
| Server placed in datacenter |
| The project server was placed in the datacenter and is running properly |
|
| DNS Problem |
| Sorry, Lost the DNS entry for the site for about an hour of so today. DynDNS delete accounts that are not active for more than a month and I neglected to login and refresh the entry this month. |
|
| oops... |
| Ok, so I enabled the project 2 days before going on holiday... this had the unfortunate effect that a small
configuration error resulted in nothing being validated. (It was running the test validator not the production one).
The workunit chain is now clear and credit has been allocated.
|
|
| The project accepts new users |
| Today I enabled user account creation. That's not to say evrrything is stable right now, but it does mean that you can try it out and see what you think. |
|
| Progress is being made |
| Well, a few months on, and I'm making progress. There's now a new task that actually makes PNGs of the images. This is a boinc job, but
only one workunit gets generated every once in a while. I've also been tweaking the software so that now it actually loads workunit parameters
from a database, thus allowing methodical stepping of one parameter at a time to decide the best parameters to run the project at. I think this will
be the first phase of the project, determining the best parameters to feed into the software. Right now each workunit is put through the system 10
times, this will likely be changing because on my dual core 3ghz machine it takes about 1.5-3 hours per run, and so 15-30 hours to get an image,
this isn't an unreasonable amount of time, except it has to return the workunit to the server so it might take a couple of weeks to actually see
results. Also, this software seems to be exhibiting the law of diminishing returns after a few runs, so I doubt that last run really does much for
the image goodness. Of course the thing that really needs to be worked out (and will be my next effort) is writing the algorithm that scores the
images. Until yesterday it used a mixture of the average pixel colour and the number of times each operator has been used to generate the socre. This
did not actually score better images higher than worse ones, and there were also some problems with scaling to more operations (as in, it made the
job files even more of a headache to load/save) so it was dropped back to just average pixel colour. This obviously isn't a very good scoring method,
but it does the trick for testing the processing chain. Once I have figured out a decent scoring algorithm, the project can probably start
accepting users signing up to help process the images. Oh: one more thing, the source code for at least the client will *probably* be released as
open source fairly soon after that, however be warned that it's horribly untidy! |
|
| Still working! New Collaborators - Ben Ferenchak. |
| I'm still here guys! Ben Ferenchak, a student at Widener University, PA, and some of his fellow classmates are collaborating with me on the new direction of this project. I'm very excited to have their assistance in working out some of the mathematics and CUDA development we hope to do. As of now we are still brushing up on our group theory, developing small test applications, and corresponding with Dr. Exoo who has been a great source of information and promising leads. We're still fairly far away from an alpha test but please stay tuned! Thanks everyone! |
|
| Project suspended |
| Having had no time to do anything with the project in the last year,
I'm switching off the scheduler. It's been a learning experience, but
now I'm finding it more efficient to wait for local machines to finish
jobs than putting in the considerable effort required to keep a BOINC
project running.
Thanks for everyon'e help with setting up and testing out the project.
It may make a comeback one day... |
|
| This is taking forever! :) |
| Sorry this is taking so long everyone. I've finally started working again, I'm finally starting to get real job interviews, and the new direction of this project is rather demanding in terms of mathematics I'm unfamiliar with. So those are the reasons that it will take a while to get the next iteration going. Please be patient! I really appreciate it! Thanks. |
|
| Project News |
|
Power outage: Tomorrow (July 15h) this server will go down for 24 hours
due to a planned power outage.
|
|
| Project News July 14, 2009 |
|
Power outage: Tomorrow (July 15h) this server will go down for 24 hours
due to a planned power outage.
|
|
| I'm back! |
| I'm back from my trip so hopefully I'll be able to sit down and pound through a few of these papers this week. As soon as I understand them well enough, I'll be implementing us a new app. On a side note, I may have a few collaborators helping me with a CUDA app so stay tuned for news on that. Thanks Guys! |
|
| Work run-down |
| During July no new simulations will be created, only those that are in progress will be recycled. Once the work has run down I shall go through the results of the C75 simulations I now have (some 350 of them to date) and decide what changes need to be made to the simulator and the simulations. At this time I will also spend some time in trying to reduce the size of the work unit files so that more can produced without consuming all available disk space. I would suggest nobody considers this project as a POTM for July or August. |
|
| I will be gone for a week. Still working on new implementation |
| I'm going to be gone for a week (until the 10th) so if anything goes wrong with the server it won't be fixed until then. I've started investigating new algorithms and approaches and I think I've decided on one but it's going to take me awhile longer to implement it so please hold tight! Thanks guys! |
|
| New Direction Needed |
| Our result turned out a bit short of the current known lower bound, so I believe that the project will have to switch directions... I'm currently looking into a few different things and I'll let everyone know once I choose one. Thanks guys! |
|
| 1 Year Old Today |
| The project is 1 year old as of today. In that time you have processed over 900,000 workunits, produced in excess of 1,300,000 results and ran a number of simulations that have survived for over 3 trillion instructions. |
|
