fredski, I will not comment on your post as others have done it, and I am not sure that the crusade to save the good CPU crunchers against the evil GPU ones is really justified.
But then you should also start a crusade between the CPU crunchers themselves, the ones that do crunch one HCC in 12 hours, and the ones like me that crunches one CPU HCC task in 50 minutes, and on and on. We are not equal in the technological performance sense but we are equal in terms of contribution and what it means helping for a better world. Do not forget that those who spend a lot of money to have stellar performance do not get one dollar for that. They are simply happy to contribute as you should be.

A HCC WU consists of a discrete X-ray Crystallography image that is passed through a standard set of filters and then is done. Thus, to extend and task, to only way to achieve that is to package more into one work unit. With version 7.05, the product is 1 transmission to WCG servers for 2 images processed so to speak.

Hope that answers your question/concern.

FYI, if you wonder about the suffix 1 in HCC1, this was a slip of the pen when the project was created. It does not imply there's a second phase in to biro.

evasion wrote:
Can you please clarify what you mean by 'cutoff'? Work is still being sent to volunteers contributing with CPU's so you should not be cutoff in anyway for the project. If you are unable to receive work for the project, please post in this thread or send an email to support@worldcommunitygrid.org

fredski wrote:
I want to clarify something here. Right now, and throughout much of this year, we have devoted resources to supporting a new platform - GPU computing. There are many algorithms and types of problems that will be able to benefit from the capabilities of GPU computing which will accelerate the pace of research performed here. I believe everyone agrees that increasing the amount of research done here is a good thing.

I dispute the notion that all of this work was done only for Help Conquer Cancer. While they are certainly the immediate beneficiary of this, we hope that they are the first in a long line of projects that will use GPU computing at World Community Grid. This is no different then when we expanded to support BOINC, Linux or Macs. As we move forward, we are watching the trends in use of ARM processors, iOS, and Android and will expand our capabilities to support wherever significant compute resources are available to use.

Having said that, not all projects will benefit from GPU computing and we expect to have a diversity of projects available. Some will only run on CPU and some will run on both CPU and GPU computers. We will need a diverse set of computers available to support these projects and all contributions are welcome and needed.

Secondly, while the work around Help Conquer Cancer has been highly visible over the past months, it is far from the only thing that we are working on. We currently have an upgrade to the AutoDock code in our alpha testing environment. This took one of our team members several months of focused work to get to this point and we are now discussing with the FightAIDS@Home project when they want to shift to the new version. Further down the road, this version of AutoDock is needed to support the GO Fight Against Malaria project.

We are also looking at an upgrade to QCHEM, we are about to start on-boarding our next research project, we have had a long series of discussions with the DDDT2 researchers about the next iteration of workunits for them, we have been extensively scrutinizing the full set of work completed for the HCMD2 project so that before we declare the project finish, we can confirm that all batches and workunits have been transferred successfully to their servers and backed up in our repository. We have intense contractual discussions underway with projects that we are hoping to run next year and we are reviewing proposals from other projects and evaluating their fitness for running here and using qualified experts to review the scientific merits of these new proposals.

We are spending a lot of time working toward adopting the BOINC 7 client so that we can support the the latest version of Windows, Linux and Mac.

Beyond that we are working with the research team who will be giving the next webcast, we are preparing for our anniversary this coming Friday, and many many more routine day to day tasks that are necessary to keep the environment running and operating.

One of our immediate challenges now is that with a wide range of performance difference between devices (this problem was growing even with CPU only devices, but is now much larger with GPU computing as well), we need to be able to scale workunit sizes to accommodate workunit durations that are suitable across that wide range of devices. It will take us a few months to work through this challenge, but it is now a necessary step that we must take.

I hope that you will see GPU computing as a strong addition to increasing the contribution of the volunteers and accelerating the research that we all are doing here at World Community Grid.

Keep up the good work... you're doing just great !!!

A satisfied customer

Crunching for a bettter world...

+1

I for one suspected that the people in the back room must be busy but I didn't quite realize how busy.

Thanks for the info and for the hard work.