Frequently asked questions about the BOINCSIMAP project

Frequently asked questions about the BOINCSIMAP project

log in

What is the difference between the simap and hmmer applications?

Both application analyze the protein sequences in SIMAP. While the simap application calculates similarities between sequences using the algorithm, hmmer (say it like "hammer") searches for , mathematically described as , within the sequences from SIMAP. See our project description to get an overview what their results are used for.

Do I have to reconfigure my BOINC manager if the BOINCSIMAP project switches between hmmer and simap workunits?

No. Your BOINC client switches automatically to the appropriate application if the workunits from BOINCSIMAP change.

How much compute power does BOINCSIMAP need in the future?

There are two subprojects, represented by the two applications: The similarity calculation will not need the whole compute power of the BOINCSIMAP project any longer, as the complete backlog from 2005/2006 (caused by updates and expansion) could be calculated until september 2006. The second subproject - the protein domain pre-calculation - will need much more compute power as there has been only a small dataset calculated yet. For the year 2007 we expect to have enough work for hmmer to utilize the whole power of BOINCSIMAP.

Why did I get more/less credits than my results claimed?

BOINCSIMAP has implemented workunit-specific credits (also called "fixed credits per workunit"). Therefore the credits you get for valid results is not calculated from the credits your (optimized/non-optimized) BOINC client claimed, but from the amount of sequences/models in the workunit (hmmer and simap) and the number of hits in the result (simap only due to the expensive post-processing of hits). This makes the credits fair and really related to the compute power you have donated to our project. As this is not the default way to grant credits in BOINC projects, we are open for discussions about our workunit-specific credits (please visit our message boards).

Why did I get more/less workunits per day than other users?

BOINCSIMAP uses a dynamic daily result quota. Every host gets a minimal quota of 10 results per CPU and day. Depending on the average turnaround time of the results the quota increases to up to 100 results per CPU and day. Thus we send more results to hosts that return results quickly, which speeds up the SIMAP update process. It you have started to calculate BOINCSIMAP workunits recently, please stay tuned. It takes some time until the average turnaround time of your host has adapted to a stable value.

Can the download/upload volume be decreased for users with low internet bandwidth?

Not really. We are currently working on a 7zip-compression of workunits and results. This will decrease the up-/downloads by several percent. But the data transfer volume is directly related to the technique of biological sequence analysis: it is necessary to transfer the sequence data to be analyzed. Local caching is not an option, as the chance to get several workunits containing the same sequences or domain models is extremely low. Therefore we would like to suggest users with low internet bandwidth to limit the BOINCSIMAP project in the BOINC manager to a percentage that is still acceptable for you.

How about graphics for the hmmer and simap apps?

Yes, we want to implement graphics for our applications, but until now we could not make it because the (few) C++ programmers in our team were busy with the functional part of the applications. Therefore we plan to provide static graphics (for a description see the ) displaying some information about the workunit currently being processed.

Why aren't there application builds optimized for CPU type xyz?

Currently the BOINC client cannot select an application based on the CPU type, but on the major platforms. Therefore we provide applications for the platforms that are typically used within BOINC projects. There are discussions ongoing in the BOINC developer community about an application selection mechanism based on the CPU type. As soon as such a method will be available we will provides different application builds for the most important CPU types.

Where can I ask a different question or discuss a problem?

In our message boards you will find more FAQ and discussion boards. Please feel free to use it actively. Writing to the boards requires a separate login, different from the BOINC account.

Return to SIMAP - Similarity Matrix of Proteins main page

Copyright © 2013 University of Vienna