He said very clearly that "RosettaLigand was successful", not that Rosetta@home was used in the work done by the pharmeceutical company. He also distinguished between the "private set of compounds" provided by that company, and the "public datasets" that BakerLab had run the program against.

So, the company is using some of the tools, and provided feedback to BakerLab both on the usefulness of the program as it is today, and some areas where it may be further improved.

If such a drug company sent out work units to your machine to study a new drug or "private compound", they would be sending the amino acid sequence of that compound out to the world, essentially making it public domain! Trust me, drug companies don't operate that way.

So, BakerLab is working to improve the tool on a non-profit basis. And drug companies are working to use such tools to attack specific diseases/viruses etc. Read my prior comments on intellectual property rights here

I have a question regarding manuscript 3 from earlier this month. Mr. Baker describes RosettaLigand as very succesful in predicting pharmacologic interactions, but results are not public due to requirements of pharmaceutical companies using these predictions. I am questioning the information on home page of this project - Rosetta@home is not for profit. That was main reason I quit Predictor@home and started Rosetta. Is therefore possible that my computers were working on some results for private company that does not want to disclose the results now? I do not wish any of my machine work for private company. I pay for electricity and hardware, therefore any work for private companies should be optional with option not to participate.

I believe that it was said somewhere last year or so that the Baker Lab is not allowed to receive payments from private companies. Also the pharmaceutical companies would not want their products run on a public network as someone could possibly decipher some of their secrets. In addition it would also probably breach the strict rules that apply to patents.

Under Vista SP1, I had enough free disk space that I was able to improve performance by telling BOINC that it could use more disk space and a higher percentage of the swap space. I have no idea whether that will work under iMac also. Months ago, I was able to improve performance to adding more RAM memory to my computer, but I'm now at the limit this model of computer can handle. Also, note that on machines with a 32-bit operating system, such as most of those sold with less that 4 GB of RAM memory already installed, 4 GB or less is the limit of what you can use even if more is installed; you need to switch to a 64-bit operating system to get beyond that limit. Most computers sold these days have the capability to switch to a 64-bit operating system, but do not come with one already.

OK, so both tasks finished successfully - 25 hrs and 33 hrs. I changed CPU load to 100% to speed up the process. Manager is asking 438 and 332 credits, they are pending so far. I will allow more swap space and see if it helps. I have 1 GB now, but I'll change to 2 GB, because I see MiniRosetta is using 865 MB virtual memory for each WU, so it is almost 2GB together. Thanks for help, I appreciate a lot.

If that helps, but not enough, you may also want to try something more than 2 GB multiplied by number of BOINC projects you are participating in for the swap space but not the RAM memory, like I did, in case BOINC is dividing the available swap space equally among the projects before deciding how much to allocate to each workunit.

iMac Intel Core2Duo 1,86 GHz, OSX 10.5.5, BOINC 5.10.45, CPU use limited to 75%.

This WU http://boinc.bakerlab.org/rosetta/result.php?resultid=211924207 and this one http://boinc.bakerlab.org/rosetta/result.php?resultid=211927271 are running for 18 and 25 hours already using Minirosetta 1.40. Should I reset the project or let it run ? Progress is frozen at 99.079% and 99.353%, remaining time shows on both WU 9 minutes 56 seconds.

That type of apparantly frozen progress is typical when you get a minirosetta workunit that takes significantly more CPU time than predicted. I got one with a predicted time of 6 CPU hours; it actually took 19.5. What predicted length of workunits have you asked for, or have you left it as the default? If I remember correctly, the default has recently been raised to 6 CPU hours. I'd let it run a few more hours before doing anything.

Update: - now the WUs show 22 and 29,5 hours, progress is at 99.245 and 99.441 respectively, so it is progressing but very very slowly. Remaining time still shows 9:56 and 9:57 to completion, iMac seems to be instable a bit, sometimes does not respond for couple seconds, but I let it run and see how far can that get. I had not selected any predicted time or any graphics CPU time - so it should be at the lowest numbers (3 hours, 10% graphics).

Since you have not selected any predicted time, it should the the default instead. You still have the option of selecting 3 hours instead, if that is what you prefer.

I'd let the workunits run for about 5 times the selection of the predicted time when you got those workunits, to see if the automatic cutoff of workunits that take too long is able to work for those workunits. For you, that should be about 30 hours CPU time, but even more wall clock time. I'd then abort them, after making sure that I had reported which workunits this happened to.

Under Vista SP1, I had enough free disk space that I was able to improve performance by telling BOINC that it could use more disk space and a higher percentage of the swap space. I have no idea whether that will work under iMac also. Months ago, I was able to improve performance to adding more RAM memory to my computer, but I'm now at the limit this model of computer can handle. Also, note that on machines with a 32-bit operating system, such as most of those sold with less that 4 GB of RAM memory already installed, 4 GB or less is the limit of what you can use even if more is installed; you need to switch to a 64-bit operating system to get beyond that limit. Most computers sold these days have the capability to switch to a 64-bit operating system, but do not come with one already.

OK, so both tasks finished successfully - 25 hrs and 33 hrs. I changed CPU load to 100% to speed up the process. Manager is asking 438 and 332 credits, they are pending so far. I will allow more swap space and see if it helps. I have 1 GB now, but I'll change to 2 GB, because I see MiniRosetta is using 865 MB virtual memory for each WU, so it is almost 2GB together. Thanks for help, I appreciate a lot.

iMac Intel Core2Duo 1,86 GHz, OSX 10.5.5, BOINC 5.10.45, CPU use limited to 75%.

This WU http://boinc.bakerlab.org/rosetta/result.php?resultid=211924207 and this one http://boinc.bakerlab.org/rosetta/result.php?resultid=211927271 are running for 18 and 25 hours already using Minirosetta 1.40. Should I reset the project or let it run ? Progress is frozen at 99.079% and 99.353%, remaining time shows on both WU 9 minutes 56 seconds.

That type of apparantly frozen progress is typical when you get a minirosetta workunit that takes significantly more CPU time than predicted. I got one with a predicted time of 6 CPU hours; it actually took 19.5. What predicted length of workunits have you asked for, or have you left it as the default? If I remember correctly, the default has recently been raised to 6 CPU hours. I'd let it run a few more hours before doing anything.

Update: - now the WUs show 22 and 29,5 hours, progress is at 99.245 and 99.441 respectively, so it is progressing but very very slowly. Remaining time still shows 9:56 and 9:57 to completion, iMac seems to be instable a bit, sometimes does not respond for couple seconds, but I let it run and see how far can that get. I had not selected any predicted time or any graphics CPU time - so it should be at the lowest numbers (3 hours, 10% graphics).

Since you have not selected any predicted time, it should the the default instead. You still have the option of selecting 3 hours instead, if that is what you prefer.

I'd let the workunits run for about 5 times the selection of the predicted time when you got those workunits, to see if the automatic cutoff of workunits that take too long is able to work for those workunits. For you, that should be about 30 hours CPU time, but even more wall clock time. I'd then abort them, after making sure that I had reported which workunits this happened to.

Under Vista SP1, I had enough free disk space that I was able to improve performance by telling BOINC that it could use more disk space and a higher percentage of the swap space. I have no idea whether that will work under iMac also. Months ago, I was able to improve performance to adding more RAM memory to my computer, but I'm now at the limit this model of computer can handle. Also, note that on machines with a 32-bit operating system, such as most of those sold with less that 4 GB of RAM memory already installed, 4 GB or less is the limit of what you can use even if more is installed; you need to switch to a 64-bit operating system to get beyond that limit. Most computers sold these days have the capability to switch to a 64-bit operating system, but do not come with one already.

iMac Intel Core2Duo 1,86 GHz, OSX 10.5.5, BOINC 5.10.45, CPU use limited to 75%.

This WU http://boinc.bakerlab.org/rosetta/result.php?resultid=211924207 and this one http://boinc.bakerlab.org/rosetta/result.php?resultid=211927271 are running for 18 and 25 hours already using Minirosetta 1.40. Should I reset the project or let it run ? Progress is frozen at 99.079% and 99.353%, remaining time shows on both WU 9 minutes 56 seconds.

That type of apparantly frozen progress is typical when you get a minirosetta workunit that takes significantly more CPU time than predicted. I got one with a predicted time of 6 CPU hours; it actually took 19.5. What predicted length of workunits have you asked for, or have you left it as the default? If I remember correctly, the default has recently been raised to 6 CPU hours. I'd let it run a few more hours before doing anything.