I came to Rosetta from another project just a few days ago. Today I caught my 8-core machine loafing along at 75%. The Win 7 CPU gadget showed me the cpu was starving for RAM. On my previous project 50%,(4GB), was plenty.Changed the setting in TOOLS to 90%,and the CPU is running full out.

---

I came to Rosetta from another project just a few days ago. Today I caught my 8-core machine loafing along at 75%. The Win 7 CPU gadget showed me the cpu was starving for RAM. On my previous project 50%,(4GB), was plenty.Changed the setting in TOOLS to 90%,and the CPU is running full out.

Hi John

Rosetta needs a lot of RAM for some work units - it's the nature of the protein-modelling beast. It's good to hear that you've got that CPU crunching here though - hope you stay around as there's a huge backlog of tasks to crunch through at the lab!

---

I came to Rosetta from another project just a few days ago. Today I caught my 8-core machine loafing along at 75%. The Win 7 CPU gadget showed me the cpu was starving for RAM. On my previous project 50%,(4GB), was plenty.Changed the setting in TOOLS to 90%,and the CPU is running full out.

Hi John

Rosetta needs a lot of RAM for some work units - it's the nature of the protein-modelling beast. It's good to hear that you've got that CPU crunching here though - hope you stay around as there's a huge backlog of tasks to crunch through at the lab!

When they fix the constant errors I will come back, until them my pc's are elsewhere.

I came to Rosetta from another project just a few days ago. Today I caught my 8-core machine loafing along at 75%. The Win 7 CPU gadget showed me the cpu was starving for RAM. On my previous project 50%,(4GB), was plenty.Changed the setting in TOOLS to 90%,and the CPU is running full out.

Hi John

Rosetta needs a lot of RAM for some work units - it's the nature of the protein-modelling beast. It's good to hear that you've got that CPU crunching here though - hope you stay around as there's a huge backlog of tasks to crunch through at the lab!

When they fix the constant errors I will come back, until them my pc's are elsewhere.

Maybe I haven't taken notice lately, but I usually don't seem many errors. And when I do, they usually (if not always) happen a few seconds after the WU starts being crunched.

To OP: RAM is as important as CPU speed in this project.

---

When they fix the constant errors I will come back, until them my pc's are elsewhere.

Maybe I haven't taken notice lately, but I usually don't seem many errors. And when I do, they usually (if not always) happen a few seconds after the WU starts being crunched.

Try this thread:
https://boinc.bakerlab.org/rosetta/forum_thread.php?id=5914

I came to Rosetta from another project just a few days ago. Today I caught my 8-core machine loafing along at 75%. The Win 7 CPU gadget showed me the cpu was starving for RAM. On my previous project 50%,(4GB), was plenty.Changed the setting in TOOLS to 90%,and the CPU is running full out.

Hi John

Rosetta needs a lot of RAM for some work units - it's the nature of the protein-modelling beast. It's good to hear that you've got that CPU crunching here though - hope you stay around as there's a huge backlog of tasks to crunch through at the lab!

When they fix the constant errors I will come back, until them my pc's are elsewhere.

Hello,Mikey:

A few days ago,I looked through my results. I had ~581 successful W/U's and 12 client errors. Call it a 2.5% failure rate. But when I figured in CPU time lost against successful time,It came in at about .002%. I can live with that.
I hope you will look into it as I did. You probably won't get exactly the same numbers,but it may be lot better than you think.

---

I came to Rosetta from another project just a few days ago. Today I caught my 8-core machine loafing along at 75%. The Win 7 CPU gadget showed me the cpu was starving for RAM. On my previous project 50%,(4GB), was plenty.Changed the setting in TOOLS to 90%,and the CPU is running full out.

Hi John

Rosetta needs a lot of RAM for some work units - it's the nature of the protein-modelling beast. It's good to hear that you've got that CPU crunching here though - hope you stay around as there's a huge backlog of tasks to crunch through at the lab!

When they fix the constant errors I will come back, until them my pc's are elsewhere.

Hello,Mikey:

A few days ago,I looked through my results. I had ~581 successful W/U's and 12 client errors. Call it a 2.5% failure rate. But when I figured in CPU time lost against successful time,It came in at about .002%. I can live with that.
I hope you will look into it as I did. You probably won't get exactly the same numbers,but it may be lot better than you think.

I have about 15 pc's running, it varies a little bit based on my basement temps, and most of them used to be here at Rosetta crunching along just fine until POOF there was a problem! The problem seems to be isolated to those pc's that use gpu's to crunch too, I do that in ALL but one of my desktops. That means coming back here will result in 99% errors and that is unacceptable to me!! Too many fish in the sea to stay with one that leaves a bad taste in ones mouth! Now IF Rosetta fixes their problems I WILL come back but not before!!

Rosetta even has a whole thread describing how they do NOT support version 7.0.25 of Boinc, yet it IS the RELEASE version of Boinc!! I WAS using version 6.10.58 or 6.10.60 so that wasn't the cause of my problems, the gpu crunching was. All of my machines are now 7.0.25, except for a couple, one is still on ver 6.10.60, while a couple are on ver 7.0.23 and one is on ver 7.0.27.

The Client Error thread has been going on since 16 Feb 2012 and there is STILL no fix!!! Next week will make it 3 months and STILL no fix!!!

I came to Rosetta from another project just a few days ago. Today I caught my 8-core machine loafing along at 75%. The Win 7 CPU gadget showed me the cpu was starving for RAM. On my previous project 50%,(4GB), was plenty.Changed the setting in TOOLS to 90%,and the CPU is running full out.

Hi John

Rosetta needs a lot of RAM for some work units - it's the nature of the protein-modelling beast. It's good to hear that you've got that CPU crunching here though - hope you stay around as there's a huge backlog of tasks to crunch through at the lab!

When they fix the constant errors I will come back, until them my pc's are elsewhere.

Hello,Mikey:

A few days ago,I looked through my results. I had ~581 successful W/U's and 12 client errors. Call it a 2.5% failure rate. But when I figured in CPU time lost against successful time,It came in at about .002%. I can live with that.
I hope you will look into it as I did. You probably won't get exactly the same numbers,but it may be lot better than you think.

I have about 15 pc's running, it varies a little bit based on my basement temps, and most of them used to be here at Rosetta crunching along just fine until POOF there was a problem! The problem seems to be isolated to those pc's that use gpu's to crunch too, I do that in ALL but one of my desktops. That means coming back here will result in 99% errors and that is unacceptable to me!! Too many fish in the sea to stay with one that leaves a bad taste in ones mouth! Now IF Rosetta fixes their problems I WILL come back but not before!!

Rosetta even has a whole thread describing how they do NOT support version 7.0.25 of Boinc, yet it IS the RELEASE version of Boinc!! I WAS using version 6.10.58 or 6.10.60 so that wasn't the cause of my problems, the gpu crunching was. All of my machines are now 7.0.25, except for a couple, one is still on ver 6.10.60, while a couple are on ver 7.0.23 and one is on ver 7.0.27.

The Client Error thread has been going on since 16 Feb 2012 and there is STILL no fix!!! Next week will make it 3 months and STILL no fix!!!

Hello,Mikey:

Well,the problem would seem to be in the GPU/software interface. I'm running 6 PC's w/34 cores and no GPU's(for crunching).As mentioned,my error rate is trifling.

I wish I could help you,but my tech level is not up to it. Best wishes anyway,and I hope you find a solution.

---

I came to Rosetta from another project just a few days ago. Today I caught my 8-core machine loafing along at 75%. The Win 7 CPU gadget showed me the cpu was starving for RAM. On my previous project 50%,(4GB), was plenty.Changed the setting in TOOLS to 90%,and the CPU is running full out.

Hi John

Rosetta needs a lot of RAM for some work units - it's the nature of the protein-modelling beast. It's good to hear that you've got that CPU crunching here though - hope you stay around as there's a huge backlog of tasks to crunch through at the lab!

When they fix the constant errors I will come back, until them my pc's are elsewhere.

Hello,Mikey:

A few days ago,I looked through my results. I had ~581 successful W/U's and 12 client errors. Call it a 2.5% failure rate. But when I figured in CPU time lost against successful time,It came in at about .002%. I can live with that.
I hope you will look into it as I did. You probably won't get exactly the same numbers,but it may be lot better than you think.

I have about 15 pc's running, it varies a little bit based on my basement temps, and most of them used to be here at Rosetta crunching along just fine until POOF there was a problem! The problem seems to be isolated to those pc's that use gpu's to crunch too, I do that in ALL but one of my desktops. That means coming back here will result in 99% errors and that is unacceptable to me!! Too many fish in the sea to stay with one that leaves a bad taste in ones mouth! Now IF Rosetta fixes their problems I WILL come back but not before!!

Rosetta even has a whole thread describing how they do NOT support version 7.0.25 of Boinc, yet it IS the RELEASE version of Boinc!! I WAS using version 6.10.58 or 6.10.60 so that wasn't the cause of my problems, the gpu crunching was. All of my machines are now 7.0.25, except for a couple, one is still on ver 6.10.60, while a couple are on ver 7.0.23 and one is on ver 7.0.27.

The Client Error thread has been going on since 16 Feb 2012 and there is STILL no fix!!! Next week will make it 3 months and STILL no fix!!!

Hello,Mikey:

Well,the problem would seem to be in the GPU/software interface. I'm running 6 PC's w/34 cores and no GPU's(for crunching).As mentioned,my error rate is trifling.

I wish I could help you,but my tech level is not up to it. Best wishes anyway,and I hope you find a solution.

Ohhhhh John...what you could REALLY do if you expanded into the World of gpu crunching!! A gpu puts out 10X the credit of a cpu for an equivalent amount of work time! Here is just ONE of my units crunched by a gpu for one project:
22043004 15842341 37654 8 May 2012 | 10:59:51 UTC 9 May 2012 | 6:20:02 UTC Completed and validated 2,241.68 9.59 8,758.00 Distributed Rainbow Table Generator (distrrtgen) v3.48 (cuda23)

You can see I crunched for 2,241.68 seconds and got 8,758.00 credits for it, that is 37 minutes of crunching to get over 8700 credits!!! Now that is NOT Rosetta and Rosetta does NOT have a gpu application, BUT Poem does and they also do medical type crunching stuff. Similar but different from Rosetta, but you CAN use a gpu at Poem. You can also use a gpu on over half a dozen projects right now with more coming on line all the time! You can help map the Milky Way, you can work on math problems, Einstein has one too. Soon World Community Grid will have one for helping fight childhood cancer, they are in the early stages of testing it now. World community Grid, WCG for short, is run by IBM so is a very good project with few problems. They have over half a dozen projects within their umbrella and you can choose to work on one or all of them, or anything in between.

Hello,Mikey:

I did Einstein@Home for about 9 months and accumulated 10.6 M points. I have 8 Nvidia GPUs,from 9800GT to 480 GTX.In all that time I never got more than 4 crunching at once,and often only 2.Drivers in,drivers out,going back to earlier versions,troubleshooting,you name it. At Rosetta,I've made adjustments to my run time and buffer.Beyond that,I just look in now and then to see what my PC'c are up to. As money allows,I'm planning to go from 34 cores to 60,but GPU's are not in the picture near term.

---

Hello,Mikey:

I did Einstein@Home for about 9 months and accumulated 10.6 M points. I have 8 Nvidia GPUs,from 9800GT to 480 GTX.In all that time I never got more than 4 crunching at once,and often only 2.Drivers in,drivers out,going back to earlier versions,troubleshooting,you name it. At Rosetta,I've made adjustments to my run time and buffer.Beyond that,I just look in now and then to see what my PC'c are up to. As money allows,I'm planning to go from 34 cores to 60,but GPU's are not in the picture near term.

If you still have those gpu's ready and able to crunch you might try here:
[url]http://boinc.freerainbowtables.com/distrrtgen/index.php[/url

I have ONE Nvidia gpu, a 560 Ti, and it is currently getting almost 500K PER DAY RAC there!!!

---

Hello,Mikey:

I did Einstein@Home for about 9 months and accumulated 10.6 M points. I have 8 Nvidia GPUs,from 9800GT to 480 GTX.In all that time I never got more than 4 crunching at once,and often only 2.Drivers in,drivers out,going back to earlier versions,troubleshooting,you name it. At Rosetta,I've made adjustments to my run time and buffer.Beyond that,I just look in now and then to see what my PC'c are up to. As money allows,I'm planning to go from 34 cores to 60,but GPU's are not in the picture near term.

If you still have those gpu's ready and able to crunch you might try here:
[url]http://boinc.freerainbowtables.com/distrrtgen/index.php[/url

I have ONE Nvidia gpu, a 560 Ti, and it is currently getting almost 500K PER DAY RAC there!!!

Yes. If you want a lot on points, add nvidia gpus to DistRgen.