Hello Team,

I just brought online another new 8 Core System (win 2003 64bit)and I'm receiving very low credits compared to my other 8 Core system (Ubuntu 64bit).

It has 2 Physical AMD 8354 Quad Core on WIN Srvr 2003 64Bit. And I'm receiving an average credit of 54.91 compared to 2825.69 on my other 8 core system. I have 2 more processors in route which will bump me up to a 16 core system by this weekend.

I am thinking of placing Ubuntu 64bit but really trying to get this working correctly in Win 2003 64 bit.

Any thoughts?

---

The average credit (RAC) will rise slowly. It is not an immediate score. Give it 1 to 2 weeks for it to give you the real RAC.

---

Good point about the RAC, Chilean, but he really does appear to be completing a lot fewer decoys on one machine. Looking at a few of the task details for that machine I see a lot of restarts following this message:

No heartbeat from core client for 30 sec - exiting

Here's a link to the Unofficial BOINC Wiki entry but unfortunately they don't offer a solution and we are already at the end of my useful contribution. Hopefully someone else will have some ideas on what to do next.

Snags

project reset???

Worth a try i guess as there's no other suggestions! ;)

P.S. Nice machine!

---

It looks like Daniel reset yesterday. Hey it's worth a try though I think I recall from looking through his task list that he appeared to have already tried that before he posted (a page of "client detached" WUs). For those reading along it's this machine that is struggling.
The problem isn't necessarily with the Rosetta application. It could be with the BOINC installation in which case reinstalling BOINC might solve the problem. Or it could be that some other process on that machine is interfering with communications between the Rosetta application and BOINC. You could test this by temporarily attaching to another project. If you get the same messages you know the problem isn't with the Rosetta application and you get another avenue for assistance. SETI is one possibility as its boards are very active with a fair number of troubleshooters on patrol upping your odds that someone has seen, and hopefully solved, this problem before. (Just be sure to subscribe to your thread as it may move off the first page in less than a day.) The BOINC boards are another source of knowledgeable troubleshooters.

Good luck and keep us informed.

Snags

It looks like Daniel reset yesterday. Hey it's worth a try though I think I recall from looking through his task list that he appeared to have already tried that before he posted (a page of "client detached" WUs). For those reading along it's this machine that is struggling.
The problem isn't necessarily with the Rosetta application. It could be with the BOINC installation in which case reinstalling BOINC might solve the problem. Or it could be that some other process on that machine is interfering with communications between the Rosetta application and BOINC. You could test this by temporarily attaching to another project. If you get the same messages you know the problem isn't with the Rosetta application and you get another avenue for assistance. SETI is one possibility as its boards are very active with a fair number of troubleshooters on patrol upping your odds that someone has seen, and hopefully solved, this problem before. (Just be sure to subscribe to your thread as it may move off the first page in less than a day.) The BOINC boards are another source of knowledgeable troubleshooters.

Good luck and keep us informed. Snags

I just checked and his machines are not identical, according to the list so that could be part of the issue, but he is also using a much different version of Boinc on the Linux and new machine. The Linux machine is using Boinc 6.2.18 and the Windows Server machine is using Boinc 6.6.36! Personally I would drop back to the same version on both until this is sorted out, the fewer variables the better.

The credit granted at Rosetta is below par, has been for a very long time.
One of the reasons I have virtually stopped doing this project, which was one of my favorites, my loyalties have moved to other places now.
It seems the way to get good credit on this project is to have a fast recent computer and return the work units very quickly, as first returned gets what they claim, it averages all work after that.

Getting 13 credits for 6 hours work is not inspiring.

Just check Rosetta against World Community Grid, Rosetta has more connected users and computers but has one third the credit output.

---

The credit granted at Rosetta is below par, has been for a very long time.
One of the reasons I have virtually stopped doing this project, which was one of my favorites, my loyalties have moved to other places now.
It seems the way to get good credit on this project is to have a fast recent computer and return the work units very quickly, as first returned gets what they claim, it averages all work after that.

Getting 13 credits for 6 hours work is not inspiring.

Just check Rosetta against World Community Grid, Rosetta has more connected users and computers but has one third the credit output.

I am getting 12 credits for about 30 to 40 minutes of work on Virtual Prairie. They are fixed credits over there.

The credit granted at Rosetta is below par, has been for a very long time.
One of the reasons I have virtually stopped doing this project, which was one of my favorites, my loyalties have moved to other places now.
It seems the way to get good credit on this project is to have a fast recent computer and return the work units very quickly, as first returned gets what they claim, it averages all work after that.

Getting 13 credits for 6 hours work is not inspiring.

Just check Rosetta against World Community Grid, Rosetta has more connected users and computers but has one third the credit output.

I am getting 12 credits for about 30 to 40 minutes of work on Virtual Prairie. They are fixed credits over there.

Well Rosetta can be worse than I have written with 13 credits an hour as a higher return, as it can be as low as 7.5 cr/h (45 credits for 6 hours) or as 'high' as 15.5 cr/h (93 credits for 6 hours).
Still on the low side for returns.

Your figures work out at 24 cr/h to 18 cr/h, so still much higher than Rosetta.

---

The credit granted at Rosetta is below par, has been for a very long time.
One of the reasons I have virtually stopped doing this project, which was one of my favorites, my loyalties have moved to other places now.
It seems the way to get good credit on this project is to have a fast recent computer and return the work units very quickly, as first returned gets what they claim, it averages all work after that.

Getting 13 credits for 6 hours work is not inspiring.

Just check Rosetta against World Community Grid, Rosetta has more connected users and computers but has one third the credit output.

Someone else mentioned the low and decreasing credits here about a month ago, which matched my experience, but since then they seem to have increased 5-10% per WU (claimed and granted).

Looking at the project comparison page at boincstats you can compare the two here and it's a bit confusing for me. The first graph showing credit comparison across all hosts shows WCG delivering 25% more than R@h, but the 2nd graph showing credit per CPU second on shared hosts indicated r@h is delivering 1-2% more.

Can someone tell me which is the most relevant, please? My impression is that the 2nd graph compares the situation in most similar circumstances, in which case r@h does slightly better, whatever the general impression is.

The credit granted at Rosetta is below par, has been for a very long time.
One of the reasons I have virtually stopped doing this project, which was one of my favorites, my loyalties have moved to other places now.
It seems the way to get good credit on this project is to have a fast recent computer and return the work units very quickly, as first returned gets what they claim, it averages all work after that.

Getting 13 credits for 6 hours work is not inspiring.

Just check Rosetta against World Community Grid, Rosetta has more connected users and computers but has one third the credit output.

Someone else mentioned the low and decreasing credits here about a month ago, which matched my experience, but since then they seem to have increased 5-10% per WU (claimed and granted).

Looking at the project comparison page at boincstats you can compare the two here and it's a bit confusing for me. The first graph showing credit comparison across all hosts shows WCG delivering 25% more than R@h, but the 2nd graph showing credit per CPU second on shared hosts indicated r@h is delivering 1-2% more.

Can someone tell me which is the most relevant, please? My impression is that the 2nd graph compares the situation in most similar circumstances, in which case r@h does slightly better, whatever the general impression is.

I am not sure either what those charts mean but this chart may help to better show how Rosetta is doing compared to the Boinc Projects http://boincstats.com/stats/project_cpcs.php

The credit granted at Rosetta is below par, has been for a very long time.
One of the reasons I have virtually stopped doing this project, which was one of my favorites, my loyalties have moved to other places now.
It seems the way to get good credit on this project is to have a fast recent computer and return the work units very quickly, as first returned gets what they claim, it averages all work after that.

Getting 13 credits for 6 hours work is not inspiring.

Just check Rosetta against World Community Grid, Rosetta has more connected users and computers but has one third the credit output.

Looking at the project comparison page at boincstats you can compare the two here and it's a bit confusing for me. The first graph showing credit comparison across all hosts shows WCG delivering 25% more than R@h, but the 2nd graph showing credit per CPU second on shared hosts indicated r@h is delivering 1-2% more.

Can someone tell me which is the most relevant, please? My impression is that the 2nd graph compares the situation in most similar circumstances, in which case r@h does slightly better, whatever the general impression is.

I am not sure either what those charts mean but this chart may help to better show how Rosetta is doing compared to the Boinc Projects http://boincstats.com/stats/project_cpcs.php

Mikey, if you click at the junction of the projects you want to compare in your link (same as the one I provided at the top of the page) it leads directly to the graphs pointed to by lizziebarry if you compare WCG to R@H.

I don't quite know what it means either, but if I were to guess it does seem to indicate WCG awards only 0.9777 of what Rosetta awards per CPU per second when hosts work on both projects. Rosetta beats (a very few) others too, while WCG seems to lose against almost everything.

I don't quite know what to make of that.

---

The first graph compares all hosts that run both or one of the two projects, the second graph only compares the hosts running both projects.

I guess that makes sense. The implication is that WCG has a different profile of machine among its active users, resulting in a higher average award than Rosetta overall, but when the same machine runs both, comparing like with like, it's Rosetta that actually awards slightly more.

Looking at Conan's sig, it looks like MW is his favourite, partly because it awards 2.74 times as much in credits per CPU second (or Rosetta only 0.365 as much). I've just taken a look at MW to see what's so attractive about it and nothing grabbed me at all, but still, who can blame him?

I know it's hardly the most principled thing in the world to do, but why doesn't this project just add 50-100% to its credits? Human nature will do the rest...

The first graph compares all hosts that run both or one of the two projects, the second graph only compares the hosts running both projects.

I guess that makes sense. The implication is that WCG has a different profile of machine among its active users, resulting in a higher average award than Rosetta overall, but when the same machine runs both, comparing like with like, it's Rosetta that actually awards slightly more.

Looking at Conan's sig, it looks like MW is his favourite, partly because it awards 2.74 times as much in credits per CPU second (or Rosetta only 0.365 as much). I've just taken a look at MW to see what's so attractive about it and nothing grabbed me at all, but still, who can blame him?

I know it's hardly the most principled thing in the world to do, but why doesn't this project just add 50-100% to its credits? Human nature will do the rest...

If they just do what you suggest, not that you REALLY suggested actually doing it, alot of people will come potentially causing a TON of work for everyone involved with the project! Lots more units will have to be made and then worked on when they are returned causing a ton of work for the Server potentially requiring more hard drive space, meaning more money spent! More people also means more folks here on the boards requiring more time spent answering all the newbie questions and then censoring all the inappropriate stuff. More people means alot more work, it also means more units getting crunched and the project getting thru its goals faster. A balancing act is what is required and most projects have trouble maintaining a happy medium. Too few credits and people leave, too many credits and too many people come!

The first graph compares all hosts that run both or one of the two projects, the second graph only compares the hosts running both projects.

I guess that makes sense. The implication is that WCG has a different profile of machine among its active users, resulting in a higher average award than Rosetta overall, but when the same machine runs both, comparing like with like, it's Rosetta that actually awards slightly more.

Looking at Conan's sig, it looks like MW is his favourite, partly because it awards 2.74 times as much in credits per CPU second (or Rosetta only 0.365 as much). I've just taken a look at MW to see what's so attractive about it and nothing grabbed me at all, but still, who can blame him?

I know it's hardly the most principled thing in the world to do, but why doesn't this project just add 50-100% to its credits? Human nature will do the rest...

If they just do what you suggest, not that you REALLY suggested actually doing it, alot of people will come potentially causing a TON of work for everyone involved with the project! Lots more units will have to be made and then worked on when they are returned causing a ton of work for the Server potentially requiring more hard drive space, meaning more money spent! More people also means more folks here on the boards requiring more time spent answering all the newbie questions and then censoring all the inappropriate stuff. More people means alot more work, it also means more units getting crunched and the project getting thru its goals faster. A balancing act is what is required and most projects have trouble maintaining a happy medium. Too few credits and people leave, too many credits and too many people come!

I'm fairly sure they'd be happy with 10x the number of computers - DB posted to that extent recently. I think the issue is that they've had a nightmare with altering credits before even though they were doing something completely logical. Increasing the credit award would reduce the value of past work and some people would no doubt be upset by that - an upset minority can have a big impact on a forum - there are old threads here proving that!

---

I know it's hardly the most principled thing in the world to do, but why doesn't this project just add 50-100% to its credits? Human nature will do the rest...

If they just do what you suggest, not that you REALLY suggested actually doing it, alot of people will come potentially causing a TON of work for everyone involved with the project! Lots more units will have to be made and then worked on when they are returned causing a ton of work for the Server potentially requiring more hard drive space, meaning more money spent! More people also means more folks here on the boards requiring more time spent answering all the newbie questions and then censoring all the inappropriate stuff. More people means alot more work, it also means more units getting crunched and the project getting thru its goals faster. A balancing act is what is required and most projects have trouble maintaining a happy medium. Too few credits and people leave, too many credits and too many people come!

Ok, but how about just 10%. See how it affects that comparison page after a few weeks or a month until it gets to the point where it can be a considered a reasonable payer, rather than how it's considered now - a low payer of credit.

As someone once said, it's not like credits mean anything, over and above the influence it has on "credit whores".

Increasing the credit award would reduce the value of past work and some people would no doubt be upset by that - an upset minority can have a big impact on a forum - there are old threads here proving that!

There's an easy way around that. Don't tell anyone! ;)

Contrary to popular opinion, it's not good to talk. Those who think it is are just a bunch of yappers [(c) Ally McBeal]

[quote]Increasing the credit award would reduce the value of past work and some people would no doubt be upset by that - an upset minority can have a big impact on a forum - there are old threads here proving that!

There's an easy way around that. Don't tell anyone! ;)

Contrary to popular opinion, it's not good to talk. Those who think it is are just a bunch of yappers [(c) Ally McBeal]

The project could also come out with a new version of the software and increase credits at that time. That way the old folks just crunched too soon for the project and those that are still hanging around get first dibs on the new system. The new version could be as small as a number change that only reflects the new credit system to as big as a whole rewrite!

The project could also come out with a new version of the software and increase credits at that time. That way the old folks just crunched too soon for the project and those that are still hanging around get first dibs on the new system. The new version could be as small as a number change that only reflects the new credit system to as big as a whole rewrite!

I haven't been around that long, but when MiniRosetta was taking over from Rosetta Beta the credits did change. They slumped by a very large amount (a third?).

So the theory is sound. The practise doesn't quite match.

Did the people who complain about a potential increase in credits now also complain when more recent credits went down? If not, they fully deserve to be ignored.

As I've said before, credits are irrelevant, except to the extent that they influence behaviour in crunching here or elsewhere. Whatever gets more people here is good for me and prioritising the past over the presentfuture is about as meaningless as it gets.

---

The project could also come out with a new version of the software and increase credits at that time. That way the old folks just crunched too soon for the project and those that are still hanging around get first dibs on the new system. The new version could be as small as a number change that only reflects the new credit system to as big as a whole rewrite!

I haven't been around that long, but when MiniRosetta was taking over from Rosetta Beta the credits did change. They slumped by a very large amount (a third?).

So the theory is sound. The practise doesn't quite match.

Did the people who complain about a potential increase in credits now also complain when more recent credits went down? If not, they fully deserve to be ignored.

As I've said before, credits are irrelevant, except to the extent that they influence behaviour in crunching here or elsewhere. Whatever gets more people here is good for me and prioritising the past over the presentfuture is about as meaningless as it gets.

I agree, however the project should agree on a standard that will be applied to all versions of Rosetta. Then we know what to expect. Credit granted from project to project will always vary. I doubt that there ever will be any consistency between all the BOINC projects.

As I've said before, credits are irrelevant, except to the extent that they influence behaviour in crunching here or elsewhere. Whatever gets more people here is good for me and prioritising the past over the presentfuture is about as meaningless as it gets.

I agree, however the project should agree on a standard that will be applied to all versions of Rosetta. Then we know what to expect. Credit granted from project to project will always vary. I doubt that there ever will be any consistency between all the BOINC projects.

Between projects, impossible of course.

Interestingly there was a news post on Milky Way that reads:
"September 11, 2009
Milkyway's credit awarded has been a bit too high for a while now and we're bringing it down to be in line more with other projects. We've been using a credit multiplier of 7.5 per double precision FLOP and other projects are using around 5.4 per double precision FLOP. We're reducing our multiplier to this, so credit should be down around 20-30%. I know everyone hates credit reductions, but it was really needed. We also have some performance improvements for CUDA and ATI applications on the way out, so hopefully it won't be too bad. As usual, try to keep the forum flaming about the credit changes somewhat civil."

The cross-project comparison page on Boincstats put Rosetta credits at 0.3723 of Milky Way, so it was needed. Some people don't know when they're (more than) well off.

---