Discuss Cosmology@Home credit here. Please keep discussion along the lines of "your 50 credit scheme goes too far" or "your 50 credit scheme doesn't go too far enough". Please do not discuss credit philosophy and grand credit schemes over which I do not have any control.

Clarification: There's been some confusion about what can and can't be discussed here. You may compare credit here to credit at other projects. However, you may not argue about "being in it for the credits" or "being in it for the science" or any shenanigans like that.

Edit X2: Discussion about credit philosophy. Post there for general credit stuff, not here.

---

Scott Kruger
Project Administrator, Cosmology@Home

Discuss Cosmology@Home credit here. Please keep discussion along the lines of "your 50 credit scheme goes too far" or "your 50 credit scheme doesn't go too far enough". Please do not discuss credit philosophy and grand credit schemes over which I do not have any control.

Nice Futurama reference!

---

Dublin, CA
Team SETI.USA

It's all part of the discussion as long as there are BOINC stats sites.

The "cross-project stats parity" nutz will always tell you that your credit system either awards too much or too little credit. They are never happy.

---

It's all part of the discussion as long as there are BOINC stats sites.

The "cross-project stats parity" nutz will always tell you that your credit system either awards too much or too little credit. They are never happy.

I agree with Angus on this one! IMHO your present 50 value is about right. We have some boxes that claim more and some less. There is also some variation in WUs of course. We use the standard BOINC client on all machines....Cheers, Rog.

It's good.
That's all !! :D

I'm happy to GET credits, it's still alpha phase as far as I know. I wouldn't worry too much Scott, they can always come back when it's mainline.

---

I'm happy to GET credits, it's still alpha phase as far as I know. I wouldn't worry too much Scott, they can always come back when it's mainline.

Is receiving credits a big issue in your situation? In looking at your active hosts all but 2.5% of completed units have been awarded credit which doesn't seem like much.

Please keep the thread on-topic, as Scott described in his original post. Don't go into personal discussions, or into discussions about the importance of credits in general.

---

BOINC.BE: For Belgians who love the smell of glowing red cpu's in the morning
Tutta55's Lair

I've said so in the other thread as well:
On my machine setup (C2D-E6750@3.6, ubuntu7.04, BOINC 5.10.8) I get too much.

I've got 3 projects with more than 40 Credits/hour in my stats: Seti-Opt, QMC and Cosmology.

Next best are around 30 C/h: CPDN, SIMAP, Einstein, SAP, ABC.

I claim about 22 C/h, and the projects roughly in that area are: RCN, malaria, WCG, Xtreme, Chess960 and Lattice

I get reasonably less (down to 14 C/h) on the following ones: Spin, Pirates, Leiden, BRaTS.

For my machine 30 credits would be the "right" amount, but I know that usually Linux isn't taken as the standard, as most crunchers use Windows. In the beginning Linux was a lot faster than Windows per WU, if that's the reason for the difference still, it's OK.

---

Grüße vom Sänger

Fourth try on this, since I keep getting modded off the board:

I *feel* I am not getting enough credit on this project. The credit awards per WU should be much greater for my special CPU/OS combination, perhaps an order of magnitude.

Saenger likes to compare credit granted on this project with other projects, with the idea that all BOINC projects should award credit in lockstep. That's his opinion.

My opinion is that THIS project is free to award whatever they want, withour regard to discussion of credit awarded by other projects.

---

Scott:

Please clarify what CAN and CANNOT be discussed in this thread, so your mod staff don't have to try to interpret for you, with the resulting inconsistent modding.

---

I've said so in the other thread as well:
On my machine setup (C2D-E6750@3.6, ubuntu7.04, BOINC 5.10.8) I get too much.

I've got 3 projects with more than 40 Credits/hour in my stats: Seti-Opt, QMC and Cosmology.

Next best are around 30 C/h: CPDN, SIMAP, Einstein, SAP, ABC.

I claim about 22 C/h, and the projects roughly in that area are: RCN, malaria, WCG, Xtreme, Chess960 and Lattice

I get reasonably less (down to 14 C/h) on the following ones: Spin, Pirates, Leiden, BRaTS.

For my machine 30 credits would be the "right" amount, but I know that usually Linux isn't taken as the standard, as most crunchers use Windows. In the beginning Linux was a lot faster than Windows per WU, if that's the reason for the difference still, it's OK.

You are lucky with that machine then Saenger,
I am a bit envious, but even so you have a range from 14 to 40 over the projects that you do, with Cosmology at the top hence you think it over claims.

My range is from 12 to 30 with Cosmology well below my top but double my lowest, so I think it is claims about right.

My AMD X2 4800+ (Windows) running Seti Enhanced (Optimised by NWSN) only gets 17 to 18 Cr/h about the same as the "standard" app, so no advantage for me. So your 40+ per hour is amazing to me.

QMC on 4800+ around 22 to 30 cr/h, on Opteron 285 (Linux) about 24 to 30 cr/h.

Cosmology on the Opteron 285 averages about 23 to 24 cr/h (a spread of 14 to 40 cr/h per WU, few of the 40 cr/h ones though).

CPDN on the Opteron between 17 to 20 cr/h (20 on the new optimised app).

Einstein on 4800+ is 16 cr/h, on the Opteron was 21 cr/h on old app now 14 cr/h on the new app.

Rosetta around 13 to 16 cr/h on both computers.

LHC around 14 to 15 cr/h on the Opteron.

Spinhenge only gives about 13 cr/h

Lattice only give me around 12 cr/h.

Therfore I say that Cosmology with an average of 23/24 for my standard Opteron (running Linux) is good and balances well with the other projects that I run (although Spinhenge and Lattice need a kick up in granted credit).

It is obvious that faster (later models) and particularly overclocked machines are going to get far more per hour than other older computers, and may think the project is therefore granting too much credit.
From the other side of the coin, from people who thought they had a fast machine till the new ones came along (and there are a great many of these), find they don't get a huge amount of credit compared to the new machines and to them the project may not be granting enough credit.

It is a two edged sword, the balance could tip either way.
Trying to get parity with the other projects will be virually impossible due to the nature of each project. There will be a band from X amount on this project to Y on that project, which is why a number of projects are trying to aim for the middle if they can.

If on this project there were lots of the shorter WU's then that would screw the average to a much higher value and then may be deemed to be granting too much.

So at the moment Cosmology credit for me is good, not too high and not too low.

Thanks and have a good day.

---

Interesting how well the Semprons did.

22/hr on opteron 170 at 2.7 ghz (76 results) - Windows
27/hr on Sempron 3100+ @ 2.33 ghz (64 results) - Linux
26/hr on x2 4200+ @ 2.75 ghz (60 results) - Windows
27/hr on Sempron 2800+ @ 2.33 ghz (64 results) - Linux
19/hr on AthlonXP @ 2.25 ghz (29 results) - Linux
23/hr on opteron 165 @ 2.7 ghz (75 results) - Windows

vs 17/hr for Einstein (blend of all)

and

13/hr for Rosetta (blend of all)


----

Then there's always this, the credit comparison matrix:

http://boinc.netsoft-online.com/e107_plugins/boinc/get_cpcs.php

----

Fill in the blank for the % reduction needed to be on par with other projects if this is the intent.

Scott:

Please clarify what CAN and CANNOT be discussed in this thread, so your mod staff don't have to try to interpret for you, with the resulting inconsistent modding.

I've clarified at the top.

---

Scott Kruger
Project Administrator, Cosmology@Home

Then there's always this, the credit comparison matrix:

http://boinc.netsoft-online.com/e107_plugins/boinc/get_cpcs.php

----

Fill in the blank for the % reduction needed to be on par with other projects if this is the intent.

That is the best place to look for considering credit adjustments. From looking at that chart it appears that this project is granting nearly double the credit it should be.

edit: This fits in with my observations too. Since I am only running this project on my test hosts, my RAC should be about 55 but it is over 100.

---

BOINC WIKI

BOINCing since 2002/12/8

Then there's always this, the credit comparison matrix:

http://boinc.netsoft-online.com/e107_plugins/boinc/get_cpcs.php

----

Fill in the blank for the % reduction needed to be on par with other projects if this is the intent.

That is the best place to look for considering credit adjustments. From looking at that chart it appears that this project is granting nearly double the credit it should be.

edit: This fits in with my observations too. Since I am only running this project on my test hosts, my RAC should be about 55 but it is over 100.

It's Amazing that you can come up with any kind of Observation by only turning in 2 or 3 Wu's every 2 to 3 Day's, yet you think the project should cut the Credits by 50% more.

All I can say is I hope Scott doesn't Shot himself in the Foot like a few of the other Projects have done.

Let's hope independence wins out over herd mentality.

---

For what its worth I have 2 P4HT hosts running the stock Boinc 5.10.13 client that continually claim well over 50 credits (up to 80) on the longest work units. Hosts 524 and 445

These same hosts claim middle of the road credit on benchmark based projects with multi-quorum the likes of LHC and Spinhenge.

---

P4HT, Windows XP, 5.10.13, multi-project: Roughly 15% of all the CaH Wus I've ran claim more than 50 credits. Conversely it means 85% claim less than 50 credits, so I'm happy.

Hopefully this won't be modded as being off topic, In the last 426 wu's processed on my acct 8 claimed a high of 40 credits most were well below that. I don't see a problem,,,,,,,

---