I looked at the BOINC wiki and it does not explain the issue which is...

I see my claimed credit does not match the granted credit. Ok... with Seti as many as four different results were compared before credit was granted... With Rosetta, which uses a quorum of 1, why does the granted credit not match the claimed credit? ( since only one computer worked on the task).

Just a little confused...

---

Credit is granted per decoy. The amount granted is an average of all the claims per decoy for that workunit type.

---

BOINC WIKI

BOINCing since 2002/12/8

Rather then use a quorum and crunch the same work more then once, Rosetta uses more of a concensus. The claimed credit is based on your BOINC benchmarks reported (which can be triple the machine's actual speed at computing Rosetta work when using an optimized client). The granted credit it based on what others reported and the average of the others crunching the same protein from different starting random seed.

This approach was found to be preferable to just granting everyone the credit their machine claimed. And having more then one machine do exactly the same work just for the sake of verifying the credit is a waste of the computing power. So, this is the "new credit system". The claimed credit is just there as a frame of reference.

---

Rosetta Moderator: Mod.Sense

I have 2 computers running rosetta@home, the first one is a 2G dual core Intel and is granted the same amount of credit(sometimes a little more) for each model it crunches,
The second is a 2G AMD this computer however,only gets half of the credits it is supposed to be asking for/allotted.

Why is this happening ???
I really would like some explanation to this as its getting me frustrated for it to be crunching rosetta @ half the credits i should be earning,

It also crunches SETI@home and receives the full credit asked for/allotted someone show me a way to fix it, or i will just be turning to SETI for crunch credits instead.

---

Credits granted by Rosetta are based on work completed. Each task returned is typically made up of many models computed. Some types of machines are able to crunch a model more efficiently then others. Large CPU cache size is one factor that helps.

Find a task on each of the machines with a very similar name, and compare the number of "decoys" shown in the result file. Now divide the number of CPU seconds by the number of decoys. That shows the number of seconds the machine took for each model. You will see that your AMD machine takes more seconds to complete a model then your other.

The credit "claimed" is based on the BOINC benchmarks, which do not at all measure the ability of your processor cache, nor the amount of memory available. The credit "granted" is based on the amount of work completed. When the two numbers differ significantly, it indicates that the benchmarks do not utilize the portions of the machine needed for Rosetta to run. The benchmarks are not specific to Rosetta. Even within Rosetta, different task types have slightly different performance characteristics.

---

Rosetta Moderator: Mod.Sense

Credits granted by Rosetta are based on work completed. Each task returned is typically made up of many models computed. Some types of machines are able to crunch a model more efficiently then others. Large CPU cache size is one factor that helps.

Find a task on each of the machines with a very similar name, and compare the number of "decoys" shown in the result file. Now divide the number of CPU seconds by the number of decoys. That shows the number of seconds the machine took for each model. You will see that your AMD machine takes more seconds to complete a model then your other.

The credit "claimed" is based on the BOINC benchmarks, which do not at all measure the ability of your processor cache, nor the amount of memory available. The credit "granted" is based on the amount of work completed. When the two numbers differ significantly, it indicates that the benchmarks do not utilize the portions of the machine needed for Rosetta to run. The benchmarks are not specific to Rosetta. Even withing Rosetta, different task types have slightly different performance characteristics.

so in layman's terms my AMD sucks at crunching rosetta and should switch to SETI anyway, and use my 2G dual core Intel for crunching solely rosetta as it can cope with the data more efficiently ?

Another way to put it is that SETI is giving your AMD machine more credit then the work produced would likely deserve (SETI utilizes many of the same areas of the machine as Rosetta, so I'm guessing the two machines probably have a similar disparity on actual work produced on SETI as well).

There is always more to a machine's ability to do work then just the Ghz rating of the CPU. And the type of work you ask the machine to do will impact the results as well. The benchmarks basically just measure the CPU's ability, but in order to get work done, you need more then just CPU.

---

Rosetta Moderator: Mod.Sense

Another way to put it is that SETI is giving your AMD machine more credit then the work produced would likely deserve.

OMG did you just slap me in the face ???(i think you did!!)
you can see the stats of both my machines, there not lacking in any respect as far as i can see, no less than most of the machines crunching for rosetta.

Sorry to say, but a 2GHz AMD system (which equates to a 3200+ or 3400+) isn't really much to write home about in this day and age. Is the AMD system dual core as well, or is it an "old" single core CPU?

The claimed credit vs granted credit is a little confusing for me as well. I've got several server machines that run R@H when they're not being used by my research group. Those machines often get more credit than they claim, but my lower end machines around the lab or at home typically get less credit granted than claimed. This is probably due to the fact that the cluster machines are using Xeon X5355 CPUs, while the other systems are PIII and AthlonXP CPUs (Xeons have plenty of cache).

The AMD 3200+ CPU has 512K of cache, while your new dual core Intel (I'm guessing an E6400 or E4400) has 2MB or 4MB of cache. That alone will contribute to better simulation runs.

OMG did you just slap me in the face ???(i think you did!!)

I was simply trying to express some doubts as to exactly how SETI issues credits. Also expressing doubt as to how many SETI credits each machine earns per unit time.

Your machines are hidden. So, no, I can't see your stats on both machines.

I suppose I could word it all the other way around and say that if the AMD machine gets credit from SETI that is on par with your Intel, then SETI must not be making much use of cache and other system resources the way Rosetta does.

Bottom line is I cannot change the credit system. I can only attempt to explain it. Since it is outcome based, there is no processor-specific bias to it. The project team has no way to know ahead of time what processor type will run best. ("best" meaning the most models per unit time per Ghz).

---

Rosetta Moderator: Mod.Sense

A single 2G AMD Athlon Thunderbird XP 2600+, i know its old and she does very well for what she is, its was just the (mater of factly)your AMD computer sucks and possibly your Intel too, that was just insulting to hear.
Also my preffs are set to show pc's i dont know why there not showing

Well I certainly meant no offense. You started out asking why two machines with similar Ghz rating yield different credit. I hope you now have a feel for why that can happen.

---

Rosetta Moderator: Mod.Sense

that was just insulting to hear.

I don't think there's any insult going on here, just statement of fact; the AMD is not a 'bad' computer, just the benchmarks don't reflect its specific performance at a particular task - this happens all the time with benchmarks; a RAID0 looks great in HDTach benches, but isn't in fact so good in most real life applications.

I believe Einstein work units favour AMD processors more, last time I checked. Also, if you are running an optimized app over at SETI, that will help the AMD score higher.

It's not something to get too stressed over, it just.. is. However, if you're planning on doing a lot of SETI and Rosetta crunching, your next CPU should probably be a large cache one. Similarly, if you were planning something that needed high memory bandwidth, you'd choose AMD.

And if you just want a computer that works, and does some BOINC on the side, either AMD or Intel will be fine ; )

---

Crunching Rosetta as a member of the Whirlpool BOINC Teams