As you can see I have 4 machines, two of which seem to be claiming slightly less credit per WU than what is granted while the other two claim less and are granted less than half what they claim. Any ideas on what this means? Is it bad and if so, how do I fix it?

---

To save anyone who might help me some trouble, here are some numbers concerning my machines and an approximate ratio of credit claimed to credit granted.

Laptop
Measured floating point speed: 1537.54 million ops/sec
Measured integer speed: 2815.93 million ops/sec
Credit Claimed/Credit Granted: 1.0

Old Mac
Measured floating point speed: 419.78 million ops/sec
Measured integer speed: 1349.27 million ops/sec
Credit Claimed/Credit Granted: 0.4-0.5

New Mac
Measured floating point speed: 2161.2 million ops/sec
Measured integer speed: 5683.94 million ops/sec
Credit Claimed/Credit Granted: 1.0

Old PC
Measured floating point speed: 1580.25 million ops/sec
Measured integer speed: 2688.1 million ops/sec
Credit Claimed/Credit Granted: .8-.9

And by the way, the "Old Mac" above with its dual 866 Mhz processors was hot stuff 4-5 years ago, but based on its numbers seems to be far outclassed by the other machines - even the garden variety PCs. Do these numbers seem right for a machine of this vintage?

---

The 'Old Mac' would be a PPC machine (Power PC), and the Rosetta code doesn't seem to run very fast on them. This is why it's not getting that much credit compared to the other machines.

In addition to that, the Benchmark only measures part of the system performance. So a newer PC with very fast memory and low memory latency will not necessarily get much higher benchmarks, but will run the Rosetta code very well (and hence get more credit).

---

Good thing I'm here for the science.

It's really disheartening to know that my pc should be generating ~80 credits per 6 hour work unit and I wind up with credits in the 40's and 50's. My RAC has gone from 320+ to about 280 now.

But, I guess everyone else is in the same boat.

Just would have expected better from a 2.4GHz A64 3800+, thats all.

Happy Holidays.

As you can see I have 4 machines, two of which seem to be claiming slightly less credit per WU than what is granted while the other two claim less and are granted less than half what they claim. Any ideas on what this means? Is it bad and if so, how do I fix it?

I have a similar AMD, gets around 13-15c/h per core. The memory was the expensive part of the setup, rather than the CPU. The clock is similar (2540), but the memory is DDR500 @ 3,3,2,7,1T.

---

Mine is single core A64, not X2.

Heard various reports about X2's. Some suggest that total performance increases only about 15%-20%.

Do you actually obtain an average total of 30 c/h? Guess it could be the memory.

Still, hurts to see my "claimed" credit reduced by 30%-40% on some wu's.

Guess it all supposed to average out over time.

I have a similar AMD, gets around 13-15c/h per core. The memory was the expensive part of the setup, rather than the CPU. The clock is similar (2540), but the memory is DDR500 @ 3,3,2,7,1T.

It started off life as a 3800, but I upgraded it to an X2 in August when the prices dropped (HostID stayed the same). o/c was about the same each time (2553, 2540), and everything else was the same too (motherboard, memory, etc).

The penalty for running two cores rather than one seems to be about 10% (so it's the same as 1.9 3800s. On CPDN/Coupled I was getting 1.7s/ts on the 3800, but 2 x 1.8s/ts on the X2, lower s/ts is better). If you look at my Synergy RAC, that's from two PCs, the X2 and a nonoverclocked Dell 820D.

The claimed credit is actually about the same or a little lower as the granted credit on workunits of 3h or more - remember that the benchmarks won't reflect the memory. 1h workunits seemed to be inefficient when I tried that, and the claimed credit was higher than the granted credit.

Different Rosetta WUs do give different credit, I think depending on the size of the protein and which algorithms are in use.

I was running one core on CPDN:SAP and the other on Rosetta, so I don't know what it would have scored if both were on Rosetta, but the impression I get is that CPDN:SAP is harder on memory bandwidth than Rosetta (i.e., two CPDNs run a little slower than one CPDN:SAP and one Rosetta). CPDN/Coupled model seems to be in the middle - not such a big hit on memory as SAP.

---

One of the computers on my account is running an AMD duo core 4400+ overclocked. For claimed credit vs. granted credit I am getting on average a 40 point loss per WU and many of the WU's that take many hours end with a computation error at the very end resulting in wasted time. What is the problem?

---

One of the computers on my account is running an AMD duo core 4400+ overclocked. For claimed credit vs. granted credit I am getting on average a 40 point loss per WU and many of the WU's that take many hours end with a computation error at the very end resulting in wasted time. What is the problem?

The erroring unit problem if you are using the screensaver or graphics is known and the advice is to not use the graphics/screensaver (a half fix is in place already, but a true fix will start in the new year)

But if it is not, chances are you maybe overclockied a little too far, ease back a bit.

A quick look at the errors are the two known ones for graphics.
You may also want to try a different Nvidia driver and/or slow down the overclock a little.

---

Team mauisun.org

Dedicated cruncher earning less than 50% of expected credit. ;-(

wu 53916548

Different Rosetta WUs do give different credit, I think depending on the size of the protein and which algorithms are in use.

To save anyone who might help me some trouble, here are some numbers concerning my machines and an approximate ratio of credit claimed to credit granted.

Laptop
Measured floating point speed: 1537.54 million ops/sec
Measured integer speed: 2815.93 million ops/sec
Credit Claimed/Credit Granted: 1.0

Old Mac
Measured floating point speed: 419.78 million ops/sec
Measured integer speed: 1349.27 million ops/sec
Credit Claimed/Credit Granted: 0.4-0.5

New Mac
Measured floating point speed: 2161.2 million ops/sec
Measured integer speed: 5683.94 million ops/sec
Credit Claimed/Credit Granted: 1.0

Old PC
Measured floating point speed: 1580.25 million ops/sec
Measured integer speed: 2688.1 million ops/sec
Credit Claimed/Credit Granted: .8-.9

And by the way, the "Old Mac" above with its dual 866 Mhz processors was hot stuff 4-5 years ago, but based on its numbers seems to be far outclassed by the other machines - even the garden variety PCs. Do these numbers seem right for a machine of this vintage?

>> Hello glans,
What your computers are doing are making a claim for the length of time they have been crunching and use the Boinc benchmark in the calculation for the credit claim.
Rosetta wants science from your work so looks at the number of 'Decoys' that have been generated in the time you have set. The more 'decoys' then usually (depends on the protein complexity) you get more credit.
The longer you let a WU run the more 'decoys' will be generated and more credit will be granted (WUs can be adjusted from 1 hour to 24 hours).
Hence a faster more modern computer can generate more 'decoys' in the same amount of time compared to an older and slower machine and so will claim more credits and also be granted more credits than a slower machine.
As Rosetta no longer uses the Boinc Benchmark the claims made by computers (the claims use the Boinc benchmark) are often no where near what is actually granted. Optimised clients also blow these claims out.
Using some data from November for myself, I find that I get
7 to 7.5 cobblestones/ hour on an Intel Celeron 2.4GHz 512 Gb RAM
9 cobblestones/ hour on an Intel Pentium 4 2.53 @ 2.75 GHz 1 Gb RAM
14 to 14.5 cobblestones/ hour per core on an AMD 4800+ 2.4 GHz 2 Gb RAM
14 to 14.8 cobblestones/ hour per core AMD Opteron 275 2.2 GHz 4 Gb RAM
15 to 15.5 cobblestones/ hour per core AMD Opteron 285 2.6 GHz 2 Gb RAM

As you can see my faster AMD machines generate higher credit claims and get granted more credit as they do more science in the same length on time as my slower Intel machines (my preference is set to crunch a WU for 6 hours).

Hope this helps.

---

@The Bad Penguin: What sort of memory do you have in your PC, what are the timings, 1T or 2T, and does it have two sticks or 4 sticks? (Note that if you have AMDs, it often drops down to the next setting with 4 sticks. So DDR400 may be downgraded to DDR333).

I don't know about Rosetta, but on the CPDN project the memory has a bigger impact on modelling speed (and hence credit) than the processor clock. Looking at the figures you're giving, I suspect memory speed is just as important here. My credit/hour is consistent with Conan's.

When I was overclocking my PC, I optimised it for memory first, and processor second.

---

@The Bad Penguin: What sort of memory do you have in your PC, what are the timings, 1T or 2T, and does it have two sticks or 4 sticks? (Note that if you have AMDs, it often drops down to the next setting with 4 sticks. So DDR400 may be downgraded to DDR333).

I don't know about Rosetta, but on the CPDN project the memory has a bigger impact on modelling speed (and hence credit) than the processor clock. Looking at the figures you're giving, I suspect memory speed is just as important here. My credit/hour is consistent with Conan's.

When I was overclocking my PC, I optimised it for memory first, and processor second.

cache size first & brute force MHz, don't think memory speed effects much, I've ran at varying mem speeds with no noticable difference, ymmv.

It could just be bad luck playing out on the task your getting (if it is over a short period)

---

Team mauisun.org

More than likely, just plain ol bad luck, I'm a guessin'. Unfortunately, seems I've run into a streak of it this past week. Not getting anywhere near expected 320 credits/day.

This is the "new/replacement" Compaq, 2 sticks (which is max) DDR2. CPU is AM2 A64 2.4 GHz 3800+.

I do understand that there will be variation, but what I've seen lately just makes me wonder if something else is going on. (i.e., utilizing credits as a measuring stick to monitor performance).

@The Bad Penguin: What sort of memory do you have in your PC, what are the timings, 1T or 2T, and does it have two sticks or 4 sticks?

cache size first & brute force MHz, don't think memory speed effects much, I've ran at varying mem speeds with no noticable difference, ymmv.

It could just be bad luck playing out on the task your getting (if it is over a short period)

You are actually using two measuring systems here. One is for claimed and the other is for granted. Claimed is based upon the BOINC benchmarks for your PC. If you are using an optimized BOINC client, it often has benchmarks which report higher performance then the standard BOINC client. The granted credit is based upon the average claimed credit per reported model, average based on what is reported back to the Rosetta server. For a given task, there is probably very little variation in credit per model granted over time. But it varies greatly from one task to the next, and rightly so, as you can see some types of tasks run through models very quickly.

So, what you are observing is really a difference between the work used for the benchmark, and the work done to complete Rosetta models. Rosetta is VERY floating point intensive. The benchmark is not as focused on floating point speed. So there is a disparity. And it varies on each machine.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

For better or for worse, I wouldn't know an optimized client if it bit me. Plain ol' vanilla BOINC.

When I first added this cruncher, it was getting on average 320 credits/day.

Recently, I've had 6 hr wu's in the high 20's and low 30's, which would average out to maybe 120 credits/day.

So, I know what the cruncher is capable of, and I know there is variation. Just seems like too much variation.

Thats why I was wondering if others had, at times, seen their expected credit drop by over 50%-60%.

For better or for worse, I wouldn't know an optimized client if it bit me. Plain ol' vanilla BOINC.

When I first added this cruncher, it was getting on average 320 credits/day.

Recently, I've had 6 hr wu's in the high 20's and low 30's, which would average out to maybe 120 credits/day.

So, I know what the cruncher is capable of, and I know there is variation. Just seems like too much variation.

Thats why I was wondering if others had, at times, seen their expected credit drop by over 50%-60%.

The only reason for a continous drop (other than rosetta@home problems) would be throttling, Intel CPU's do this especially the P4's or Cool & Quiet for the AMD, A64 do this, it may sit at a slower speed say the default 800MHz or some inbetween value it thinks is all it needs.
I put that loosely but you get the idea, various CPU and setups have different systems.

[well unless something is really wrong ;-)]

---

Team mauisun.org

For better or for worse, I wouldn't know an optimized client if it bit me. Plain ol' vanilla BOINC.

When I first added this cruncher, it was getting on average 320 credits/day.

Recently, I've had 6 hr wu's in the high 20's and low 30's, which would average out to maybe 120 credits/day.

So, I know what the cruncher is capable of, and I know there is variation. Just seems like too much variation.

Thats why I was wondering if others had, at times, seen their expected credit drop by over 50%-60%.

@ The_Bad_Penguin,
I have an AMD 4800+ which also runs at 2.4GHz (per core). I get an average of about 14 to 14.5 cobblestones per hour. If I was just doing Rosetta (my computer does 5 different ones) I would expect about 14 cr/h x 24 h = 336 cobblestones for the day (per core). I am using a standard Boinc client.
For the past week or more I have not been anywhere near this output due to one core slowing down by about 75%, after every 4th second on one core the other core ticked over one second.
I found that a programme at the service level was loading and running but as it had no name I had no idea what it was/is, so I unticked it, applied and rebooted and now all running at top speed again.
Could it be possible that you have another process running on your computer that is slowing it down just like I had?

---

Thats a good thought! I will look into it.

@ The_Bad_Penguin,

I found that a programme at the service level was loading and running but as it had no name I had no idea what it was/is, so I unticked it, applied and rebooted and now all running at top speed again.

Could it be possible that you have another process running on your computer that is slowing it down just like I had?

Not to pick nits, but I just wanted to point out that if other processes are running, it should take your tasks longer then their target CPU time to complete. I mean if you've got a 3 hour time preference, it would take some fraction longer then 3 hours to accumulate 3hrs of runtime.

Also, if another task is consuming CPU, this would explain a drop in RAC, but not a lower granted credit then claimed. The task knows it didn't have 100% of the CPU, and the time keeping should be very precise. In other words, if you know you didn't crunch as hard on a task, it wouldn't be claiming high credit. Your claim would be lower, and granted should roughly match it.

Ironically, looks like the penguin's last two WUs have claimed and granted in-line now. Go figure!

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/